Research ArticleEstimation of Rock Burst Grades Using Rock Mass Strength
Yalei Wang1 Jinming Xu 1 Junshuai Xu1 and Chuanjiang Zhong2
1Department of Civil Engineering Shanghai University Shanghai 200444 China2China Railway 17 Bureau Group Co Ltd Taiyuan 030006 China
Correspondence should be addressed to Jinming Xu xjming211163com
Received 8 December 2019 Revised 5 April 2020 Accepted 27 April 2020 Published 18 July 2020
Academic Editor Rafael J Bergillos
Copyright copy 2020 Yalei Wang et al is is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited
e traditional rock burst estimation method is usually based on the σc (rock strength) in practice while the actual occurrence ofrock burst depends more on the structure and strength of the rock mass In this work the actual rock bursts occurred in a railwaytunnel project in Western China and the σcm (rock mass strength) was calculated by the generalized HoekndashBrown criterionAccording to the actual situation of rock bursts a modified rock burst estimation criterion using the ratio of σcm to σmax(maximum geostress) was proposed e influence of randomness on the reliability of rock burst estimation criterion wasconsidered e estimation results based on the traditional and modified method were furthermore compared with those of theactual rock bursts e results show that σcm calculated by the generalized HoekndashBrown criterion may be considered well in therock type and strength construction condition and structure features of the rock mass the estimation results of rock burst usingthe ratio of σc to σmax are quite different from the actual situation while those using the ratio of the σcm to σmax coincided relativelywith the actual rock bursts the ratios of σcm to σmax which are greater than 0167 0066 to 0167 0012 to 0066 and less than 0012are corresponded to the slight medium strong and violent grades of the rock bursts respectively the randomness of dataselection has certain influence on the rock burst estimation criterion but the variation range is small the modified estimationcriterion of rock burst proposed in this work has a good reliability e results presented herein are important for tunnelconstruction and the prevention of rock burst in the high geostress areas
1 Introduction
When excavating in ground prone to rock burst it is es-sential to estimate the intensity of rock burst in order toimplement suitable ground control measures A rock burst isa visible rock mass damage to excavations and represents amajor concern in tunneling projects since it is associatedwith casualties and accidents is deadly phenomenoninitiates in high geostress and brittle rock conditions withextremely complex mechanical features is may lead to aviolent expulsion of rock from the surrounding rock massUsually the rock burst occurs with a sudden release of elasticstrain energy in different ways including slabbing spallingejecting or throwing [1 2]
At present rock burst is usually estimated based on theσc (rock strength) in practice and there are manyachievements that have been made in this field For exampleTao [3] proposed a rock burst criterion based on the ratio of
σc to σmax (maximum geostress) while Gu et al [4] proposeda comprehensive criterion for the occurrence of rock burstbased on σc and rock integrity Later on Zhang and Fu [5]modified the TaondashGu criterion Liu et al [6] introduced amembership function to investigate the influence of σc onrock burst occurrence Miao et al [7] used the Tao dis-criminant criterion and combined with other rock param-eters to comprehensively estimate the rock burst gradesZhou et al [8] Li et al [9] and Afraei et al [10] regarded theσc as one of the main factors to estimate rock burst intensityXue et al [11] adopted the ratio of σc to geostress as anindicator to estimate the rock burst
However the actual occurrence of rock burst dependsnot only on the σc but also on the structure of the rock massand the σcm Rock mass structure which may change theevolution mode of rock burst activity is the important factorcausing rock burst [12 13]ere are also some reports aboutthe influence of rock mass structure on rock burst For
HindawiAdvances in Civil EngineeringVolume 2020 Article ID 2517459 14 pageshttpsdoiorg10115520202517459
example Zhou et al [14] found that the rock mass structureplane plays an important role in controlling rock burst andanalyzed the mechanism of rock burst caused by thestructure plane Mohamad et al [15] found that when rockburst occurred a large number of flying rocks are producedin the rock mass plane with small joint spacing and largeaperture Du et al [16] recognized that the weakeningfeature of rock mass structure surface is the main factor toestimate whether rock burst is instantaneous or lagged
At present the HoekndashBrown (H-B) strength criterion forcalculating σcm based on rock mass structure is widelyrecognized Hoek and Brown [17] gave a nonlinear empiricalrelationship between the rock mass type rock mass qualitycriterion and σcm Hoek and Brown [18] obtained the re-lationship between the ultimate principal stress of rock massand the rock mass failure through statistical analysis of rocktriaxial test data and the rock mass tests Sharifzadeh et al[19] calculated the shear strength of rock mass usingHoekndashBrown failure criterion and geological strength cri-terion (GSI)Wu et al [20] proposed amethod to analyze theGSI and the disturbance factor of the rock mass and cal-culated the RMS based on this method Ma et al [21]thereafter estimated the rock burst by a ratio of the H-Bstrength-based σcm to the horizontal stress perpendicular tothe tunnel axis
e traditional rock burst estimation method usuallyestimates the rock burst grades based on the ratio of σc toσmax and the estimated results are often different from theactual situation Nevertheless the actual rock burst dependsnot only on the geostress state and the strength of rock butalso on the rock mass structure and the construction dis-turbance In this work the geostress state was determined bythe hollow inclusion stress relief method e σc was ob-tained by the laboratory uniaxial compressive test e rockmass structure was characterized by GSI according to theactual situation e σcm was calculated by the generalizedH-B strength criterion e 18 measurement points wererandomly selected from the 22 field measurement points tocompare the actual rock burst situation According to theactual construction situation and rock burst situation amodified rock burst estimation criterion based on σmax andσcm was thereafter investigated e reliability of themodified rock burst estimation criterion was verified by theremaining 4 measurement points and the influence ofrandomness of data selection on the stability of the modifiedrock burst estimation criterion was considerede researchresults have important reference value for tunnel con-struction and determination of rock burst preventionmeasures
2 Measurement of Geostress
21 Engineering Background e geostress measurementslocated at the construction site of a railway tunnel inWestern China e measurement area has extremely harshclimate where the mountain ranges extend longitudinally inthe north-south direction and the terrain fluctuates greatlyere are many regional great and active faults with frequentearthquakes high tectonic stresses and well-developed
joints Due to the fact that the surrounding rock mass is hardand brittle it is expected that rock burst may occur e 22geostress measurements are all taken from the railwaytunnel
22 Process of Geostress Measurement e geostress datawere obtained by using the hollow inclusion stress reliefmethod As shown in Figure 1 a great hole inclined by 3deg to5deg with a diameter of 130mm was drilled during the test Aconcentric small hole with a diameter of 36mm was thendrilled from the bottom of the great hole
As shown in Figure 2 after completing the small hole adry towel was used to wipe the hole and to arrange theadhesion After mixing the binder the butter was smearedevenly on the orientation instrument the surface of thehollow inclusion stress gauge was grinded with sandpaperthe binder was poured into the inner cavity of the stressmeter and the stress gauge was propelled into the boreholewith the mounting rod e entire installation process wascontrolled within 20 minutes to maintain the fluidity of thecolloid
After the colloid was completely solidified the inclina-tion and orientation of the borehole the position of theorifice (earth coordinates) and the mounting angle weremeasured As shown in Figure 3 when the stress gauge wasreleased the data were collected by the YJZ-16+ intelligentdigital strain gauge and the core with the strain gauge isremoved e data are received by the KJ327-F type of themine pressure monitoring system and the core with thestress gauge was placed into the confining calibrationinstrument
As shown in Figure 4 after gradually applying theconfining pressure on the core the stress-strain curve wasobtained and the elastic modulus and Poissonrsquos ratio werethereafter computed
In this work the geostress data from the 22 in situmeasurement points were used in the later analysis where 18of them are randomly selected to compare the actual rockburst situation and estimation results by using the tradi-tional method e improvement of the traditional esti-mation method was conducted by analyzing the actual rockburst grades e correctness of the modified method will beverified by the remaining 4 points e field results of theσmax of these 18 measurement points and the correspondinglaboratory results from the uniaxial compressive tests areshown in Table 1
3 Estimation of Rock Burst Tendency
31 Actual Situation of Rock Burst During the tunnelconstruction the appearances of the 18 measurement pointsare shown in Figure 5
Five parameters ie motion sound aging impact onconstruction and influence depth were used to compre-hensively classify the rock burst grades by the code forhydropower engineering geological investigation (CHEGI)suggested by the National Standards Compilation Group ofPeoplersquos Republic of China [22] e measurement point 1
2 Advances in Civil Engineering
was taken as an example to concisely describe the details inestimating rock burst grade At the measurement point 1many large rocks flew out rapidly accompanied by the rock
powder ejection a strong burst sound was heard the rockburst lasted for a long time there is a great influence on theconstruction of the tunnel the rock burst pit is distributed
(a) (b)
Figure 1 e large hole with a diameter of 130mm (a) e large hole drilling process (b) e appearance of the large hole
(a) (b) (c)
(d) (e) (f )
Figure 2 Stress gauge installation process (a) Prepare to clean the hole (b) Mixing glue (c) Apply butter (d) Insert pin (e) Apply glue(f ) Place the stress gauge
Advances in Civil Engineering 3
continuously with the influencing distances of more than2merefore the rock burst at this point was determined atthe strong level Based on the actual situations of rock burststhe rock burst grades at 18 measurement points were de-termined by the above method and are shown in Table 2
32RockBurst EstimationResultsUsingTraditionalCriterione CHEGI criterion based on the ratio of σc to σmax is oftenused to estimate the rock burst in practice e corre-sponding criterion is shown in Table 3
In this work the σcm σc and rock burst grades wereobtained by using the hollow inclusion stress relief methodlaboratory uniaxial compressive tests and the ratio of σc toσmax respectively e results are shown in Table 4 Forcomparison purposes the observation results of field rockbursts are also listed in Table 4 From Table 4 it can be seenthat results by using the CHEGI rock burst estimationcriterion are quite different from those at the field situations
33 Modified Criterion of Rock Burst Estimation In thetraditional criteria for rock burst classification the strengthin the strength-stress ratio method generally refers to theuniaxial compressive strength However the actual occur-rence of rock burst in practice relates much with the rockmass structure and the strength of the corresponding rockmass erefore it is necessary to improve the traditionalmethod to pay attention both to the rock mass structurefeatures and to the σcm in the rock burst estimation
e rock mass structure features are generally charac-terized by the geological strength criterion or GSI Based onGSI and other relevant parameters (such as the disturbancecoefficient and uniaxial compressive strength) the gener-alized H-B strength criterion is often used to calculate theσcm and the ratio of σcm to σc may be used as estimationcriterion of the rock burst erefore the rock burst esti-mation based on σcm will be proposed in the following
331 Determination of GSI e geological strength crite-rion or GSI introduced by Marinos and Hoek [23] andconsidering the structural features weathering conditionand the surface features of the rock mass could better reflectthe geological situation of rock mass As shown in Figure 6based on the discontinuity structure and surface conditionof rock mass the average value of GSI may be estimated In
(a) (b) (c)
Figure 3 Stress relief process (a) Core sampling (b) Data collection (c) e core with stress gauge
Figure 4 Core elastic modulus and Poissonrsquos ratio acquisition
Table 1 σmax and σc at different measurement points
this figure ldquoNArdquo means that it is not applicable within thisrange
As shown in Figure 6 the surface quality of rock massstructure may be divided into five categories using theweathering condition of rock mass and the surface featuresof joints which are Very Good Good Fair Poor and VeryPoor Among them Very Goodmeans very rough fresh andunweathered surface of rock mass Good means rough mildweathered iron surface of rock mass Fair means mediumweathered and altered surface of rock mass Poor meanssmooth highly weathered rock mass surface with a denseoverburden or filler or angular fragments Very Poor meanssmooth severely weathered rock mass surface with a softclay coating or filler e corresponding range of values isfrom 100 to 0 in order of high to low and the higher thevalue the better the quality grade of rock mass surface estructural features of the rock mass are divided into sixcategories using the order of the integrity of the rock massstructure surface which are Intact or Massive Blocky VeryBlocky BlockyDisturbedSeamy Disintegrated and Lam-inated or Sheared Wherein the Intact or Massive means acomplete rock mass or a large rock mass structure with fewlarge spacing and discontinuity the Blocky means a goodand original rock mass structure composed of cubic blocksformed by three mutually orthogonal joint faces VeryBlocky means a partially disturbed rock mass structurewhich composed of multifaceted angular blocks formed byat least 4 sets of joints BlockyDisturbedSeamy means a
(a) (b) (c) (d) (e) (f )
(g) (h) (i) (j) (k) (l)
(m) (n) (o) (p) (q) (r)
Figure 5 Appearances after rock bursts at different measurement locations (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4 (e) Point 5 (f )Point 6 (g) Point 7 (h) Point 8 (i) Point 9 (j) Point 10 (k) Point 11 (l) Point 12 (m) Point 13 (n) Point 14 (o) Point 15 (p) Point 16 (q)Point17 (r) Point 18
Table 2 Field rock burst grades at different measurement points
Table 3 Rock burst grades using the ratio of σc to σmax in CHEGIcriterionlowast
IndexGrades of rock burst
Slight Medium Strong Violentσcσmax 4sim7 2sim4 1sim2 lt1lowastere is no groundwater activity in the area
Advances in Civil Engineering 5
rock mass structure in which a plurality of sets of discon-tinuous surfaces are mutually cut to form an angular rockmass and undergoes a fold activity a layer or a flank planecontinuous Disintegrated means a severely fractured rockmass which contains a mixture of angular and circular rockblocks LaminatedSheared means the lack of a massive rockmass structure due to the weak schistosity or the closespacing of the shear planes e corresponding range ofvalues is from 100 to 0 in order of high to low and the higherthe value the better the integrity of the rock mass
e MarinosndashHoek method in estimating GSI was usedin this work e measurement point 1 was taken as an
example to concisely describe the process in estimating GSIIn the measurement point 1 the structural fissures andunloading fissures are well developed the structural fissuresare often distributed in an ldquoXrdquo shape the joints generally donot extend the fissures are 1 to 10mm wide with a maxi-mum width of 15mm the fissures are mostly half-open andfilled visibly the joint spacing is more than 2m Accordinglythe surface condition of the rock mass at this point wasdetermined as ldquoGoodrdquo and the value was estimated as 62 therock mass structure was determined as a blocky structurewith a value of 71
As shown in Figure 7 when the intersection of thevertical line of the rock surface condition and the horizontalline of the rock structure characteristic was between the twoGSI values linear interpolation was often used for calcu-lation Accordingly the GSI of the measurement point 1 wasestimated as 58 According to the above method the GSIvalues of the remaining measurement points were estimatedand the estimated GSIs at 18 measurement points are listedin Table 5
332 Determination of Relevant Parameters Other relevantparameters mainly relate to the rock mass disturbance andσc e related parameter of the rock mass disturbancedegree is coefficient D which represents the disturbance ofthe rock mass Considering the actual situation of the tunnelexcavationDrsquos of all the measurement points are assumed tobe 05 Rock uniaxial compressive strength was determinedthe σc at each measurement point is determined by thelaboratory uniaxial compression test (see early-mentionedTable 1)
333 Calculation of Rock Mass Strength Using HoekndashBrownCriterion Based on Griffithrsquos theory Hoek et al [24] in-vestigated the relations between the ultimate principal stressin a rock mass and the rock mass failure from the statistical
Table 4 Rock burst grades at various measurement locations using the ratio of σc to σmax
No of measurement points σcσmax Estimated results using CHEGI criterion Actual field grades of rock burst
1 082 Violent Strong2 151 Strong Medium3 380 Medium Slight4 087 Violent Strong5 111 Strong Medium6 048 Violent Strong7 083 Violent Strong8 152 Strong Medium9 326 Slight Medium10 188 Strong Medium11 222 Medium Medium12 117 Strong Strong13 253 Medium Medium14 274 Medium Medium15 299 Medium Medium16 143 Strong Medium17 300 Medium Medium18 135 Strong Medium
Rock mass structureRock mass surface conditions
Verygood
Intactmassive
Blocky
Very blocky
Blockydisturbed
seamy
Disintegrated
Laminatedsheared
90
80
70
60
50
40
30
NA NA
Good Fair Poor Verypoor
20
10NA NA
Figure 6 Estimation of geological strength criterion [23]
6 Advances in Civil Engineering
results of the rock triaxial tests and the rock mass tests andobtained the generalized H-B strength guidelines
In the generalized H-B strength criterion the σcm iscalculated using the following formula
s exp(GSI minus 100)
(9 minus 3D)1113890 1113891 (1)
a [exp(minusGSI15) minus exp(minus203)]
6 + 05 (2)
σcm σcsa (3)
where σc is the uniaxial compressive strength of intact rock sis the empirical parameter reflecting the fracture degree ofthe rock mass ranging from 0 to 1 respectively taking 0 forthe fully fractured rockmass and 1 for the intact rock mass ais the empirical parameter that reflects the features of therock mass GSI is the geological strength criterion of the rockmass D is a parameter reflecting the disturbance degree ofthe field rockmass influencing by external factors and rangesfrom 0 to 1 respectively taking 0 for the undisturbed rockmass and 1 for the completely disturbed rockmass σcm is theuniaxial compressive strength of the rock mass
At measurement point 1 substituting GSI 58 andD 05 into equations (1) and (2) s and a will be 0003698and 05033 respectively Substituting s a and σc intoequation (3) the σcm is 242MPa e calculation results ofthe rock mass strength σcm of 18 measurement points areshown in Table 6
334 Rock Burst Estimation Based on Rock Mass StrengthIn computing the ratio of the rock mass strength σcm to thegeostress Ma et al [21] represented the geostress as thehorizontal stress perpendicular to the tunnel axis In thecurrent study the maximum stress σmax was still used torepresent the geostress considering the availability inpractice and the comparability with the existing specifica-tions (ie the code for hydropower engineering geologicalinvestigation 2016)
e ratios of σcm to σmax at different measurement pointsare listed in Table 7 and shown in Figure 5 e on-site fieldrock burst grades from the actual observations are alsoshown in Table 7
As shown in Table 7 among the 18 randomly chosenpoints the number of measurement points for slight rockburst medium rock burst strong rock burst and violentrock burst is 1 11 6 and 0 respectively Because the pointnumber of violent rock burst is zero it is difficult to ac-curately determine the extent of the ratio for the violent rockburst In order to facilitate the division of the rock burst theratio of σcm to σmax is taken as 0 to be temporarily used toestimate the violent rock burst
From Figure 8 it can be seen that the rock burst is lighterwhen the ratio of σcm to σmax is greater and vice versaerefore three boundary lines may be used to divide theratio of these measurement points into four parts corre-sponding to the grades of the slight medium strong andviolent rock bursts e determination processes of thesethree boundary lines are as follows
90
80
70
60
Figure 7 Schematic diagram of linear interpolation
(a) e determination of the boundary line between theslight and medium rock bursts In this case one and 12measurement points belong respectively to the slightand medium rock bursts Because y3 01850 andy9 01497 represent the minimum and maximumratios of the slight and medium rock bursts respec-tively at the measurement points 3 and 16 the mid-value of y3 and y16 or y 016735 is used as theboundary ratio between these two grades of rock bursts
(b) e determination of the boundary line between themedium and strong rock bursts Because y8 00738and y12 00573 represent the minimum and maxi-mum ratios of the medium and strong rock burstsrespectively at the measurement points 8 and 12 themidvalue of y8 and y12 or y 006555 is used as theboundary ratio between the grades of the rock bursts
(c) e determination of the boundary line between thestrong and violent rock bursts In this case
y6 00236 and y 0 are taken as the minimum andmaximum ratios for the strong and violent rockbursts respectively at the measurement point 6 andothers e midvalue of y6 and 0 or y 001180 isregarded as the boundary ratio between the strongand violent rock bursts
To consider the facility in applications it is better totransfer these threshold values into ones with three digitsafter decimal points Accordingly the ratio intervals aremore than 0167 (0066 0167] (0012 0066] and no morethan 0012 respectively representing the slight mediumstrong and violent rock bursts A rock burst estimationmethod (see Table 8) based on the RMS-to-MG ratio isthereafter obtained
335 Verification of Estimation Criterion for Rock BurstIn order to address the reliability of the abovementionedestimation criterion for rock bursts the remaining 4 field
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
example Zhou et al [14] found that the rock mass structureplane plays an important role in controlling rock burst andanalyzed the mechanism of rock burst caused by thestructure plane Mohamad et al [15] found that when rockburst occurred a large number of flying rocks are producedin the rock mass plane with small joint spacing and largeaperture Du et al [16] recognized that the weakeningfeature of rock mass structure surface is the main factor toestimate whether rock burst is instantaneous or lagged
At present the HoekndashBrown (H-B) strength criterion forcalculating σcm based on rock mass structure is widelyrecognized Hoek and Brown [17] gave a nonlinear empiricalrelationship between the rock mass type rock mass qualitycriterion and σcm Hoek and Brown [18] obtained the re-lationship between the ultimate principal stress of rock massand the rock mass failure through statistical analysis of rocktriaxial test data and the rock mass tests Sharifzadeh et al[19] calculated the shear strength of rock mass usingHoekndashBrown failure criterion and geological strength cri-terion (GSI)Wu et al [20] proposed amethod to analyze theGSI and the disturbance factor of the rock mass and cal-culated the RMS based on this method Ma et al [21]thereafter estimated the rock burst by a ratio of the H-Bstrength-based σcm to the horizontal stress perpendicular tothe tunnel axis
e traditional rock burst estimation method usuallyestimates the rock burst grades based on the ratio of σc toσmax and the estimated results are often different from theactual situation Nevertheless the actual rock burst dependsnot only on the geostress state and the strength of rock butalso on the rock mass structure and the construction dis-turbance In this work the geostress state was determined bythe hollow inclusion stress relief method e σc was ob-tained by the laboratory uniaxial compressive test e rockmass structure was characterized by GSI according to theactual situation e σcm was calculated by the generalizedH-B strength criterion e 18 measurement points wererandomly selected from the 22 field measurement points tocompare the actual rock burst situation According to theactual construction situation and rock burst situation amodified rock burst estimation criterion based on σmax andσcm was thereafter investigated e reliability of themodified rock burst estimation criterion was verified by theremaining 4 measurement points and the influence ofrandomness of data selection on the stability of the modifiedrock burst estimation criterion was considerede researchresults have important reference value for tunnel con-struction and determination of rock burst preventionmeasures
2 Measurement of Geostress
21 Engineering Background e geostress measurementslocated at the construction site of a railway tunnel inWestern China e measurement area has extremely harshclimate where the mountain ranges extend longitudinally inthe north-south direction and the terrain fluctuates greatlyere are many regional great and active faults with frequentearthquakes high tectonic stresses and well-developed
joints Due to the fact that the surrounding rock mass is hardand brittle it is expected that rock burst may occur e 22geostress measurements are all taken from the railwaytunnel
22 Process of Geostress Measurement e geostress datawere obtained by using the hollow inclusion stress reliefmethod As shown in Figure 1 a great hole inclined by 3deg to5deg with a diameter of 130mm was drilled during the test Aconcentric small hole with a diameter of 36mm was thendrilled from the bottom of the great hole
As shown in Figure 2 after completing the small hole adry towel was used to wipe the hole and to arrange theadhesion After mixing the binder the butter was smearedevenly on the orientation instrument the surface of thehollow inclusion stress gauge was grinded with sandpaperthe binder was poured into the inner cavity of the stressmeter and the stress gauge was propelled into the boreholewith the mounting rod e entire installation process wascontrolled within 20 minutes to maintain the fluidity of thecolloid
After the colloid was completely solidified the inclina-tion and orientation of the borehole the position of theorifice (earth coordinates) and the mounting angle weremeasured As shown in Figure 3 when the stress gauge wasreleased the data were collected by the YJZ-16+ intelligentdigital strain gauge and the core with the strain gauge isremoved e data are received by the KJ327-F type of themine pressure monitoring system and the core with thestress gauge was placed into the confining calibrationinstrument
As shown in Figure 4 after gradually applying theconfining pressure on the core the stress-strain curve wasobtained and the elastic modulus and Poissonrsquos ratio werethereafter computed
In this work the geostress data from the 22 in situmeasurement points were used in the later analysis where 18of them are randomly selected to compare the actual rockburst situation and estimation results by using the tradi-tional method e improvement of the traditional esti-mation method was conducted by analyzing the actual rockburst grades e correctness of the modified method will beverified by the remaining 4 points e field results of theσmax of these 18 measurement points and the correspondinglaboratory results from the uniaxial compressive tests areshown in Table 1
3 Estimation of Rock Burst Tendency
31 Actual Situation of Rock Burst During the tunnelconstruction the appearances of the 18 measurement pointsare shown in Figure 5
Five parameters ie motion sound aging impact onconstruction and influence depth were used to compre-hensively classify the rock burst grades by the code forhydropower engineering geological investigation (CHEGI)suggested by the National Standards Compilation Group ofPeoplersquos Republic of China [22] e measurement point 1
2 Advances in Civil Engineering
was taken as an example to concisely describe the details inestimating rock burst grade At the measurement point 1many large rocks flew out rapidly accompanied by the rock
powder ejection a strong burst sound was heard the rockburst lasted for a long time there is a great influence on theconstruction of the tunnel the rock burst pit is distributed
(a) (b)
Figure 1 e large hole with a diameter of 130mm (a) e large hole drilling process (b) e appearance of the large hole
(a) (b) (c)
(d) (e) (f )
Figure 2 Stress gauge installation process (a) Prepare to clean the hole (b) Mixing glue (c) Apply butter (d) Insert pin (e) Apply glue(f ) Place the stress gauge
Advances in Civil Engineering 3
continuously with the influencing distances of more than2merefore the rock burst at this point was determined atthe strong level Based on the actual situations of rock burststhe rock burst grades at 18 measurement points were de-termined by the above method and are shown in Table 2
32RockBurst EstimationResultsUsingTraditionalCriterione CHEGI criterion based on the ratio of σc to σmax is oftenused to estimate the rock burst in practice e corre-sponding criterion is shown in Table 3
In this work the σcm σc and rock burst grades wereobtained by using the hollow inclusion stress relief methodlaboratory uniaxial compressive tests and the ratio of σc toσmax respectively e results are shown in Table 4 Forcomparison purposes the observation results of field rockbursts are also listed in Table 4 From Table 4 it can be seenthat results by using the CHEGI rock burst estimationcriterion are quite different from those at the field situations
33 Modified Criterion of Rock Burst Estimation In thetraditional criteria for rock burst classification the strengthin the strength-stress ratio method generally refers to theuniaxial compressive strength However the actual occur-rence of rock burst in practice relates much with the rockmass structure and the strength of the corresponding rockmass erefore it is necessary to improve the traditionalmethod to pay attention both to the rock mass structurefeatures and to the σcm in the rock burst estimation
e rock mass structure features are generally charac-terized by the geological strength criterion or GSI Based onGSI and other relevant parameters (such as the disturbancecoefficient and uniaxial compressive strength) the gener-alized H-B strength criterion is often used to calculate theσcm and the ratio of σcm to σc may be used as estimationcriterion of the rock burst erefore the rock burst esti-mation based on σcm will be proposed in the following
331 Determination of GSI e geological strength crite-rion or GSI introduced by Marinos and Hoek [23] andconsidering the structural features weathering conditionand the surface features of the rock mass could better reflectthe geological situation of rock mass As shown in Figure 6based on the discontinuity structure and surface conditionof rock mass the average value of GSI may be estimated In
(a) (b) (c)
Figure 3 Stress relief process (a) Core sampling (b) Data collection (c) e core with stress gauge
Figure 4 Core elastic modulus and Poissonrsquos ratio acquisition
Table 1 σmax and σc at different measurement points
this figure ldquoNArdquo means that it is not applicable within thisrange
As shown in Figure 6 the surface quality of rock massstructure may be divided into five categories using theweathering condition of rock mass and the surface featuresof joints which are Very Good Good Fair Poor and VeryPoor Among them Very Goodmeans very rough fresh andunweathered surface of rock mass Good means rough mildweathered iron surface of rock mass Fair means mediumweathered and altered surface of rock mass Poor meanssmooth highly weathered rock mass surface with a denseoverburden or filler or angular fragments Very Poor meanssmooth severely weathered rock mass surface with a softclay coating or filler e corresponding range of values isfrom 100 to 0 in order of high to low and the higher thevalue the better the quality grade of rock mass surface estructural features of the rock mass are divided into sixcategories using the order of the integrity of the rock massstructure surface which are Intact or Massive Blocky VeryBlocky BlockyDisturbedSeamy Disintegrated and Lam-inated or Sheared Wherein the Intact or Massive means acomplete rock mass or a large rock mass structure with fewlarge spacing and discontinuity the Blocky means a goodand original rock mass structure composed of cubic blocksformed by three mutually orthogonal joint faces VeryBlocky means a partially disturbed rock mass structurewhich composed of multifaceted angular blocks formed byat least 4 sets of joints BlockyDisturbedSeamy means a
(a) (b) (c) (d) (e) (f )
(g) (h) (i) (j) (k) (l)
(m) (n) (o) (p) (q) (r)
Figure 5 Appearances after rock bursts at different measurement locations (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4 (e) Point 5 (f )Point 6 (g) Point 7 (h) Point 8 (i) Point 9 (j) Point 10 (k) Point 11 (l) Point 12 (m) Point 13 (n) Point 14 (o) Point 15 (p) Point 16 (q)Point17 (r) Point 18
Table 2 Field rock burst grades at different measurement points
Table 3 Rock burst grades using the ratio of σc to σmax in CHEGIcriterionlowast
IndexGrades of rock burst
Slight Medium Strong Violentσcσmax 4sim7 2sim4 1sim2 lt1lowastere is no groundwater activity in the area
Advances in Civil Engineering 5
rock mass structure in which a plurality of sets of discon-tinuous surfaces are mutually cut to form an angular rockmass and undergoes a fold activity a layer or a flank planecontinuous Disintegrated means a severely fractured rockmass which contains a mixture of angular and circular rockblocks LaminatedSheared means the lack of a massive rockmass structure due to the weak schistosity or the closespacing of the shear planes e corresponding range ofvalues is from 100 to 0 in order of high to low and the higherthe value the better the integrity of the rock mass
e MarinosndashHoek method in estimating GSI was usedin this work e measurement point 1 was taken as an
example to concisely describe the process in estimating GSIIn the measurement point 1 the structural fissures andunloading fissures are well developed the structural fissuresare often distributed in an ldquoXrdquo shape the joints generally donot extend the fissures are 1 to 10mm wide with a maxi-mum width of 15mm the fissures are mostly half-open andfilled visibly the joint spacing is more than 2m Accordinglythe surface condition of the rock mass at this point wasdetermined as ldquoGoodrdquo and the value was estimated as 62 therock mass structure was determined as a blocky structurewith a value of 71
As shown in Figure 7 when the intersection of thevertical line of the rock surface condition and the horizontalline of the rock structure characteristic was between the twoGSI values linear interpolation was often used for calcu-lation Accordingly the GSI of the measurement point 1 wasestimated as 58 According to the above method the GSIvalues of the remaining measurement points were estimatedand the estimated GSIs at 18 measurement points are listedin Table 5
332 Determination of Relevant Parameters Other relevantparameters mainly relate to the rock mass disturbance andσc e related parameter of the rock mass disturbancedegree is coefficient D which represents the disturbance ofthe rock mass Considering the actual situation of the tunnelexcavationDrsquos of all the measurement points are assumed tobe 05 Rock uniaxial compressive strength was determinedthe σc at each measurement point is determined by thelaboratory uniaxial compression test (see early-mentionedTable 1)
333 Calculation of Rock Mass Strength Using HoekndashBrownCriterion Based on Griffithrsquos theory Hoek et al [24] in-vestigated the relations between the ultimate principal stressin a rock mass and the rock mass failure from the statistical
Table 4 Rock burst grades at various measurement locations using the ratio of σc to σmax
No of measurement points σcσmax Estimated results using CHEGI criterion Actual field grades of rock burst
1 082 Violent Strong2 151 Strong Medium3 380 Medium Slight4 087 Violent Strong5 111 Strong Medium6 048 Violent Strong7 083 Violent Strong8 152 Strong Medium9 326 Slight Medium10 188 Strong Medium11 222 Medium Medium12 117 Strong Strong13 253 Medium Medium14 274 Medium Medium15 299 Medium Medium16 143 Strong Medium17 300 Medium Medium18 135 Strong Medium
Rock mass structureRock mass surface conditions
Verygood
Intactmassive
Blocky
Very blocky
Blockydisturbed
seamy
Disintegrated
Laminatedsheared
90
80
70
60
50
40
30
NA NA
Good Fair Poor Verypoor
20
10NA NA
Figure 6 Estimation of geological strength criterion [23]
6 Advances in Civil Engineering
results of the rock triaxial tests and the rock mass tests andobtained the generalized H-B strength guidelines
In the generalized H-B strength criterion the σcm iscalculated using the following formula
s exp(GSI minus 100)
(9 minus 3D)1113890 1113891 (1)
a [exp(minusGSI15) minus exp(minus203)]
6 + 05 (2)
σcm σcsa (3)
where σc is the uniaxial compressive strength of intact rock sis the empirical parameter reflecting the fracture degree ofthe rock mass ranging from 0 to 1 respectively taking 0 forthe fully fractured rockmass and 1 for the intact rock mass ais the empirical parameter that reflects the features of therock mass GSI is the geological strength criterion of the rockmass D is a parameter reflecting the disturbance degree ofthe field rockmass influencing by external factors and rangesfrom 0 to 1 respectively taking 0 for the undisturbed rockmass and 1 for the completely disturbed rockmass σcm is theuniaxial compressive strength of the rock mass
At measurement point 1 substituting GSI 58 andD 05 into equations (1) and (2) s and a will be 0003698and 05033 respectively Substituting s a and σc intoequation (3) the σcm is 242MPa e calculation results ofthe rock mass strength σcm of 18 measurement points areshown in Table 6
334 Rock Burst Estimation Based on Rock Mass StrengthIn computing the ratio of the rock mass strength σcm to thegeostress Ma et al [21] represented the geostress as thehorizontal stress perpendicular to the tunnel axis In thecurrent study the maximum stress σmax was still used torepresent the geostress considering the availability inpractice and the comparability with the existing specifica-tions (ie the code for hydropower engineering geologicalinvestigation 2016)
e ratios of σcm to σmax at different measurement pointsare listed in Table 7 and shown in Figure 5 e on-site fieldrock burst grades from the actual observations are alsoshown in Table 7
As shown in Table 7 among the 18 randomly chosenpoints the number of measurement points for slight rockburst medium rock burst strong rock burst and violentrock burst is 1 11 6 and 0 respectively Because the pointnumber of violent rock burst is zero it is difficult to ac-curately determine the extent of the ratio for the violent rockburst In order to facilitate the division of the rock burst theratio of σcm to σmax is taken as 0 to be temporarily used toestimate the violent rock burst
From Figure 8 it can be seen that the rock burst is lighterwhen the ratio of σcm to σmax is greater and vice versaerefore three boundary lines may be used to divide theratio of these measurement points into four parts corre-sponding to the grades of the slight medium strong andviolent rock bursts e determination processes of thesethree boundary lines are as follows
90
80
70
60
Figure 7 Schematic diagram of linear interpolation
(a) e determination of the boundary line between theslight and medium rock bursts In this case one and 12measurement points belong respectively to the slightand medium rock bursts Because y3 01850 andy9 01497 represent the minimum and maximumratios of the slight and medium rock bursts respec-tively at the measurement points 3 and 16 the mid-value of y3 and y16 or y 016735 is used as theboundary ratio between these two grades of rock bursts
(b) e determination of the boundary line between themedium and strong rock bursts Because y8 00738and y12 00573 represent the minimum and maxi-mum ratios of the medium and strong rock burstsrespectively at the measurement points 8 and 12 themidvalue of y8 and y12 or y 006555 is used as theboundary ratio between the grades of the rock bursts
(c) e determination of the boundary line between thestrong and violent rock bursts In this case
y6 00236 and y 0 are taken as the minimum andmaximum ratios for the strong and violent rockbursts respectively at the measurement point 6 andothers e midvalue of y6 and 0 or y 001180 isregarded as the boundary ratio between the strongand violent rock bursts
To consider the facility in applications it is better totransfer these threshold values into ones with three digitsafter decimal points Accordingly the ratio intervals aremore than 0167 (0066 0167] (0012 0066] and no morethan 0012 respectively representing the slight mediumstrong and violent rock bursts A rock burst estimationmethod (see Table 8) based on the RMS-to-MG ratio isthereafter obtained
335 Verification of Estimation Criterion for Rock BurstIn order to address the reliability of the abovementionedestimation criterion for rock bursts the remaining 4 field
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
was taken as an example to concisely describe the details inestimating rock burst grade At the measurement point 1many large rocks flew out rapidly accompanied by the rock
powder ejection a strong burst sound was heard the rockburst lasted for a long time there is a great influence on theconstruction of the tunnel the rock burst pit is distributed
(a) (b)
Figure 1 e large hole with a diameter of 130mm (a) e large hole drilling process (b) e appearance of the large hole
(a) (b) (c)
(d) (e) (f )
Figure 2 Stress gauge installation process (a) Prepare to clean the hole (b) Mixing glue (c) Apply butter (d) Insert pin (e) Apply glue(f ) Place the stress gauge
Advances in Civil Engineering 3
continuously with the influencing distances of more than2merefore the rock burst at this point was determined atthe strong level Based on the actual situations of rock burststhe rock burst grades at 18 measurement points were de-termined by the above method and are shown in Table 2
32RockBurst EstimationResultsUsingTraditionalCriterione CHEGI criterion based on the ratio of σc to σmax is oftenused to estimate the rock burst in practice e corre-sponding criterion is shown in Table 3
In this work the σcm σc and rock burst grades wereobtained by using the hollow inclusion stress relief methodlaboratory uniaxial compressive tests and the ratio of σc toσmax respectively e results are shown in Table 4 Forcomparison purposes the observation results of field rockbursts are also listed in Table 4 From Table 4 it can be seenthat results by using the CHEGI rock burst estimationcriterion are quite different from those at the field situations
33 Modified Criterion of Rock Burst Estimation In thetraditional criteria for rock burst classification the strengthin the strength-stress ratio method generally refers to theuniaxial compressive strength However the actual occur-rence of rock burst in practice relates much with the rockmass structure and the strength of the corresponding rockmass erefore it is necessary to improve the traditionalmethod to pay attention both to the rock mass structurefeatures and to the σcm in the rock burst estimation
e rock mass structure features are generally charac-terized by the geological strength criterion or GSI Based onGSI and other relevant parameters (such as the disturbancecoefficient and uniaxial compressive strength) the gener-alized H-B strength criterion is often used to calculate theσcm and the ratio of σcm to σc may be used as estimationcriterion of the rock burst erefore the rock burst esti-mation based on σcm will be proposed in the following
331 Determination of GSI e geological strength crite-rion or GSI introduced by Marinos and Hoek [23] andconsidering the structural features weathering conditionand the surface features of the rock mass could better reflectthe geological situation of rock mass As shown in Figure 6based on the discontinuity structure and surface conditionof rock mass the average value of GSI may be estimated In
(a) (b) (c)
Figure 3 Stress relief process (a) Core sampling (b) Data collection (c) e core with stress gauge
Figure 4 Core elastic modulus and Poissonrsquos ratio acquisition
Table 1 σmax and σc at different measurement points
this figure ldquoNArdquo means that it is not applicable within thisrange
As shown in Figure 6 the surface quality of rock massstructure may be divided into five categories using theweathering condition of rock mass and the surface featuresof joints which are Very Good Good Fair Poor and VeryPoor Among them Very Goodmeans very rough fresh andunweathered surface of rock mass Good means rough mildweathered iron surface of rock mass Fair means mediumweathered and altered surface of rock mass Poor meanssmooth highly weathered rock mass surface with a denseoverburden or filler or angular fragments Very Poor meanssmooth severely weathered rock mass surface with a softclay coating or filler e corresponding range of values isfrom 100 to 0 in order of high to low and the higher thevalue the better the quality grade of rock mass surface estructural features of the rock mass are divided into sixcategories using the order of the integrity of the rock massstructure surface which are Intact or Massive Blocky VeryBlocky BlockyDisturbedSeamy Disintegrated and Lam-inated or Sheared Wherein the Intact or Massive means acomplete rock mass or a large rock mass structure with fewlarge spacing and discontinuity the Blocky means a goodand original rock mass structure composed of cubic blocksformed by three mutually orthogonal joint faces VeryBlocky means a partially disturbed rock mass structurewhich composed of multifaceted angular blocks formed byat least 4 sets of joints BlockyDisturbedSeamy means a
(a) (b) (c) (d) (e) (f )
(g) (h) (i) (j) (k) (l)
(m) (n) (o) (p) (q) (r)
Figure 5 Appearances after rock bursts at different measurement locations (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4 (e) Point 5 (f )Point 6 (g) Point 7 (h) Point 8 (i) Point 9 (j) Point 10 (k) Point 11 (l) Point 12 (m) Point 13 (n) Point 14 (o) Point 15 (p) Point 16 (q)Point17 (r) Point 18
Table 2 Field rock burst grades at different measurement points
Table 3 Rock burst grades using the ratio of σc to σmax in CHEGIcriterionlowast
IndexGrades of rock burst
Slight Medium Strong Violentσcσmax 4sim7 2sim4 1sim2 lt1lowastere is no groundwater activity in the area
Advances in Civil Engineering 5
rock mass structure in which a plurality of sets of discon-tinuous surfaces are mutually cut to form an angular rockmass and undergoes a fold activity a layer or a flank planecontinuous Disintegrated means a severely fractured rockmass which contains a mixture of angular and circular rockblocks LaminatedSheared means the lack of a massive rockmass structure due to the weak schistosity or the closespacing of the shear planes e corresponding range ofvalues is from 100 to 0 in order of high to low and the higherthe value the better the integrity of the rock mass
e MarinosndashHoek method in estimating GSI was usedin this work e measurement point 1 was taken as an
example to concisely describe the process in estimating GSIIn the measurement point 1 the structural fissures andunloading fissures are well developed the structural fissuresare often distributed in an ldquoXrdquo shape the joints generally donot extend the fissures are 1 to 10mm wide with a maxi-mum width of 15mm the fissures are mostly half-open andfilled visibly the joint spacing is more than 2m Accordinglythe surface condition of the rock mass at this point wasdetermined as ldquoGoodrdquo and the value was estimated as 62 therock mass structure was determined as a blocky structurewith a value of 71
As shown in Figure 7 when the intersection of thevertical line of the rock surface condition and the horizontalline of the rock structure characteristic was between the twoGSI values linear interpolation was often used for calcu-lation Accordingly the GSI of the measurement point 1 wasestimated as 58 According to the above method the GSIvalues of the remaining measurement points were estimatedand the estimated GSIs at 18 measurement points are listedin Table 5
332 Determination of Relevant Parameters Other relevantparameters mainly relate to the rock mass disturbance andσc e related parameter of the rock mass disturbancedegree is coefficient D which represents the disturbance ofthe rock mass Considering the actual situation of the tunnelexcavationDrsquos of all the measurement points are assumed tobe 05 Rock uniaxial compressive strength was determinedthe σc at each measurement point is determined by thelaboratory uniaxial compression test (see early-mentionedTable 1)
333 Calculation of Rock Mass Strength Using HoekndashBrownCriterion Based on Griffithrsquos theory Hoek et al [24] in-vestigated the relations between the ultimate principal stressin a rock mass and the rock mass failure from the statistical
Table 4 Rock burst grades at various measurement locations using the ratio of σc to σmax
No of measurement points σcσmax Estimated results using CHEGI criterion Actual field grades of rock burst
1 082 Violent Strong2 151 Strong Medium3 380 Medium Slight4 087 Violent Strong5 111 Strong Medium6 048 Violent Strong7 083 Violent Strong8 152 Strong Medium9 326 Slight Medium10 188 Strong Medium11 222 Medium Medium12 117 Strong Strong13 253 Medium Medium14 274 Medium Medium15 299 Medium Medium16 143 Strong Medium17 300 Medium Medium18 135 Strong Medium
Rock mass structureRock mass surface conditions
Verygood
Intactmassive
Blocky
Very blocky
Blockydisturbed
seamy
Disintegrated
Laminatedsheared
90
80
70
60
50
40
30
NA NA
Good Fair Poor Verypoor
20
10NA NA
Figure 6 Estimation of geological strength criterion [23]
6 Advances in Civil Engineering
results of the rock triaxial tests and the rock mass tests andobtained the generalized H-B strength guidelines
In the generalized H-B strength criterion the σcm iscalculated using the following formula
s exp(GSI minus 100)
(9 minus 3D)1113890 1113891 (1)
a [exp(minusGSI15) minus exp(minus203)]
6 + 05 (2)
σcm σcsa (3)
where σc is the uniaxial compressive strength of intact rock sis the empirical parameter reflecting the fracture degree ofthe rock mass ranging from 0 to 1 respectively taking 0 forthe fully fractured rockmass and 1 for the intact rock mass ais the empirical parameter that reflects the features of therock mass GSI is the geological strength criterion of the rockmass D is a parameter reflecting the disturbance degree ofthe field rockmass influencing by external factors and rangesfrom 0 to 1 respectively taking 0 for the undisturbed rockmass and 1 for the completely disturbed rockmass σcm is theuniaxial compressive strength of the rock mass
At measurement point 1 substituting GSI 58 andD 05 into equations (1) and (2) s and a will be 0003698and 05033 respectively Substituting s a and σc intoequation (3) the σcm is 242MPa e calculation results ofthe rock mass strength σcm of 18 measurement points areshown in Table 6
334 Rock Burst Estimation Based on Rock Mass StrengthIn computing the ratio of the rock mass strength σcm to thegeostress Ma et al [21] represented the geostress as thehorizontal stress perpendicular to the tunnel axis In thecurrent study the maximum stress σmax was still used torepresent the geostress considering the availability inpractice and the comparability with the existing specifica-tions (ie the code for hydropower engineering geologicalinvestigation 2016)
e ratios of σcm to σmax at different measurement pointsare listed in Table 7 and shown in Figure 5 e on-site fieldrock burst grades from the actual observations are alsoshown in Table 7
As shown in Table 7 among the 18 randomly chosenpoints the number of measurement points for slight rockburst medium rock burst strong rock burst and violentrock burst is 1 11 6 and 0 respectively Because the pointnumber of violent rock burst is zero it is difficult to ac-curately determine the extent of the ratio for the violent rockburst In order to facilitate the division of the rock burst theratio of σcm to σmax is taken as 0 to be temporarily used toestimate the violent rock burst
From Figure 8 it can be seen that the rock burst is lighterwhen the ratio of σcm to σmax is greater and vice versaerefore three boundary lines may be used to divide theratio of these measurement points into four parts corre-sponding to the grades of the slight medium strong andviolent rock bursts e determination processes of thesethree boundary lines are as follows
90
80
70
60
Figure 7 Schematic diagram of linear interpolation
(a) e determination of the boundary line between theslight and medium rock bursts In this case one and 12measurement points belong respectively to the slightand medium rock bursts Because y3 01850 andy9 01497 represent the minimum and maximumratios of the slight and medium rock bursts respec-tively at the measurement points 3 and 16 the mid-value of y3 and y16 or y 016735 is used as theboundary ratio between these two grades of rock bursts
(b) e determination of the boundary line between themedium and strong rock bursts Because y8 00738and y12 00573 represent the minimum and maxi-mum ratios of the medium and strong rock burstsrespectively at the measurement points 8 and 12 themidvalue of y8 and y12 or y 006555 is used as theboundary ratio between the grades of the rock bursts
(c) e determination of the boundary line between thestrong and violent rock bursts In this case
y6 00236 and y 0 are taken as the minimum andmaximum ratios for the strong and violent rockbursts respectively at the measurement point 6 andothers e midvalue of y6 and 0 or y 001180 isregarded as the boundary ratio between the strongand violent rock bursts
To consider the facility in applications it is better totransfer these threshold values into ones with three digitsafter decimal points Accordingly the ratio intervals aremore than 0167 (0066 0167] (0012 0066] and no morethan 0012 respectively representing the slight mediumstrong and violent rock bursts A rock burst estimationmethod (see Table 8) based on the RMS-to-MG ratio isthereafter obtained
335 Verification of Estimation Criterion for Rock BurstIn order to address the reliability of the abovementionedestimation criterion for rock bursts the remaining 4 field
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
continuously with the influencing distances of more than2merefore the rock burst at this point was determined atthe strong level Based on the actual situations of rock burststhe rock burst grades at 18 measurement points were de-termined by the above method and are shown in Table 2
32RockBurst EstimationResultsUsingTraditionalCriterione CHEGI criterion based on the ratio of σc to σmax is oftenused to estimate the rock burst in practice e corre-sponding criterion is shown in Table 3
In this work the σcm σc and rock burst grades wereobtained by using the hollow inclusion stress relief methodlaboratory uniaxial compressive tests and the ratio of σc toσmax respectively e results are shown in Table 4 Forcomparison purposes the observation results of field rockbursts are also listed in Table 4 From Table 4 it can be seenthat results by using the CHEGI rock burst estimationcriterion are quite different from those at the field situations
33 Modified Criterion of Rock Burst Estimation In thetraditional criteria for rock burst classification the strengthin the strength-stress ratio method generally refers to theuniaxial compressive strength However the actual occur-rence of rock burst in practice relates much with the rockmass structure and the strength of the corresponding rockmass erefore it is necessary to improve the traditionalmethod to pay attention both to the rock mass structurefeatures and to the σcm in the rock burst estimation
e rock mass structure features are generally charac-terized by the geological strength criterion or GSI Based onGSI and other relevant parameters (such as the disturbancecoefficient and uniaxial compressive strength) the gener-alized H-B strength criterion is often used to calculate theσcm and the ratio of σcm to σc may be used as estimationcriterion of the rock burst erefore the rock burst esti-mation based on σcm will be proposed in the following
331 Determination of GSI e geological strength crite-rion or GSI introduced by Marinos and Hoek [23] andconsidering the structural features weathering conditionand the surface features of the rock mass could better reflectthe geological situation of rock mass As shown in Figure 6based on the discontinuity structure and surface conditionof rock mass the average value of GSI may be estimated In
(a) (b) (c)
Figure 3 Stress relief process (a) Core sampling (b) Data collection (c) e core with stress gauge
Figure 4 Core elastic modulus and Poissonrsquos ratio acquisition
Table 1 σmax and σc at different measurement points
this figure ldquoNArdquo means that it is not applicable within thisrange
As shown in Figure 6 the surface quality of rock massstructure may be divided into five categories using theweathering condition of rock mass and the surface featuresof joints which are Very Good Good Fair Poor and VeryPoor Among them Very Goodmeans very rough fresh andunweathered surface of rock mass Good means rough mildweathered iron surface of rock mass Fair means mediumweathered and altered surface of rock mass Poor meanssmooth highly weathered rock mass surface with a denseoverburden or filler or angular fragments Very Poor meanssmooth severely weathered rock mass surface with a softclay coating or filler e corresponding range of values isfrom 100 to 0 in order of high to low and the higher thevalue the better the quality grade of rock mass surface estructural features of the rock mass are divided into sixcategories using the order of the integrity of the rock massstructure surface which are Intact or Massive Blocky VeryBlocky BlockyDisturbedSeamy Disintegrated and Lam-inated or Sheared Wherein the Intact or Massive means acomplete rock mass or a large rock mass structure with fewlarge spacing and discontinuity the Blocky means a goodand original rock mass structure composed of cubic blocksformed by three mutually orthogonal joint faces VeryBlocky means a partially disturbed rock mass structurewhich composed of multifaceted angular blocks formed byat least 4 sets of joints BlockyDisturbedSeamy means a
(a) (b) (c) (d) (e) (f )
(g) (h) (i) (j) (k) (l)
(m) (n) (o) (p) (q) (r)
Figure 5 Appearances after rock bursts at different measurement locations (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4 (e) Point 5 (f )Point 6 (g) Point 7 (h) Point 8 (i) Point 9 (j) Point 10 (k) Point 11 (l) Point 12 (m) Point 13 (n) Point 14 (o) Point 15 (p) Point 16 (q)Point17 (r) Point 18
Table 2 Field rock burst grades at different measurement points
Table 3 Rock burst grades using the ratio of σc to σmax in CHEGIcriterionlowast
IndexGrades of rock burst
Slight Medium Strong Violentσcσmax 4sim7 2sim4 1sim2 lt1lowastere is no groundwater activity in the area
Advances in Civil Engineering 5
rock mass structure in which a plurality of sets of discon-tinuous surfaces are mutually cut to form an angular rockmass and undergoes a fold activity a layer or a flank planecontinuous Disintegrated means a severely fractured rockmass which contains a mixture of angular and circular rockblocks LaminatedSheared means the lack of a massive rockmass structure due to the weak schistosity or the closespacing of the shear planes e corresponding range ofvalues is from 100 to 0 in order of high to low and the higherthe value the better the integrity of the rock mass
e MarinosndashHoek method in estimating GSI was usedin this work e measurement point 1 was taken as an
example to concisely describe the process in estimating GSIIn the measurement point 1 the structural fissures andunloading fissures are well developed the structural fissuresare often distributed in an ldquoXrdquo shape the joints generally donot extend the fissures are 1 to 10mm wide with a maxi-mum width of 15mm the fissures are mostly half-open andfilled visibly the joint spacing is more than 2m Accordinglythe surface condition of the rock mass at this point wasdetermined as ldquoGoodrdquo and the value was estimated as 62 therock mass structure was determined as a blocky structurewith a value of 71
As shown in Figure 7 when the intersection of thevertical line of the rock surface condition and the horizontalline of the rock structure characteristic was between the twoGSI values linear interpolation was often used for calcu-lation Accordingly the GSI of the measurement point 1 wasestimated as 58 According to the above method the GSIvalues of the remaining measurement points were estimatedand the estimated GSIs at 18 measurement points are listedin Table 5
332 Determination of Relevant Parameters Other relevantparameters mainly relate to the rock mass disturbance andσc e related parameter of the rock mass disturbancedegree is coefficient D which represents the disturbance ofthe rock mass Considering the actual situation of the tunnelexcavationDrsquos of all the measurement points are assumed tobe 05 Rock uniaxial compressive strength was determinedthe σc at each measurement point is determined by thelaboratory uniaxial compression test (see early-mentionedTable 1)
333 Calculation of Rock Mass Strength Using HoekndashBrownCriterion Based on Griffithrsquos theory Hoek et al [24] in-vestigated the relations between the ultimate principal stressin a rock mass and the rock mass failure from the statistical
Table 4 Rock burst grades at various measurement locations using the ratio of σc to σmax
No of measurement points σcσmax Estimated results using CHEGI criterion Actual field grades of rock burst
1 082 Violent Strong2 151 Strong Medium3 380 Medium Slight4 087 Violent Strong5 111 Strong Medium6 048 Violent Strong7 083 Violent Strong8 152 Strong Medium9 326 Slight Medium10 188 Strong Medium11 222 Medium Medium12 117 Strong Strong13 253 Medium Medium14 274 Medium Medium15 299 Medium Medium16 143 Strong Medium17 300 Medium Medium18 135 Strong Medium
Rock mass structureRock mass surface conditions
Verygood
Intactmassive
Blocky
Very blocky
Blockydisturbed
seamy
Disintegrated
Laminatedsheared
90
80
70
60
50
40
30
NA NA
Good Fair Poor Verypoor
20
10NA NA
Figure 6 Estimation of geological strength criterion [23]
6 Advances in Civil Engineering
results of the rock triaxial tests and the rock mass tests andobtained the generalized H-B strength guidelines
In the generalized H-B strength criterion the σcm iscalculated using the following formula
s exp(GSI minus 100)
(9 minus 3D)1113890 1113891 (1)
a [exp(minusGSI15) minus exp(minus203)]
6 + 05 (2)
σcm σcsa (3)
where σc is the uniaxial compressive strength of intact rock sis the empirical parameter reflecting the fracture degree ofthe rock mass ranging from 0 to 1 respectively taking 0 forthe fully fractured rockmass and 1 for the intact rock mass ais the empirical parameter that reflects the features of therock mass GSI is the geological strength criterion of the rockmass D is a parameter reflecting the disturbance degree ofthe field rockmass influencing by external factors and rangesfrom 0 to 1 respectively taking 0 for the undisturbed rockmass and 1 for the completely disturbed rockmass σcm is theuniaxial compressive strength of the rock mass
At measurement point 1 substituting GSI 58 andD 05 into equations (1) and (2) s and a will be 0003698and 05033 respectively Substituting s a and σc intoequation (3) the σcm is 242MPa e calculation results ofthe rock mass strength σcm of 18 measurement points areshown in Table 6
334 Rock Burst Estimation Based on Rock Mass StrengthIn computing the ratio of the rock mass strength σcm to thegeostress Ma et al [21] represented the geostress as thehorizontal stress perpendicular to the tunnel axis In thecurrent study the maximum stress σmax was still used torepresent the geostress considering the availability inpractice and the comparability with the existing specifica-tions (ie the code for hydropower engineering geologicalinvestigation 2016)
e ratios of σcm to σmax at different measurement pointsare listed in Table 7 and shown in Figure 5 e on-site fieldrock burst grades from the actual observations are alsoshown in Table 7
As shown in Table 7 among the 18 randomly chosenpoints the number of measurement points for slight rockburst medium rock burst strong rock burst and violentrock burst is 1 11 6 and 0 respectively Because the pointnumber of violent rock burst is zero it is difficult to ac-curately determine the extent of the ratio for the violent rockburst In order to facilitate the division of the rock burst theratio of σcm to σmax is taken as 0 to be temporarily used toestimate the violent rock burst
From Figure 8 it can be seen that the rock burst is lighterwhen the ratio of σcm to σmax is greater and vice versaerefore three boundary lines may be used to divide theratio of these measurement points into four parts corre-sponding to the grades of the slight medium strong andviolent rock bursts e determination processes of thesethree boundary lines are as follows
90
80
70
60
Figure 7 Schematic diagram of linear interpolation
(a) e determination of the boundary line between theslight and medium rock bursts In this case one and 12measurement points belong respectively to the slightand medium rock bursts Because y3 01850 andy9 01497 represent the minimum and maximumratios of the slight and medium rock bursts respec-tively at the measurement points 3 and 16 the mid-value of y3 and y16 or y 016735 is used as theboundary ratio between these two grades of rock bursts
(b) e determination of the boundary line between themedium and strong rock bursts Because y8 00738and y12 00573 represent the minimum and maxi-mum ratios of the medium and strong rock burstsrespectively at the measurement points 8 and 12 themidvalue of y8 and y12 or y 006555 is used as theboundary ratio between the grades of the rock bursts
(c) e determination of the boundary line between thestrong and violent rock bursts In this case
y6 00236 and y 0 are taken as the minimum andmaximum ratios for the strong and violent rockbursts respectively at the measurement point 6 andothers e midvalue of y6 and 0 or y 001180 isregarded as the boundary ratio between the strongand violent rock bursts
To consider the facility in applications it is better totransfer these threshold values into ones with three digitsafter decimal points Accordingly the ratio intervals aremore than 0167 (0066 0167] (0012 0066] and no morethan 0012 respectively representing the slight mediumstrong and violent rock bursts A rock burst estimationmethod (see Table 8) based on the RMS-to-MG ratio isthereafter obtained
335 Verification of Estimation Criterion for Rock BurstIn order to address the reliability of the abovementionedestimation criterion for rock bursts the remaining 4 field
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
this figure ldquoNArdquo means that it is not applicable within thisrange
As shown in Figure 6 the surface quality of rock massstructure may be divided into five categories using theweathering condition of rock mass and the surface featuresof joints which are Very Good Good Fair Poor and VeryPoor Among them Very Goodmeans very rough fresh andunweathered surface of rock mass Good means rough mildweathered iron surface of rock mass Fair means mediumweathered and altered surface of rock mass Poor meanssmooth highly weathered rock mass surface with a denseoverburden or filler or angular fragments Very Poor meanssmooth severely weathered rock mass surface with a softclay coating or filler e corresponding range of values isfrom 100 to 0 in order of high to low and the higher thevalue the better the quality grade of rock mass surface estructural features of the rock mass are divided into sixcategories using the order of the integrity of the rock massstructure surface which are Intact or Massive Blocky VeryBlocky BlockyDisturbedSeamy Disintegrated and Lam-inated or Sheared Wherein the Intact or Massive means acomplete rock mass or a large rock mass structure with fewlarge spacing and discontinuity the Blocky means a goodand original rock mass structure composed of cubic blocksformed by three mutually orthogonal joint faces VeryBlocky means a partially disturbed rock mass structurewhich composed of multifaceted angular blocks formed byat least 4 sets of joints BlockyDisturbedSeamy means a
(a) (b) (c) (d) (e) (f )
(g) (h) (i) (j) (k) (l)
(m) (n) (o) (p) (q) (r)
Figure 5 Appearances after rock bursts at different measurement locations (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4 (e) Point 5 (f )Point 6 (g) Point 7 (h) Point 8 (i) Point 9 (j) Point 10 (k) Point 11 (l) Point 12 (m) Point 13 (n) Point 14 (o) Point 15 (p) Point 16 (q)Point17 (r) Point 18
Table 2 Field rock burst grades at different measurement points
Table 3 Rock burst grades using the ratio of σc to σmax in CHEGIcriterionlowast
IndexGrades of rock burst
Slight Medium Strong Violentσcσmax 4sim7 2sim4 1sim2 lt1lowastere is no groundwater activity in the area
Advances in Civil Engineering 5
rock mass structure in which a plurality of sets of discon-tinuous surfaces are mutually cut to form an angular rockmass and undergoes a fold activity a layer or a flank planecontinuous Disintegrated means a severely fractured rockmass which contains a mixture of angular and circular rockblocks LaminatedSheared means the lack of a massive rockmass structure due to the weak schistosity or the closespacing of the shear planes e corresponding range ofvalues is from 100 to 0 in order of high to low and the higherthe value the better the integrity of the rock mass
e MarinosndashHoek method in estimating GSI was usedin this work e measurement point 1 was taken as an
example to concisely describe the process in estimating GSIIn the measurement point 1 the structural fissures andunloading fissures are well developed the structural fissuresare often distributed in an ldquoXrdquo shape the joints generally donot extend the fissures are 1 to 10mm wide with a maxi-mum width of 15mm the fissures are mostly half-open andfilled visibly the joint spacing is more than 2m Accordinglythe surface condition of the rock mass at this point wasdetermined as ldquoGoodrdquo and the value was estimated as 62 therock mass structure was determined as a blocky structurewith a value of 71
As shown in Figure 7 when the intersection of thevertical line of the rock surface condition and the horizontalline of the rock structure characteristic was between the twoGSI values linear interpolation was often used for calcu-lation Accordingly the GSI of the measurement point 1 wasestimated as 58 According to the above method the GSIvalues of the remaining measurement points were estimatedand the estimated GSIs at 18 measurement points are listedin Table 5
332 Determination of Relevant Parameters Other relevantparameters mainly relate to the rock mass disturbance andσc e related parameter of the rock mass disturbancedegree is coefficient D which represents the disturbance ofthe rock mass Considering the actual situation of the tunnelexcavationDrsquos of all the measurement points are assumed tobe 05 Rock uniaxial compressive strength was determinedthe σc at each measurement point is determined by thelaboratory uniaxial compression test (see early-mentionedTable 1)
333 Calculation of Rock Mass Strength Using HoekndashBrownCriterion Based on Griffithrsquos theory Hoek et al [24] in-vestigated the relations between the ultimate principal stressin a rock mass and the rock mass failure from the statistical
Table 4 Rock burst grades at various measurement locations using the ratio of σc to σmax
No of measurement points σcσmax Estimated results using CHEGI criterion Actual field grades of rock burst
1 082 Violent Strong2 151 Strong Medium3 380 Medium Slight4 087 Violent Strong5 111 Strong Medium6 048 Violent Strong7 083 Violent Strong8 152 Strong Medium9 326 Slight Medium10 188 Strong Medium11 222 Medium Medium12 117 Strong Strong13 253 Medium Medium14 274 Medium Medium15 299 Medium Medium16 143 Strong Medium17 300 Medium Medium18 135 Strong Medium
Rock mass structureRock mass surface conditions
Verygood
Intactmassive
Blocky
Very blocky
Blockydisturbed
seamy
Disintegrated
Laminatedsheared
90
80
70
60
50
40
30
NA NA
Good Fair Poor Verypoor
20
10NA NA
Figure 6 Estimation of geological strength criterion [23]
6 Advances in Civil Engineering
results of the rock triaxial tests and the rock mass tests andobtained the generalized H-B strength guidelines
In the generalized H-B strength criterion the σcm iscalculated using the following formula
s exp(GSI minus 100)
(9 minus 3D)1113890 1113891 (1)
a [exp(minusGSI15) minus exp(minus203)]
6 + 05 (2)
σcm σcsa (3)
where σc is the uniaxial compressive strength of intact rock sis the empirical parameter reflecting the fracture degree ofthe rock mass ranging from 0 to 1 respectively taking 0 forthe fully fractured rockmass and 1 for the intact rock mass ais the empirical parameter that reflects the features of therock mass GSI is the geological strength criterion of the rockmass D is a parameter reflecting the disturbance degree ofthe field rockmass influencing by external factors and rangesfrom 0 to 1 respectively taking 0 for the undisturbed rockmass and 1 for the completely disturbed rockmass σcm is theuniaxial compressive strength of the rock mass
At measurement point 1 substituting GSI 58 andD 05 into equations (1) and (2) s and a will be 0003698and 05033 respectively Substituting s a and σc intoequation (3) the σcm is 242MPa e calculation results ofthe rock mass strength σcm of 18 measurement points areshown in Table 6
334 Rock Burst Estimation Based on Rock Mass StrengthIn computing the ratio of the rock mass strength σcm to thegeostress Ma et al [21] represented the geostress as thehorizontal stress perpendicular to the tunnel axis In thecurrent study the maximum stress σmax was still used torepresent the geostress considering the availability inpractice and the comparability with the existing specifica-tions (ie the code for hydropower engineering geologicalinvestigation 2016)
e ratios of σcm to σmax at different measurement pointsare listed in Table 7 and shown in Figure 5 e on-site fieldrock burst grades from the actual observations are alsoshown in Table 7
As shown in Table 7 among the 18 randomly chosenpoints the number of measurement points for slight rockburst medium rock burst strong rock burst and violentrock burst is 1 11 6 and 0 respectively Because the pointnumber of violent rock burst is zero it is difficult to ac-curately determine the extent of the ratio for the violent rockburst In order to facilitate the division of the rock burst theratio of σcm to σmax is taken as 0 to be temporarily used toestimate the violent rock burst
From Figure 8 it can be seen that the rock burst is lighterwhen the ratio of σcm to σmax is greater and vice versaerefore three boundary lines may be used to divide theratio of these measurement points into four parts corre-sponding to the grades of the slight medium strong andviolent rock bursts e determination processes of thesethree boundary lines are as follows
90
80
70
60
Figure 7 Schematic diagram of linear interpolation
(a) e determination of the boundary line between theslight and medium rock bursts In this case one and 12measurement points belong respectively to the slightand medium rock bursts Because y3 01850 andy9 01497 represent the minimum and maximumratios of the slight and medium rock bursts respec-tively at the measurement points 3 and 16 the mid-value of y3 and y16 or y 016735 is used as theboundary ratio between these two grades of rock bursts
(b) e determination of the boundary line between themedium and strong rock bursts Because y8 00738and y12 00573 represent the minimum and maxi-mum ratios of the medium and strong rock burstsrespectively at the measurement points 8 and 12 themidvalue of y8 and y12 or y 006555 is used as theboundary ratio between the grades of the rock bursts
(c) e determination of the boundary line between thestrong and violent rock bursts In this case
y6 00236 and y 0 are taken as the minimum andmaximum ratios for the strong and violent rockbursts respectively at the measurement point 6 andothers e midvalue of y6 and 0 or y 001180 isregarded as the boundary ratio between the strongand violent rock bursts
To consider the facility in applications it is better totransfer these threshold values into ones with three digitsafter decimal points Accordingly the ratio intervals aremore than 0167 (0066 0167] (0012 0066] and no morethan 0012 respectively representing the slight mediumstrong and violent rock bursts A rock burst estimationmethod (see Table 8) based on the RMS-to-MG ratio isthereafter obtained
335 Verification of Estimation Criterion for Rock BurstIn order to address the reliability of the abovementionedestimation criterion for rock bursts the remaining 4 field
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
rock mass structure in which a plurality of sets of discon-tinuous surfaces are mutually cut to form an angular rockmass and undergoes a fold activity a layer or a flank planecontinuous Disintegrated means a severely fractured rockmass which contains a mixture of angular and circular rockblocks LaminatedSheared means the lack of a massive rockmass structure due to the weak schistosity or the closespacing of the shear planes e corresponding range ofvalues is from 100 to 0 in order of high to low and the higherthe value the better the integrity of the rock mass
e MarinosndashHoek method in estimating GSI was usedin this work e measurement point 1 was taken as an
example to concisely describe the process in estimating GSIIn the measurement point 1 the structural fissures andunloading fissures are well developed the structural fissuresare often distributed in an ldquoXrdquo shape the joints generally donot extend the fissures are 1 to 10mm wide with a maxi-mum width of 15mm the fissures are mostly half-open andfilled visibly the joint spacing is more than 2m Accordinglythe surface condition of the rock mass at this point wasdetermined as ldquoGoodrdquo and the value was estimated as 62 therock mass structure was determined as a blocky structurewith a value of 71
As shown in Figure 7 when the intersection of thevertical line of the rock surface condition and the horizontalline of the rock structure characteristic was between the twoGSI values linear interpolation was often used for calcu-lation Accordingly the GSI of the measurement point 1 wasestimated as 58 According to the above method the GSIvalues of the remaining measurement points were estimatedand the estimated GSIs at 18 measurement points are listedin Table 5
332 Determination of Relevant Parameters Other relevantparameters mainly relate to the rock mass disturbance andσc e related parameter of the rock mass disturbancedegree is coefficient D which represents the disturbance ofthe rock mass Considering the actual situation of the tunnelexcavationDrsquos of all the measurement points are assumed tobe 05 Rock uniaxial compressive strength was determinedthe σc at each measurement point is determined by thelaboratory uniaxial compression test (see early-mentionedTable 1)
333 Calculation of Rock Mass Strength Using HoekndashBrownCriterion Based on Griffithrsquos theory Hoek et al [24] in-vestigated the relations between the ultimate principal stressin a rock mass and the rock mass failure from the statistical
Table 4 Rock burst grades at various measurement locations using the ratio of σc to σmax
No of measurement points σcσmax Estimated results using CHEGI criterion Actual field grades of rock burst
1 082 Violent Strong2 151 Strong Medium3 380 Medium Slight4 087 Violent Strong5 111 Strong Medium6 048 Violent Strong7 083 Violent Strong8 152 Strong Medium9 326 Slight Medium10 188 Strong Medium11 222 Medium Medium12 117 Strong Strong13 253 Medium Medium14 274 Medium Medium15 299 Medium Medium16 143 Strong Medium17 300 Medium Medium18 135 Strong Medium
Rock mass structureRock mass surface conditions
Verygood
Intactmassive
Blocky
Very blocky
Blockydisturbed
seamy
Disintegrated
Laminatedsheared
90
80
70
60
50
40
30
NA NA
Good Fair Poor Verypoor
20
10NA NA
Figure 6 Estimation of geological strength criterion [23]
6 Advances in Civil Engineering
results of the rock triaxial tests and the rock mass tests andobtained the generalized H-B strength guidelines
In the generalized H-B strength criterion the σcm iscalculated using the following formula
s exp(GSI minus 100)
(9 minus 3D)1113890 1113891 (1)
a [exp(minusGSI15) minus exp(minus203)]
6 + 05 (2)
σcm σcsa (3)
where σc is the uniaxial compressive strength of intact rock sis the empirical parameter reflecting the fracture degree ofthe rock mass ranging from 0 to 1 respectively taking 0 forthe fully fractured rockmass and 1 for the intact rock mass ais the empirical parameter that reflects the features of therock mass GSI is the geological strength criterion of the rockmass D is a parameter reflecting the disturbance degree ofthe field rockmass influencing by external factors and rangesfrom 0 to 1 respectively taking 0 for the undisturbed rockmass and 1 for the completely disturbed rockmass σcm is theuniaxial compressive strength of the rock mass
At measurement point 1 substituting GSI 58 andD 05 into equations (1) and (2) s and a will be 0003698and 05033 respectively Substituting s a and σc intoequation (3) the σcm is 242MPa e calculation results ofthe rock mass strength σcm of 18 measurement points areshown in Table 6
334 Rock Burst Estimation Based on Rock Mass StrengthIn computing the ratio of the rock mass strength σcm to thegeostress Ma et al [21] represented the geostress as thehorizontal stress perpendicular to the tunnel axis In thecurrent study the maximum stress σmax was still used torepresent the geostress considering the availability inpractice and the comparability with the existing specifica-tions (ie the code for hydropower engineering geologicalinvestigation 2016)
e ratios of σcm to σmax at different measurement pointsare listed in Table 7 and shown in Figure 5 e on-site fieldrock burst grades from the actual observations are alsoshown in Table 7
As shown in Table 7 among the 18 randomly chosenpoints the number of measurement points for slight rockburst medium rock burst strong rock burst and violentrock burst is 1 11 6 and 0 respectively Because the pointnumber of violent rock burst is zero it is difficult to ac-curately determine the extent of the ratio for the violent rockburst In order to facilitate the division of the rock burst theratio of σcm to σmax is taken as 0 to be temporarily used toestimate the violent rock burst
From Figure 8 it can be seen that the rock burst is lighterwhen the ratio of σcm to σmax is greater and vice versaerefore three boundary lines may be used to divide theratio of these measurement points into four parts corre-sponding to the grades of the slight medium strong andviolent rock bursts e determination processes of thesethree boundary lines are as follows
90
80
70
60
Figure 7 Schematic diagram of linear interpolation
(a) e determination of the boundary line between theslight and medium rock bursts In this case one and 12measurement points belong respectively to the slightand medium rock bursts Because y3 01850 andy9 01497 represent the minimum and maximumratios of the slight and medium rock bursts respec-tively at the measurement points 3 and 16 the mid-value of y3 and y16 or y 016735 is used as theboundary ratio between these two grades of rock bursts
(b) e determination of the boundary line between themedium and strong rock bursts Because y8 00738and y12 00573 represent the minimum and maxi-mum ratios of the medium and strong rock burstsrespectively at the measurement points 8 and 12 themidvalue of y8 and y12 or y 006555 is used as theboundary ratio between the grades of the rock bursts
(c) e determination of the boundary line between thestrong and violent rock bursts In this case
y6 00236 and y 0 are taken as the minimum andmaximum ratios for the strong and violent rockbursts respectively at the measurement point 6 andothers e midvalue of y6 and 0 or y 001180 isregarded as the boundary ratio between the strongand violent rock bursts
To consider the facility in applications it is better totransfer these threshold values into ones with three digitsafter decimal points Accordingly the ratio intervals aremore than 0167 (0066 0167] (0012 0066] and no morethan 0012 respectively representing the slight mediumstrong and violent rock bursts A rock burst estimationmethod (see Table 8) based on the RMS-to-MG ratio isthereafter obtained
335 Verification of Estimation Criterion for Rock BurstIn order to address the reliability of the abovementionedestimation criterion for rock bursts the remaining 4 field
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
results of the rock triaxial tests and the rock mass tests andobtained the generalized H-B strength guidelines
In the generalized H-B strength criterion the σcm iscalculated using the following formula
s exp(GSI minus 100)
(9 minus 3D)1113890 1113891 (1)
a [exp(minusGSI15) minus exp(minus203)]
6 + 05 (2)
σcm σcsa (3)
where σc is the uniaxial compressive strength of intact rock sis the empirical parameter reflecting the fracture degree ofthe rock mass ranging from 0 to 1 respectively taking 0 forthe fully fractured rockmass and 1 for the intact rock mass ais the empirical parameter that reflects the features of therock mass GSI is the geological strength criterion of the rockmass D is a parameter reflecting the disturbance degree ofthe field rockmass influencing by external factors and rangesfrom 0 to 1 respectively taking 0 for the undisturbed rockmass and 1 for the completely disturbed rockmass σcm is theuniaxial compressive strength of the rock mass
At measurement point 1 substituting GSI 58 andD 05 into equations (1) and (2) s and a will be 0003698and 05033 respectively Substituting s a and σc intoequation (3) the σcm is 242MPa e calculation results ofthe rock mass strength σcm of 18 measurement points areshown in Table 6
334 Rock Burst Estimation Based on Rock Mass StrengthIn computing the ratio of the rock mass strength σcm to thegeostress Ma et al [21] represented the geostress as thehorizontal stress perpendicular to the tunnel axis In thecurrent study the maximum stress σmax was still used torepresent the geostress considering the availability inpractice and the comparability with the existing specifica-tions (ie the code for hydropower engineering geologicalinvestigation 2016)
e ratios of σcm to σmax at different measurement pointsare listed in Table 7 and shown in Figure 5 e on-site fieldrock burst grades from the actual observations are alsoshown in Table 7
As shown in Table 7 among the 18 randomly chosenpoints the number of measurement points for slight rockburst medium rock burst strong rock burst and violentrock burst is 1 11 6 and 0 respectively Because the pointnumber of violent rock burst is zero it is difficult to ac-curately determine the extent of the ratio for the violent rockburst In order to facilitate the division of the rock burst theratio of σcm to σmax is taken as 0 to be temporarily used toestimate the violent rock burst
From Figure 8 it can be seen that the rock burst is lighterwhen the ratio of σcm to σmax is greater and vice versaerefore three boundary lines may be used to divide theratio of these measurement points into four parts corre-sponding to the grades of the slight medium strong andviolent rock bursts e determination processes of thesethree boundary lines are as follows
90
80
70
60
Figure 7 Schematic diagram of linear interpolation
(a) e determination of the boundary line between theslight and medium rock bursts In this case one and 12measurement points belong respectively to the slightand medium rock bursts Because y3 01850 andy9 01497 represent the minimum and maximumratios of the slight and medium rock bursts respec-tively at the measurement points 3 and 16 the mid-value of y3 and y16 or y 016735 is used as theboundary ratio between these two grades of rock bursts
(b) e determination of the boundary line between themedium and strong rock bursts Because y8 00738and y12 00573 represent the minimum and maxi-mum ratios of the medium and strong rock burstsrespectively at the measurement points 8 and 12 themidvalue of y8 and y12 or y 006555 is used as theboundary ratio between the grades of the rock bursts
(c) e determination of the boundary line between thestrong and violent rock bursts In this case
y6 00236 and y 0 are taken as the minimum andmaximum ratios for the strong and violent rockbursts respectively at the measurement point 6 andothers e midvalue of y6 and 0 or y 001180 isregarded as the boundary ratio between the strongand violent rock bursts
To consider the facility in applications it is better totransfer these threshold values into ones with three digitsafter decimal points Accordingly the ratio intervals aremore than 0167 (0066 0167] (0012 0066] and no morethan 0012 respectively representing the slight mediumstrong and violent rock bursts A rock burst estimationmethod (see Table 8) based on the RMS-to-MG ratio isthereafter obtained
335 Verification of Estimation Criterion for Rock BurstIn order to address the reliability of the abovementionedestimation criterion for rock bursts the remaining 4 field
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
(a) e determination of the boundary line between theslight and medium rock bursts In this case one and 12measurement points belong respectively to the slightand medium rock bursts Because y3 01850 andy9 01497 represent the minimum and maximumratios of the slight and medium rock bursts respec-tively at the measurement points 3 and 16 the mid-value of y3 and y16 or y 016735 is used as theboundary ratio between these two grades of rock bursts
(b) e determination of the boundary line between themedium and strong rock bursts Because y8 00738and y12 00573 represent the minimum and maxi-mum ratios of the medium and strong rock burstsrespectively at the measurement points 8 and 12 themidvalue of y8 and y12 or y 006555 is used as theboundary ratio between the grades of the rock bursts
(c) e determination of the boundary line between thestrong and violent rock bursts In this case
y6 00236 and y 0 are taken as the minimum andmaximum ratios for the strong and violent rockbursts respectively at the measurement point 6 andothers e midvalue of y6 and 0 or y 001180 isregarded as the boundary ratio between the strongand violent rock bursts
To consider the facility in applications it is better totransfer these threshold values into ones with three digitsafter decimal points Accordingly the ratio intervals aremore than 0167 (0066 0167] (0012 0066] and no morethan 0012 respectively representing the slight mediumstrong and violent rock bursts A rock burst estimationmethod (see Table 8) based on the RMS-to-MG ratio isthereafter obtained
335 Verification of Estimation Criterion for Rock BurstIn order to address the reliability of the abovementionedestimation criterion for rock bursts the remaining 4 field
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
measurement points are used for verification e appear-ances after the rock bursts at these points are shown inFigure 9
Using the traditional method based on the ratio of σc toσmax the estimation results at these four verification pointsare obtained (see Table 9)
Using the modified criterion the ratios of σcm to σmax atthese four verification points were obtained (see Table 10)
Comparing the index σcσmax in Table 9 and the indexσcmσmax in Table 10 with the rock burst grade estimationcriteria in Table 3 and Table 8 respectively the rock burstestimation results of the CHEGIrsquos criterion and modifiedcriterion at four verification points were obtained (see Ta-ble 11) Table 11 also lists the field observations
As shown in Table 11 the estimation results using theCHEGI criterion based on the ratio of σc to σmax are quitedifferent from those in the actual situations if the criterionbased on the ratio of σcm to σmax is used the estimationresults for the rock bursts will be consistent with the fieldobservations Accordingly the rock burst estimation onlyconsidering the rock strength regardless of the rock massstructure is not suitable In the modified criterion muchattention is paid on the rockmass structure in computing theratio of σcm to σmax e estimation results based on themodified criterion are consistent with the actual situationand may be used for rock burst estimation
34 4e Influence of Randomness of Data Selection on RockBurst Estimation Criterion Using the results of 22 field
measurement points 18 of them were randomly selected tomatch the actual situation of rock bursts and a new rockburst criterion was proposed In order to investigate theinfluence of the randomness of data selection on the esti-mation criterion of rock burst 12 sets of data each of whichincludes 18 measurement points were randomly selectedfrom 22 field measurement points Using the above methodthe corresponding rock burst estimation criteria were ob-tained and compared e results are shown in Table 12
It can be seen from Table 12 that among the estimationcriteria of slight rock burst grade in 12 sets 11 of them aremore than 0167 and the other is greater than 0156 withlittle change in range the estimation criterion of mediumrock burst grade is 0066 to 0167 in 8 sets 0062 to 0167 in 3sets and 0066 to 0156 in 1 set the change of estimationcriterion of strong rock burst grade is similar to that ofmedium rock burst grade with 8 sets ranging from 0012 to0066 3 sets ranging from 0012 to 0062 and 1 set rangingfrom 0019 to 0066 the estimation criterion of violent rockburst grade is nomore than 0012 in 9 sets and less than 0019
Table 8 Rock burst estimation criterion based on the ratio of σcm to σc
Estimation index Slight rock burst Medium rock burst Strong rock burst Violent rock burstσcmσmax gt0167 0066sim0167 0012sim0066 le0012
(a) (b) (c) (d)
Figure 9 Appearances after rock bursts at 4 verification measurement points (a) Point 1 (b) Point 2 (c) Point 3 (d) Point 4
Table 9 Calculation of rock burst grades at 4 verification points using the ratio of σc to σmax
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
in 3 sets the range of change is a little big but the probabilityof occurrence of a situation less than 0019 is small Ac-cordingly the criteria of slight medium strong and violentrock bursts were determined to be greater than 0167 0066to 0165 0012 to 0066 and less than 0012 respectivelywhich have a good stability
Comparing the data in Tables 12 and 8 it is found thatthe rock burst estimation criterion obtained in this work hasa relatively high stability It can be seen from Tables 7 10and 11 that there is only one slight rock burst measurementpoint in the 22 field measurement points If the slight rockburst measurement point was not selected the maximumratio of RMS to MG in medium rock burst measurementpoints will have to be taken as the boundary value betweenslight and medium rock bursts which will have a certainimpact on these two rock burst estimation criteria Howeverthe probability of not selecting this slight measurement pointfrom 22 points is 1818 so the probability value is relativelysmall Moreover measurement points of the medium rockburst were relatively large which will weaken this impact tosome extent In the later studies the accuracy of theboundary value between the slight and medium rock burstsmay be modified by increasing the number of slight rockburst measurement points
4 Discussions
(1) Considering the rock mass structure could changethe evolution mode of rock burst activity and therock mass (GSI) degradation could be used foravoiding the risk of rock burst [25 26] the effect of
GSI on the estimation of rock burst grades wasexplored e verification point 3 was taken as anexample (see Figure 10) where various GSI values(50 51 52 53 54 55 56 57 58 59 and 60) wereselectedFrom Figure 10 it can be seen that the estimationindex of rock burst grade increases approximatelylinearly with the increase in GSI and R2 (goodness offit) is 09916 the effect of GSI on the estimation ofrock burst grades is relatively great and σcmσmaxincreases by 71 when GSI increases by 1 the rockburst grade of verification point 3 is strong if GSI isbetween 50 and 53 while the rock burst grade ismedium if GSI is between 54 and 60 erefore theeffect of GSI on the rock burst grade needs to be paidmore attention
(2) Considering the HoekndashBrown criterion may over-estimate the strength of rock mass [27 28] variousmethods for estimating rock mass strength wereconducted Taking the measurement point 1 as anexample the values of c (the weight of rock) Q(rock mass quality rating) and RMR (rock massrating) are 26 kNm3 012 and 25 respectively eestimation results of rock mass strength are shownin Table 13From Table 13 it can be seen that the rock massstrength is mainly distributed in the range of 24 to90MPa and 5 to 20 times less than the rock massstrength the estimations using Hoekrsquos and Kala-marasrsquos methods are relatively close while those
Table 12 Estimation criteria of rock burst grades under different conditions
Table 11 Comparison of rock burst estimation results between the CHEGI and modified criterion
No of verification pointsEstimated results of the
modified criterion Determination results of on-site rock burstEstimated results of the
CHEGI criterionσcmσmax Rock burst grade σcmσmax Rock burst grade
1 00423 Strong Strong 0813 Violent2 01105 Medium Medium 1983 Strong3 00836 Medium Medium 1501 Strong4 00478 Strong Strong 0919 Violent
10 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
using Bartonrsquos and Singhrsquos methods seem to be toohigh Hence the rock mass strength estimated byHoekrsquos method is relatively reasonable
(3) In order to explore the reliability of the modifiedmethod in this paper the rock burst grade estimationresults were compared using various criteria (seeTable 14) As can be seen from Table 14 the clas-sification of Bartonrsquos criterion which simply clas-sified the rock burst into the mild and heavy rockburst grades was not accurate enough even thoughRehman et al [32] considered the rock is jointed andproposed an index of SRFQ the classification of rockburst was also inaccurate based on Bartonrsquos crite-rion Tao [3] divided the mild rock burst into themedium and slight ones while there is no subdivi-sions in heavy rock bursts the rock burst grades weredivided into four grades (slight medium strong andviolent) by the CHEGIrsquos criterion but this criterionignored the effect of the rock mass structure on therock burst In the current study the rock burst gradesare divided into four grades using the modifiedcriterion where the rock mass structure was wellconsidered
e estimation results at four verification points areshown in Table 15 As can be seen from Table 15 therock burst grades of verification points 1 and 4 es-timated by various criteria are relatively consistentwhile the rock burst grades at verification points 2and 3 estimated by Bartonrsquos Rehmanrsquos and Taorsquoscriteria are all greater than those estimated by themodified criterion for these four verification points
the rock burst grades estimated by the CHEGIrsquoscriterion are greater (in an one grade) than thoseestimated by the modified criterionEstimated results using various criteria with theactual rock burst were compared and are shown inTable 16From Table 16 it can be seen that the estimationresults using Bartonrsquos Rehmanrsquos and Taorsquos criteriaare overestimated and have a lower accuracycompared with the actual rock burst results theestimated ones using the CHEGIrsquos criterion are alsooverestimated the estimated rock burst grades usingthe modified criterion in the current study are ingood agreement with the actual ones
(4) When determining the boundary value of adjacentrock burst grades it is not enough to consider onlythe maximum or minimum value of the rock burstgrades as the boundary value and the reliability ofthe rock burst estimation index obtained by thismethod is greatly influenced by the factors of arti-ficial selection of data In this work the method oftaking the midvalue may well consider the data oftwo adjacent rock bursts which makes the estima-tion criterion of rock burst closer to the real valueand has better generalization ability In the futurethe boundary value of rock burst grades may befurther optimized by increasing the number anddiversity of rock burst data
(5) Taking the Neelum-Jhelum Hydroelectric Project inPakistan as an example ([12] and [33]) the esti-mation results using various criteria were compared
Table 13 Estimation results of rock mass strength
Hoek et al [24] Barton [29] Singh [30] Kalamaras and Bieniawski [31]Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa) Formula σcm (MPa)σcmσc sa 242 σcm 05c (Qσc100)13 475 σcm 07cQ13 898 σcmσc (RMR minus 15)170 239
y = 00051x + 00441R2 = 09916
000
002
004
006
008
010
012
50 51 52 53 54 55 56 57 58 59 60GSI
σ cm
σm
ax
y = 0066
Strong
Medium
Figure 10 e relationship between GSI and the ratio of σcm to σmax
Advances in Civil Engineering 11
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
to further analyze the reliability of the modifiedcriterion e results are shown in Table 17
It can be seen from Table 17 that the rock burst esti-mation grades using traditional criteria are quite differentfrom the actual ones the estimation results using themodified criterion in this study are much closer to the actualrock burst the estimation results using the modified
criterion have a good reliability for the measurement pointswith GSI greater than 60
5 Conclusions
In this work after using the generalized H-B strength cri-terion to calculate the σcm a modified rock burst estimation
Table 16 Comparison between actual rock burst grades and estimated results
lt25 10sim20 Heavy 10sim20 Heavy 1sim2 Strong 0012sim0066 Stronglt25 Heavy lt1 Violent le0012 Violent
lowastSRF and SRFQ are stress reduction factor and modified stress reduction factor respectively
Table 15 Estimation results of various criteria
No of verificationpoints σcσmax SRF SRFQ σcσmax
Bartonrsquoscriterion
Rehmanrsquoscriterion
Taorsquoscriterion
CHEGIrsquoscriterion
Modifiedcriterion
1 0813 1374 1457 00423 Heavy Heavy Heavy Violent Strong2 1983 1035 1017 01105 Heavy Heavy Heavy Strong Medium3 1501 1134 1180 00836 Heavy Heavy Heavy Strong Medium4 0919 1893 1255 00478 Heavy Heavy Heavy Violent Strong
12 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
criterion based on the ratio of σcm to σmax was proposedeestimation results of this method were compared with thoseof the traditional method based on the ratio of σc to σmax Itshows that
(1) e generalized H-B criterion in calculating σcm mayreflect both the structural features of the rock massesand actual situations of the excavation disturbances
(2) e estimation results by using the ratio of σc to σmaxin the traditional method are quite different from theactual situation while the estimation results by usingthe ratio of σcm to σmax in this work are consistentwith the actual situation
(3) If a ratio of σcm to σc is used for the rock burstestimation the ratio intervals in the slight mediumstrong and violent rock burst grades are more than0167 (0066 0167] (0012 0066] and no morethan 0012 respectively
(4) e randomness of the selection of rock burstmeasurement points has a certain influence on thedetermination of the estimation criterion of rockburst grades but the overall change of the range isnot large e rock burst estimation criterion pro-posed in this work has a good reliability
Data Availability
e data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
e authors declare that they have no conflicts of interest
Acknowledgments
Financial supports for the study were provided by theNatural Sciences Foundation Committee of China underGrant no 41472254 and the Science and Technology Re-search and Development Program of China Railway Con-struction Corporation Limited under Grant no 17-C13ese supports are gratefully acknowledged
References
[1] A C Adoko C Gokceoglu L Wu and Q J ZuoldquoKnowledge-based and data-driven fuzzy modeling forrockburst predictionrdquo International Journal of Rock Me-chanics and Mining Sciences vol 61 pp 86ndash95 2013
[2] A C Adoko and T Zvarivadza ldquoA bayesian approach forpredicting rockburstrdquo in Proccedings of the 52nd US RockMechanicsGeomechanics Symposium Alexandria VA USAJune 2018
[3] Z Y Tao ldquoRockburst and its evaluation method in highground stress fieldrdquo Yangtze River vol 18 no 5 pp 25ndash321987
[4] M C Gu F L He and C Z Chen ldquoStudy on rock burst inqinling tunnelrdquo Chinese Journal of Rock Mechanics and En-gineering vol 21 no 9 pp 1324ndash1329 2002
[5] J J Zhang and B J Fu ldquoRock burst and its criteria andcontrolrdquo Chinese Journal of Rock Mechanics and Engineeringvol 27 no 10 pp 2034ndash2042 2008
[6] L Liu Z Q Chen and L G Wang ldquoRock burst laws in deepmines based on combinedmodel of membership function anddominance-based rough setrdquo Journal of Central South Uni-versity vol 22 no 9 pp 3591ndash3597 2015
[7] S-J Miao M-F Cai Q-F Guo and Z-J Huang ldquoRock burstprediction based on in-situ stress and energy accumulationtheoryrdquo International Journal of Rock Mechanics and MiningSciences vol 83 pp 86ndash94 2016
[8] K P Zhou Y Lin H W Deng J L Li and C L LiuldquoPrediction of rock burst classification using cloudmodel withentropy weightrdquo Transactions of Nonferrous Metals Society ofChina vol 26 no 7 pp 1995ndash2002 2016
[9] T Z Li Y X Li and X L Yang ldquoRock burst prediction basedon genetic algorithms and extreme learningmachinerdquo Journalof Central South University vol 24 no 9 pp 2105ndash2113 2017
[10] S Afraei K Shahriar and S H Madani ldquoStatistical assess-ment of rock burst potential and contributions of consideredpredictor variables in the taskrdquo Tunnelling and UndergroundSpace Technology vol 72 pp 250ndash271 2018
[11] Y G Xue Z Q Li S C Li D H Qiu Y F Tao and L WangldquoPrediction of rock burst in underground caverns based onrough set and extensible comprehensive evaluationrdquo Bulletinof Engineering Geology and the Environment vol 78 no 1pp 417ndash429 2019
[12] W Z Chen C S Ma H M Tian and J P Yang ldquoDiscussionon rockburst predictive method applying to TBM tunnelconstructionrdquo Rock and Soil Mechanics vol 38 no 2pp 241ndash249 2017
[13] G L Feng X T Feng B R Chen Y X Xiao and Z N ZhaoldquoEffects of structural planes on the microseismicity associatedwith rockburst development processes in deep tunnels of theJinping-II hydropower station Chinardquo Tunnelling and Un-derground Space Technology vol 84 pp 273ndash280 2019
[14] H Zhou F Z Meng C Q Zhang D W Hu F G Yang andJ J Lu ldquoAnalysis of rockburst mechanisms induced bystructural planes in deep tunnelsrdquo Bulletin of EngineeringGeology and the Environment vol 74 pp 1435ndash1451 2019
[15] E T Mohamad C S Yi B R Murlidhar and R Saad ldquoEffectof geological structure on flyrock prediction in constructionblastingrdquo Geological and Geotechnical Engineering vol 36no 4 pp 2217ndash2235 2018
[16] Y Du Y T Zheng M W Xie Y J Jiang and Q Q LiuldquoStrength weakening characteristic of rock burst structuralplanesrdquo Chinese Journal of Engineering vol 40 no 3pp 269ndash275 2018
[17] E Hoek and E T Brown ldquoEmpirical strength criterion forrock massesrdquo Journal of Geotechnical and GeoenvironmentalEngineering vol 106 no 9 pp 1013ndash1035 1980
[18] E Hoek and E T Brown ldquoPractical estimates of rock massstrengthrdquo International Journal of Rock Mechanics andMining Sciences vol 34 no 8 pp 1165ndash1186 1997
[19] M Sharifzadeh M Sharifi and S M Delbari ldquoBack analysisof an excavated slope failure in highly fractured rock mass thecase study of kargar slope failure (Iran)rdquo Environmental EarthSciences vol 60 no 1 pp 183ndash192 2010
[20] L Wu A C Adoko and B Li ldquoAn illustration of determiningquantitatively the rock mass quality parameters of the Hoek-Brown failure criterionrdquo Rock Mechanics and Rock Engi-neering vol 51 no 4 pp 1063ndash1076 2018
[21] C S Ma W Z Chen X J Tan H M Tian J P Yang andJ X Yu ldquoNovel rockburst criterion based on the TBM tunnel
Advances in Civil Engineering 13
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
14 Advances in Civil Engineering
construction of the neelum-jhelum (nj) hydroelectric projectin Pakistanrdquo Tunnelling and Underground Space Technologyvol 81 pp 391ndash402 2018
[22] National Standards Compilation Group of Peoplersquos Re-public of China GB 50287mdash2016 Code for HydropowerEngineering Geological Investigation National StandardsCompilation Group of Peoplersquos Republic of China BeijingChina 2016
[23] P Marinos and E Hoek ldquoGSI a geologically friendly tool forrock mass strength estimationrdquo in Proceedings of the 2000International Conference on Geotechnical and GeologicalEngineering Melbourne Australia November 2000
[24] E Hoek C Carranza-Torres and B Corkum ldquoHoek-Brownfailure criterion-2002 editionrdquo in Proceedings of the NARMS-TAC Conference Toronto Canada July 2002
[25] P Konicek K Soucek L Stas and R Singh ldquoLong-holedestress blasting for rockburst control during deep under-ground coal miningrdquo International Journal of Rock Mechanicsand Mining Sciences vol 61 pp 141ndash153 2013
[26] A Mazaira and P Konicek ldquoIntense rockburst impacts indeep underground construction and their preventionrdquo Ca-nadian Geotechnical Journal vol 52 no 10 pp 1426ndash14392015
[27] P K Kaiser ldquoUnderground rock engineering to match therockrsquos behaviorrdquo in Proceedings of the 50th US Rock Me-chanicsGeomechanics Symposium Houston TX USA June2016
[28] V Marinos P Marinos and E Hoek ldquoe geological strengthindex applications and limitationsrdquo Bulletin of EngineeringGeology and the Environment vol 64 pp 55ndash65 2005
[29] N Barton ldquoSome new q value correlations to assist in sitecharacterisation and tunnel designrdquo International Journal ofRock Mechanics and Mining Sciences vol 39 no 2pp 185ndash216 2002
[30] B Singh ldquoIndian case studies of squeezing grounds andexperiences of application of bartonrsquos q-systemrdquo in Pro-ceedings of the Workshop on Norwegian Method of TunnellingCSMRS New Delhi India September 1993
[31] G S Kalamaras and Z T Bieniawski ldquoA rock mass strengthconcept for coal seams incorporating the effect of timerdquo inProceedings of the 8th ISRM Congress Tokyo Japan Sep-tember 1995
[32] H Rehman A M Naji J-J Kim and H Yoo ldquoExtension oftunneling quality index and rock mass rating systems fortunnel support design through back calculations in highlystressed jointed rock mass an empirical approach based ontunneling data from Himalayardquo Tunnelling and UndergroundSpace Technology vol 85 pp 29ndash42 2019
[33] A M Naji M Z Emad H Rehman and H Yoo ldquoGeologicaland geomechanical heterogeneity in deep hydropower tun-nels a rock burst failure case studyrdquo Tunnelling and Un-derground Space Technology vol 84 pp 507ndash521 2019
[34] E Grimstad and N Barton ldquoUpdating the q-system forNMTrdquo in Proceedings of the International Symposium onSprayed Concrete Modern Use of Wet Mix Sprayed Concretefor Underground Support Oslo Norway 1993
![A New Theoretical Model of Rock Burst-Prone Roadway Support … · 12 hours ago · burst start-up theory [20, 21]. All of the abovementioned models and theories are established on](https://static.documents.pub/doc/80x56/60f00bd604ac715043373397/a-new-theoretical-model-of-rock-burst-prone-roadway-support-12-hours-ago-burst.jpg)