35
D( )Transport and Road Research Laboratory
u TECHNICAL INFORMATION ANDLIBRARY SERVICES
Pleasereturn this document to the Library, or requestan extension of loan, bythedate last stamped below
ACCESSIONNUMBER
s 303
TRANSPORT AND ROAD RESEARCH LABORATORY
Departmentof Transport
RESEARCH REPORT 246
TRAFFIC INDUCED VIBRATIONS IN BUILDINGS
by G R Watts
The views expressed in this Report are not necessarily those of theDepartment of Transport
Vehicles and Environment DivisionVehicles GroupTransport and Road Research LaboratoryCrowthorne, Berkshire, RG 11 6AU1990
ISSN 0266-5247
Abstract
1 Introduction
2 Types of traffic vibration
3 Vibration nuisance
3.1 National environmental survey
3.2 Survey of airborne vibration
3.2.1 Survey method
3.2.2 Results from questionnairesurvey
3.2.3 Prediction of airbornevibration nuisance
3.3 Survey of ground-borne vibration
3.3.1 Method
3.3.2 Results
3.4 Prediction of ground-bornevibration
3.4.1 Tests with HGVS
3.4.2 Propagation tests indifferent soils
3.4.3 Results
3.4.4 Predictive model
3.5 Effects of vibration on sensitiveequipment and tasks
3.5.1 Effects on equipment
3.5.2 Task interference
3.6 Discussion of vibration nuisancestudies
4 Vibration damage
4.1 Possible damage mechanisms
4.1.1 Direct effects
4.1.2 Indirect effects
4.2 Design of experiments
4.3 Fatigue study
4.3.1 Site description
4.3.2 Simulation of vibration
Page
1
1
1
2
2
2
2
4
4
6
6
7
9
9
9
10
11
13
13
13
13
14
15
15
15
15
16
16
16
4.3.3 Monitoring techniques
4.3.4 Results
4.4 Pairwise comparisons
4.4.1 Description of sites
4.4.2 Exposure to vibration andnoise
4.4.3 Comparison of buildingdamage
4.4.4 Results
4.5 Heritage buildings
4.5.1 Description of buildings
4.5.2 Measurement and surveymethods
4.5.3 Results
4.6 Review of data on vibrationdamage in heritage buildings
4.6.1 Results
4.7 Discussion of vibration damagestudies
5 Methods to ameliorate the effects oftraffic vibration
5.1 Airborne vibration
5.2 Ground-borne vibration
5.2.1 Reduction at source
5.2.2 Attenuation methods
6 Conclusions
6.1 Vibration levels
6.2 Effects on buildings
6.3 Alleviation of traffic vibration
7 Acknowledgements
8 References
@ CROWN COPYRIGHT 1990
Page
17
17
18
18
19
19
20
20
20
20
21
24
25
25
26
26
27
27
27
28
28
28
29
29
29
The views expressed in this report are nornecessarily those of the Department of Transport.
Extracts from rhe rexr may be reproduced,except for commercial purposes, provided the
source is acknowledged.
TRAFFIC INDUCED VIBRATIONS IN BUILDINGS
ABSTRACT
Traffic vibration is a common source ofenvironmental nuisance affecting residents. Thisreport summarises TRRL studies of the effects ofthese vibrations on people, buildings andequipment and includes results from other relevantinvestigations. The first section describes thenature of the problem as revealed by questionnairesurveys and details the methods for predicting thedegree of disturbance likely to be caused by bothairborne and ground-borne vibrations. The effectsof vibration on sensitive equipment and criticaltasks are also considered. The second sectionreports on a number of investigations into theeffects of traffic vibration on buildings. Studiesincluded a fatigue test on a vacant property,comparisons of structural defects in housesexposed to high levels of vibration with similarproperties exposed to relatively low levels, andcase studies of heritage buildings adjacent toheavily trafficked roads. It is concluded thatalthough traffic vibration can cause severenuisance to occupants there is no evidence tosupport the assertion that traffic vibration can alsocause significant damage to buildings. Lastly,possible methods for reducing traffic vibrationnuisance are described.
1 INTRODUCTION
Traffic induced vibration in buildings is a commonsource of nuisance affecting residents and undercertain circumstances degrading the performanceof precision measuring equipment. The scale ofthe problem experienced by residents wasindicated in a broadly based survey ofenvironmental disturbances caused by road traffic(Morton-Williams et al., 1978). It was found that37 percent of residents experienced trafficvibration and 8 percent were seriously bothered (iebothered ‘very much’ or ‘quite a lot’). Someinteresting comparisons were made with othertypes of disturbance in this survey. For example, itwas found that traffic noise was heard inside byalmost everyone in the survey, but the percentageseriously bothered by traffic noise (9 percent) wassimilar to the percentage seriously bothered byvibration (8 percent). Thus although trafficvibration is only noticed by a minority of people itseriously bothers a similar number of respondentsto traffic noise. Traffic vibration, therefore,represents a serious environmental disturbanceaffecting large numbers of people. For this reasonTRRL has been engaged on a programme ofresearch to find ways of assessing the effects of
vibration, and methods of reducing its impacts.During the course of this work several reports andconference papers have been published. Thiscompendium report summarizes the results ofthese studies and includes relevant informationfrom the literature. The report updates andexpands an earlier review of the subject area byWhiff in and Leonard (I g71 ).
The first section of the reportdescribesthe natureof the disturbance as revealed by surveys and themethods that have been developed to predict thedegree of disturbance from physical measures.The results should prove useful in assessing theenvironmental impacts of traffic managementschemes or the construction of new roads. Thesecond part of the reDort addresses the importantissue of” whether damage to buildings can becaused by exposure to traffic vibration. Manyresidents believe this to be the case and inparticular there is concern that vibration fromheavy vehicles is damaging heritage buildingswhich may be in a weakened state due to othercauses (Civic Trust, 1970; Crockett, 1966). Thereport summarises a number of studies includingfatigue test in which a recently vacated housewas exposed to high levels of simulated traffic
a
vibration, and a series of case studies of heritagebuildings located adjacent to heavily traffickedroads. Finally, possible methods to ameliorate theeffects of airborne and ground-borne trafficvibration are described.
2 TYPES OF TRAFFIC VIBRATION
Passing vehicles can induce vibrations in buildingsin two major ways. Low frequency soundproduced by large vehicle engines and exhaustshas dominant frequencies in the 50–100 Hz rangecorresponding to the fundamental firing frequency.Inside buildings, low frequency sound can excitethe resonant frequencies of rooms by acousticcoupling through windows and doors. This mayproduce detectable vibrations in building elementsparticularly if they are light and flexible (Martin etal, 1978). High levels of vibration can bemeasured on window panes fronting heavilytrafficked roads and this can give rise to annoyingrattles (Watts, 1984). At the most exposedlocations acoustically induced floor vibrations canbecome perceptible (Watts, 1987; Martin, 1978).However vibration levels in the hard structure ofthe building are much lower.
Ground-borne vibration has dominant frequenciesin a lower frequency band, typically 8–20 Hz.
1
These vibrations are produced by the varyingforces between tyre and road and can becomeperceptible in buildings if heavy vehicles pass overirregularities in the road near the properties. Bothcompression and shear waves are produced andtheir amplitudes and attenuation with distancedepend on a number of factors including the soilcomposition and the nature of the geologicalstrata. Since this vibration enters buildings throughthe foundations, the hard structure of the buildingis normally affected to a greater degree than is thecase for airborne vibration. The principalcomponent of vibration in the ground and atfoundation level is in the vertical direction and thisis readily amplified on suspended wooden floorson upper stories since the natural frequency isoften close to that of the ground-borne wave.Consequently these vibrations are most often feltwhen standing or sitting near the middle of suchfloors. In addition, any horizontal vibration of thebuilding foundations is amplified at the upperlevels of the building.
3 VIBRATION NUISANCE
Vibration nuisance often results from airbornevibration although ground-borne vibration ispotentially a more severe problem under the worstcombination of conditions. This is because ground-borne vibration has been found to produce thegreatest motion in floors and walls and to affectthe whole building whereas airborne vibrationgenerally affects only the front rooms. Theinformation on residents’ dissatisfaction withtraffic vibration was obtained from a nationalsurvey, taken in 1972, of the environmentaleffects of traffic and, more recently, by aquestionnaire survey specifically designed toexamine vibration effects and a jury experimentwhere the response to a range of vehicles wasrecorded. In this section of the report the resultsof these surveys are presented and discussed.Additionally, methods are given for predicting theaverage vibration nuisance along sections of roadwhere airborne vibrations are dominant, and thepeak vertical velocity at the foundations due toground-borne vibration. These values can becompared with established thresholds forperception to determine if disturbance tooccupants is likely to result.
3.1 NATIONAL ENVIRONMENTAL
SURVEY
This large scale survey was based on a sample of5,686 residents and was a cross-section of theadult population of England (Morton-Williams et al,1978). It enabled vibration disturbance from trafficto be compared with the other majorenvironmental traffic-related nuisances such as
noise, fumes and dust and dirt. Figure 1 showsthe percentage of respondents bothered to varyingdegrees by the type of disturbance. In terms ofthe number of people bothered it can be seen thatvibration disturbance is not as prevalent as thatdue to noise and dust and dirt. If the percentagedisturbed is plotted against traffic flow (Figure 2)it can be seen that at higher traffic levels vibrationbecomes relatively more disturbing compared withother nuisances. This probably results from thefact that once vibration is perceived any furtherincrease in level of exposure rapidly becomesintrusive whereas in the case of noise there is amore gradual increase in annoyance withincreasing level. This consideration will beexamined further in Section 3.6 where vibrationthresholds are discussed.
3.2 SURVEY OF AIRBORNE VIBRATION
This survey (Watts, 1984) was specificallydesigned to study the nuisance resulting fromexposure to traffic vibration. The objective was toobtain information on the nature and extent of theproblem and to determine the most appropriatemethod of predicting disturbance from physicalmeasures such as noise and vibration. It wasfound that most of the disturbance was caused byairborne vibration.
3.2.1 Survey method
Approximately thirty people at each of fiftyresidential sites were interviewed. The sites werechosen in the south of England and the Midlandsand ranged from quiet residential roads to heavilytrafficked dual-carriageway. Questions on thetypes of vibration noticed, possible damage toproperty caused by vibration and the types ofvehicle and the operating conditions that hadproduced noticeable building vibration wereincluded. An overall rating of the vibrationnuisance was obtained using a seven point scaleviz: —
NOT AT ALL O 1 2 3 4 5 6 EXTREMELY
BOTHERED ~ BOTHERED
The site median vibration nuisance ratings weredetermined from these scores and were used asan overall measure of annoyance at the differentsites. An identical scale was used to obtain anoverall rating of noise nuisance. At one house persite external noise and window vibration wererecorded for 15 minutes every hour over 24 hours.At a later stage the effects of ground-bornevibration were assessed by recording vibrationnear the facade and in the middle of the groundfloor at a small number of houses where ground-borne vibration was likely to be perceptible. Thelevels derived from the analysis of noise dataincluded linear (un-weighted) levels for thefrequency ranges 25–4000 Hz and 40–125 Hz,
2
80 “
70
60
50
wEE% 40%s
30
20
10
Vibration
. . . . . . . . . .. . . . . . . . . . .. . . . . . . . . .
Fumes Dust and dirt
. . . . . .
Parking
Fig. 1 Percentage bothered in the home by type of disturbancecaused by traffic
Source: Morton - Williams et al 1978
I Key:
Bothered -very much
quite a lot m
I not very much ~
I not at all n
Noise
Dust and dirt/“’
-4 -d--0-----
. ..9●.**
Parking. ..*
●. . . .--...9==.-- ..--=
Vibration
Fumes
01 I I I I 1 1 1
25 50 100 200 400 800 1600
Vehicles per hour during peak period
F ig.2 Percentage of respondents bothered in the home by variousdisturbances by traffic flowSource: Morton - Williams et al 1978
3
weighted levels A, B and C and octave levels at63 and 80 Hz.
3.2.2 Results from questionnaire survey
Results were obtained from over 1600 completedquestionnaires. The percentage of respondentswho noticed various traffic-induced vibrations intheir homes is given Table 1. A large percentage(62 percent) noticed windows or doors rattling orbuzzing and 16 percent were aware of ornamentsvibrating. Vibrations were also perceived directlythrough tactile stimulation, for example 30 percenthad noticed that the floor shook or trembled.Table 2 lists the percentage of residents whoreported various types of damage thought to becaused by road traffic. Fewer people reportedserious structural damage (cracks in brickwork ordamaged foundations) than architectural defectssuch as cracks in plaster finishes. From Table 3 itcan be seen that an important reason forrespondents being bothered by vibration was thepossibility that traffic induced vibration haddamaged (20 percent) or could damage (55percent) their homes. Large lorries were mostoften mentioned as causing vibration (73 percent)and buses were the next most frequently reportedvehicle (51 percent).
In a further analysis (Watts, 1985a) the averagepercentage of residents bothered by vibration andnoise at various levels of noise exposure (usingthe L1o (1 8-hour) dB(A) scale) was calculated(Figure 3). The large fluctuation at low exposurelevels is probably due to sampling error since onlya small number of sites was used to compute thepercentage bothered. It was considered that asigmoid curve was the most appropriate functionto describe these data and the best fit wasobtained by taking the Iogit transformation of thepercentages and using least squares analysis. Itcan be seen from the figure that at a given valueof Llo (1 8-hour) a higher percentage ofrespondents were disturbed by noise thanvibration effects and this is true throughout thenoise exposure range sampled. This result is inagreement with the findings of the national surveywhere it is likely that traffic flow was acting as aproxy for vibration exposure (see Section 3.1 andFigure 2). A similar trend of percentage botheredwith noise (ie more than ‘moderately annoyed’)with exposure has also been reported (Fields andHall, 1987).
3.2.3 Prediction of airborne vibration
nuisance
In an attempt to determine a method of predictingthe nuisance caused by airborne traffic vibration,median vibration scores were correlated withvarious noise, window vibration and traffic flowparameters. The median scores at each site werecomputed from the individual vibration nuisance
TABLE 1
Percentages of respondents who noticedvarious vibrations
I
Vibration effectPercentage
noticing effect
Windows or doors rattling orbuzzing
Floors shaking or tremblingOrnaments rattling or buzzingTraffic causing the bed to shakeMuffled sensation in the ears or
fluttering sensation in the chestFeeling vibration in the air
62.229.515.713.6
18.930,2
TABLE 2
Percentages of respondents reporting damagethought to be caused by road traffic
Damage reportedPercentage
reporting damageI
Roof tiles falling or movingCracks in plaster on walls or
ceilingsCracks in brickworkCracked windowsSubsidenceDamaged foundations
31.6
25.810.019.913.7
7.6
TABLE 3
Percentages of respondents bothered byvibration for various reasons
ReasonPercentage
bothered
It has damaged this house/flatIt could damage this house/flatIt interferes with sleepIt makes you jump, or frightens youIt gets on your nervesIt feels unpleasantIt reminds you of the trafficIt interferes with the TV picture
19.654.735.927.044.641.755.927.3
4
90
80
70
60
umLw; 50nm~E 40u&n
30
20
10
======== A Vibration bothers very much or quite a lot— ❑ Noise bothers very much or quite a lot
❑
A
❑
❑
A/***..G.*...”......*”Best fitting curves defined by: ~NPercentage bothered = 100. Ae
l+AeBN. ..**-
. . . . Where N is the Llo (18 hour) level in dB●..s
●. . .. ..0 For noise bother A = 0.00110 and B = 0.0962
A For vibration bother A = 0.000157 and B = 0.117
I I 1 1 I I
55 60 65 70 75 80
Noise exposure Llo (18 hour) dB(A)
Fig.3 Percentages of respondents bothered by noise and vibration caused
by traffic
ratings and therefore are an indicator of overallsite nuisance. It was found that the highestcorrelations were obtained with the various18-hour noise measures which are listed in Table4. Of the measure examined, the Leq 18-hour wasmOst closely associated with the median vibrationnuisance rating (r= 0.71 ). It was expected that alow frequency noise measure would be mostclosely related to these disturbance scores sincethis noise has been shown to be responsible forthe common manifestations of vibrations such aswindow and door rattles. These noise measureswere in fact marginally less well correlated thanthe dB(A) index, but there are no statisticallysignificant differences between any of thesecorrelations. Figure 4 shows a scatter plot for theLIO (18 hour) dB(A) index. This index is widelyused for the assessment of traffic nuisance. Theregression line is based on data from 49 sites, not50, since at one quiet site adverse reactions werethought to have been caused by the recent
opening of a vehicle testing station (see Watts,1984). A reason for the small range in the sizes ofthe correlation coefficients in the present study isthe high level of association between most of thenoise measures over the 50 sites.
The results from a further study involving a groupof residents who rated vibration disturbanceconfirmed the main result (Watts, 1985b). Thejurors were seated in a living room fronting aheavily trafficked road and were asked to makenuisance ratings of selected vehicles in the trafficstream. Outside and inside noise levels wererecorded and it was possible to relate noisemeasured on various scales to median ratings foreach vibration event. As in the 50 site surveys, itwas found that there was a relatively small rangein the correlation coefficients for the various noisemeasures and that the dB(A) scale was among themost highly correlated measures.
TABLE 4
Regression coefficients relating noise exposure measures with median
18-hour noiseexposure measure (x)
63 Hz octaveLL:L,
80 Hz octaveLeq
Linear 40–1 25 HzLeq
Linear 25–4000 HzLeq
Weighted ALL:LSL,
Weighted BLeq
Weighted CLL:LI
Log of number of eventsgreater than90 dB (LIN)95 dB (LIN)
vibration score y as dependent variable
Regression coefficientsy=ax+b
a
–10.98–10.63–11.72
–9.86
—
—
2.16
3.24
0.449.28
– 10.06–11.41
–11.73
—
—
3.133.004.09
0.2741.503
3.3 SURVEY OF GROUND-BORNE
VIBRATION
Results from the 50 site survey demonstrated that18-hour noise levels at the facades of dwellingscorrelated reasonably well with ratings of vibrationnuisance, indicating the importance of acousticallycoupled vibration. However the best correlationcoefficient achieved was 0.71, which implies thatonly 50 percent of the variance in the site medianscores is explained by this measure. It wasconsidered that a better level of association mightresult if measures of ground-borne vibration wereincluded as part of the physical descriptor. Sincethe principal component of this source of vibrationis in the vertical direction, the most likelymanifestation of this type of vibration wasconsidered to be the vertical movements ofsuspended floors. This would be particularlyimportant where the forcing frequency of thevibration was close to the natural frequency of thefloor since relatively high amplitudes could then
b
0.1770.1630.166
0.164
0.190
0.201
0.1940.1710.1780.186
0.199
0.2010.1890.192
1.3631.259
Correlationcoefficient
r
0.630.600.61
0.61
0.65
0.66
0.710.680.690.69
0.68
0.670.650.66
0.690.68
Standarderror of
estimate
0.880.900.89
0.90
0.86
0.85
0.800.830.820.81
0.83
0.840.850.85
0.820.83
occur. If these vibrations were above thethreshold of perception then this could lead todisturbance in addition to any annoyance due toacoustic excitation and may result in higher thanexpected ratings. Perceptible ground-bornevibrations would be expected in dwellings situateda few metres from roads with uneven roadsurfaces and carrying HGVS. Further investigationsof the road surface profile were therefore made atthe survey sites to explore this possibility (Watts,1987).
3.3.1 Method
At each of the 50 sites the TRRL Bump-integrator(Jordan and Young, 1980) was used to obtain anaveraged measure of surface unevenness. Thisinstrument consists of a single-wheeled trailerwhich is towed along the road at a constant speedof 32 kmihr. When in operation, the downwardmovement of the wheel relative to the chassis ismeasured and integrated. The unevenness index
6
Outlier not included in the)— computation of the regression
line
●
●
●
● ●
●
●
●
●
● ● -/0
-0 ●● ●
●
●
●
● ✚●
●
Regression line: site median rating‘0.199x L1odB(A) -11.27
62 66 70 74 78
Noise exposure LIO (18 hour) dB(A)
Fig.4 Median vibration nuisance rating and LIO (18 hour) dB(A)
‘r’ is computed by dividing the integrated verticalmovement by the distance travelled. As before,the site median vibration nuisance score was usedas the dependent variable in the regressionanalyses. It was expected that the contribution tothe nuisance produced by the road surface wouldbe primarily a function of the number of HGVS,the distance of the house facades from thenearside kerb (d), the unevenness index (r) and thespeed of vehicles on the road (Rudder, 1978).Other parameters such as soil properties and theresponse of the road structure to the dynamicloads produced by vehicles were also expected tobe influential but these factors could not be readilydetermined at this stage and they were thereforenot included. Stepwise multiple regression wasperformed using the SPSS suite of programs (Nieet al., 1975). In this method the independent
variables are entered one at a time. The variablethat explains the greatest amount of varianceunexplained by the variables already in theequation enters the equation at each step. In thisway an optimal prediction equation was developedwith as few terms as possible.
It was found that generally higher correlationswere obtained by taking logarithmic transforms ofthe explanatory variables.
3.3.2 Results
The best fitting regression coefficients at eachstage of the stepwise regression are given inTable 5. Most of the variance is explained afterstep 2 when the number of HGVS passing the sitein an 18 hour day and the distance of the facades
7
TABLE 5
Results of stepwise regression analysis relatingannoyance caused by traffic vibration at 50 sites
to various site factors
Step
1
2
3
4
Median vibration annoyancerating predicted by: –
0.8:30 Ioglo (1 + No of HGVS)+0.918
0.926 Ioglo (1 +No of HGVS)–2.17 log,Od +2.69
1.10 Ioglo (1 +No df HGVS)–1.86 loglOd +2.16 Ioglor
–2.88
0.845 Ioglo (1 + No of HGVS)–2.28 loglOd +2.30 Ioglor+
3.41 * Ioglo (speed limit)–7.64
Multiplecorrelationcoefficient
(R)
0.573
0.727
0.757
0.774
Independent variables:
(No of HGVS) is the number of heavy goodsvehicles with three or more axles passing the sitein an 18 hour day.‘d’ is the distance from the facade to the nearside
kerb,‘r’ is the unevenness coefficient.‘Speed limit’ is limit prevailing at site.* Regression coefficient significant at 5 per centlevel.
from the kerb are included in the equation. Addingthe surface unevenness term only accounts for afurther 4.6 percent of the variance of the scoresand the coefficient is only just statisticallysignificant at the 5 percent level. The speed term(the logarithm of the speed limit at the site) is notsignificant and only accounts for a further 2.5percent of the variance. Further measurements ofvibration near the house foundations revealed thatground-borne vibrations were only likely to beperceptible at a relatively small number of houseswhich were close to significant surfaceirregularities. This explains the observation thatthe measure of road roughness, which is averagedover a length of road of the order of a 100 m longat each site, does not contribute substantially tothe variation in median disturbance scoresbetween sites.
The peak accelerations recorded near foundationlevel at dwellings where vibrations were wellabove perception level are given in Table 6. Thedata listed refer to four dwellings in Swindon andone dwelling in the London Borough of Brent.
The dominant frequencies were generally low, andthey are typical of ground-borne vibration. Atthese sites, ground-borne, rather than airborne,vibration produced the highest peak levels ofvibration in the hard structure of buildings andconsequently probably had the greatest potentialto cause structural damage.
TABLE 6
Peak accelerations measured at sites with perceptible vibration
Site
Swindon A
B
c
D
London Borough of Brent
Vehicle producingvibration
5 axle artic
2 axle rigid
3 axle cement mixer
4 axle rigid
Double decker bus
Position
Foundation*Floor* *FoundationFloorFoundationFloorFoundationFloorFoundationFloor
Peak verticalacceleration
level(mms-z)
75175130164110114
42785796
Dominantfrequency
(Hz)
121217412.712.51312.56025.51024
* On ground within 0.5 m of foundations.* * Middle c)f ground floor at front of house.
3.4 PREDICTION OF GROUND-BORNE
VIBRATION LEVELS
Although ground-borne vibration problems are notlikely to be as widespread as those produced byairborne vibration, maximum amplitudes can reachrelatively high levels, well above the level ofperception in unfavorable circumstances, andcould cause anxieties about property damage.Vigorous complaints might therefore be expectedunder the worst combination of conditions.Ground-borne vibration effects are much moredifficult to predict than those due to airbornevibration which can be simply estimated from anacoustic measure such as L1o (18-hour) dB(A) (seeSection 3.2.3). Ground vibrations are dependentupon a number of factors which include vehiclecharacteristics such as axle load, suspensiondesign and operating speed, the road surfaceprofile and the nature of the ground between theroad base and the building foundations. Theresponse of the building to the vibrations occurringin the ground near the foundations adds a furtherdegree of complexity to the problem.
Despite the obvious difficulties of determining apractical prediction method, a relatively simpleprediction technique has been developed whichenables peak vertical vibration levels at thefoundations of buildings to be determined. Tracktests at TRRL with a wide range of HGVSestablished the trends in peak vibration levels withvehicle speed, load and size of irregularity (Watts,1988a). These results were then generalized todifferent site conditions by determining theamplitudes and attenuation rates of vibrationgenerated in different soils using a controlledimpact method. By determining the averageeffects of these factors it was possible toestimate the likely range of the maximumamplitude (or peak) of vertical particle velocity(PPV) at the foundations of buildings for a varietyof site conditions.
It should be noted that occasionally joints inconcrete roads can give rise to perceptiblevibrations due to slab movement as heavy vehiclespass by. Investigations of this effect are describedin Watts ( 1987). There are obvious difficulties inattempting to model the generation of vibrations inthese situations and consequently predictions thatare developed below are for the much morecommon situation where a significant irregularityin the road surface is the cause of vibrationeffects.
The following parts of this Section review thedevelopment of this technique but a fullerdescription can be found in a paper by Watts(1989a).
3,4.1 Tests with HGVS
Eight HGVS, comprising rigid and articulatedvehicles, were tested by running the vehicles overartificial humps and a depression on the TRRLresearch track. The suspension systems for thetrailers covered a wide range and included twoand three axle bogies with steel and airsuspension and a two axle rubber sprung system.The vehicles were tested at speeds up to 80 kmlhfully laden and empty over a range of profiles.These were designed to represent a wide varietyof surface defects resulting from poorly backfilledholes and trenches on public roads. Figure 5 givesdetails of these profiles. The particle velocitiesproduced on the track surface by groundvibrations were measured by triaxial geophonearrays fixed at 2 metres and 6 metres from thenearside wheel paths.
Direction of travel~ I
.... . .. ... ....~?6m
t 25mm&\\\\\\\\\\\\\\\\~\\\\\\~\\\\\\m --
. . . , . . . .. .. ... .. ... .. . .. . . . . . . . . . .. .. . ... ... .. .. ...1.24m .
I
. .
.. .. ... .. . . . . .“1. : ..:...... . .
FNote different vertical scale to show —
form of profile
Note: Except for these test irregularities, the track hada level surface to within t 7mm within 5m of the midprofile position
Fig.5 Test profiles
3.4.2 Propagation tests in different soils.
Previous measurements had indicated that trafficvibrations generated in soft ground such asalluvium and peat soils were much greater thanwas the case for firmer soils under broadly similarconditions (Watts, 1988b). It was thereforeessential to make corrections for groundconditions when extrapolating from resultsobtained on the research track where thesubgrade is firm sand and gravel deposits. Thiswas achieved by measuring the transfer functionbetween a suitable force input to the road and theresulting ground vibration for representative soiltypes ranging from very soft to very firm. Oncedetermined, these functions would allow PPVS tobe calculated for a particular site by factoring thePPV expected on the research track by the ratioI H.(f) I / I H,(f) 1, where I H,(f) I and I Ht(f) I are themoduli of the transfer function or nobilities at thesite and on the track and ‘f’ is the forcing
9
frequency. This frequency is typically between10–1 2 Hz and results from the ‘wheel hop’ modeof vibration of the HGV suspension (ie theoscillatory motion of the wheels between thevehicle body and road surface).
To determine the transfer functions, the FallingWeight Deflectometer (FWD), which was designedto measure road pavement characteristics,(Sorensen and Mayven, 1982), was used toproduce very carefully controlled road surfaceimpacts. The FWD’S electronics were adapted toenable the recording of the time histories of thedynamic force at the road surface and theresulting ground vibration from the impact of thefailing weight. Measurements were made at 13sites on a range of soils from very soft (peat) tohard (chalk rock). Triaxial geophone arrays wereplaced just beneath the surface of the ground atdistances of 3, 6, 12, 25 and 50 metres from theFWD where possible. The expected values of PPVat building foundations were obtained frommeasuremt?nts in unloaded soils by applying asuitable factor which was obtained fromcomparing results obtained on buildings with thoseexpected in the same ground at similar distances.
3.4.3 Results
Figure 6 shows a plot of vertical PPV at 6magainst speed for the range of fully laden HGVSrunning over the 0.6 m x 25 mm high profile.There is a scatter of results indicating differencesin the vibration produced by the various vehicles,but the trend with speed is clearly defined. Forsimplicity the differences between vehicles are notincluded in the model and the linear regression linewas used to establish the relationship with speed.Further tests were made with smaller two axlevehicles ranging from a small estate car to a tipperlorry. As expected these vehicles generallyproduced lower peak levels. This can be seen inFigure 7 where PPV is plotted against grossvehicle weight for the same profile and crossingspeed that were used in the HGV tests. There isan obvious trend of rising PPV with increasingvehicle weight. Since the main aim was to predictthe maximum likely PPV in a stream of traffic,these smaller vehicles were not considered in thedevelopment of the prediction model.
In Figure 8 the vertical PPVS at 6 m produced byeach fully laden HGV traveling over the profiles at48 kmlh are plotted against the maximum heightor depth of the profile. It can be seen that thisdimension of the profile is a reasonable indicatorof the likely PPV despite the wide range of profileshapes employed. Linear regression analysisshowed that the average PPV at 6 m for a 25 mmhigh profile was 0.7 mm/s, the slope being0.028 mm/s per mm increase in height or depth.The trend with speed over the 25 mm high profileis linear to a good approximation (Figure 6) and
2.0
1.5 [
>nn
0.5
0
I ..
0 20 40 60 80
Speed(km/h)
Fig.6 Vertical PPV at 6m with vehicle speed over0.6m x 25mm hump - TRRL hump
0.8
0.6[
o0 5
●
7.5 tonne(common weight limit forlorry bans)
1 1 I10 15 20
Gross vehicle weight (tonne)
Variation of vertical peak particle velocityat 6m with gross vehicle weight for loaded 2 axlevehicles crossing a 0.6x 25mm hump at 48 km/h
2.0 r1.5
1.0
0.5
.
.
16 points:
i
8 ooints ●
I I I I
o 15 30 45 60
Profile height or depth (mm)
Vertical PPV at 6mdepth at 58 km/h -
with profile height orTRRL track
10
this was found to be the case for the otherprofiles. In many cases on public roads, theirregularity is significant in only one wheel pathand this is usually on the nearside. Tests haveshown that in this situation PPVS are generallysignificantly lower than the values recorded whenidentical profiles are in both paths. The factor0.75 was considered appropriate for scalingpredicted values in these cases.
Figure 9 shows how the moduli of the transferfunctions at 12 Hz vary with distance for sixdifferent ground conditions. The Figure clearlyshows that there is a large difference of up to twoorders of magnitude between the response of asoft soil (peat) and very hard ground (chalk rock).Linear regression lines were fitted to thelogarithmic transform of the data at 12 Hz foreach site and the correlation coefficients rangedfrom 0.922 to 0.998, indicating good agreementwith an attenuation model based on a simplepower law r’, where x is the power coefficient.Table 7 shows that there is a range of powercoefficients even for similar soil types. This tablealso gives the average moduli of the transferfunction at 6 m for six ground conditions, togetherwith the corresponding values expected at buildingfoundations.
1.0 3.2 10 32 -100
Distance (m)
Fig.9 Transfer function for vertical PPV at12 Hz by distance
3.4.4 Predictive model
The method is based on making predictions ofvertical PPV at 6 m using the observed trendswith amplitude of road surface irregularity andspeed, and the most appropriate ground scalingfactor. By combining these factors the expectedvalue of maximum vertical PPV at a buildingfoundation can be calculated as: –
0.028 .a. (v/48 ).t.p. (r/6)x
4.0r /
1.0
0 1.0 2.0 3.0 4.0
Measured PPV (mm/s)
Fig. 10 Measured and predicted maximumvertical PPVS at building foundations dueto HGVS passing over road sutiace irregularities
Where a = maximum height or depth of the surfacedefect in mm, v = maximum expected speed ofHGVS in kmlh and t = ground scaling factor (seeTable 7). If the surface defect occurs in one wheelpath only then p = 0.75, otherwise p =1, r is thedistance of the foundation from the defect inmetres and x = power factor, which can beobtained from Table 7 for the most appropriatesoil type.
To check the accuracy of the formula, predictionswere made at a number of sites where long-termmeasurements of vibration had been made. Figure10 compares actual and predicted results,indicating that predicted results for the maximumvertical PPV are generally in reasonable agreementwith the measurements.
The outlier is for a site thought to be on softalluvium although the measured value suggests aground scaling factor for a firmer soil would bemore appropriate.
This prediction method should be useful initially indetermining whether there is likely to be a groundvibration problem arising from road surface defectsand the possible scale of the effects. Little isknown about the subjective response to ground-borne vibration from traffic and so it is necessaryin the first instance to consider whether thesevibrations are likely to be perceptible. If themaximum PPV at foundation level is significantlyin excess of the threshold of perception thendisturbance to occupiers and even complaints maybe expected. The threshold of perception isdiscussed in Section 3.6, but is of the order of0.3 mm/s.
11
TABLE 7
Effect of ground characteristics on transmission of vibration
Modulus of transfer function ( x 103)Power coefficient for in mm/s per kN
attenuation withGround type distance ‘ x‘ ● * In ground at 6 m Expected
Number value atCondition of sites foundations
Description (if known) tested Range Average Range Average 6m
Peat *
}soft ‘;
— –1.19 1a9 41.9Alluvium
—
–0.79 tO –o. ao –0.79 72.5 to a2.O 77.3 77.3London clay 3 –o.99to –1.13 –1.06 20.9 to 56.3 33a 33aSand/gravel 3 –0.69 to –o. a2 –0.74 9.92 to 11.0 10.3 10.3Boulder clay 3 –0.71 tO –1.la –0.93 2.43 to 6.67 4.73 4.73Chalk rock 1 — – 1 .oa — 1.14 1.14
Ground***scalingfactort=~
IH,l
3.a47.073.100.940.430.10
* For peat soil transfer function values and power coefficient are for 10 Hz.* * Power law for attenuation with distance is rx where r is distance from source.
● ● * lH~l and IH,I are the moduli of the site and track transfer functions.
3.5 EFFECTS ON SENSITIVE
EQUIPMENT AND TASKS
~ further type of nuisance can arise whensensitive equipment is affected by traffic vibration(Wooton, 1975). Such problems are usuallyconcerned with specialist buildings such aslaboratories or workshops. For example there havebeen claims with regard to computer installations,a hydraulics laboratory (where vibration couldeffect the water surfaces of the large modelestuaries), a machine tool laboratory andinstrument workshop. Disturbance to equipmentcan also occur in school and universitylaboratories (Whiff in and Leonard, 1971) wheredifficulty can be experienced in reading sensitivegalvanometers and operating chemical balancesand various types of high magnificationmicroscope. The vibration effects due to peopleusing the building are sometimes overcome bymounting such equipment on slabs or framesattached to solid walls or concrete pillars.However, this does not prevent the equipmentfrom being affected by ground-borne trafficvibration since these vibrations enter the buildingthrough the foundations and propagate readilythrough the hard structure of the building.
3.5.1 Effects on equipment
Information has been collected on the vibrationlevels which impair the operation of various typesof laboratory measuring equipment (Ferahian andWard, 1970; Instrument Society of America,1975; Whiff in and Leonard, 1971). It appears thatthe satisfactory operation of electron microscopesis particularly dependent on low frequencyvibration levels. For example one manufacturer hasspecified a PPV of 0.46 mm/s in the region10–1 5 Hz corresponding to the principlefrequencies of ground-borne traffic vibration, whileanother has set the maximum permissible level aslow as 0.04 mm/s. More recently Holmberg et al.( 1983) have reported a pilot survey amongcomputer manufacturers which was designed toestablish threshold values of vibration abovewhich equipment may malfunction. Disk storageunits are considered the most vibration sensitivedevices in computer systems. This is because theaccess heads are typically supported by a thincushion of air only 2 pm above the rotating disk. Ifdue to vibration the access head touches the diska failure would occur which could damage the diskresulting in expensive replacements and loss ofstored data. Unfortunately threshold values givenby manufacturers were defined in different waysand seldom were the measurement positionsspecified. The results indicated that thresholdvalues for computer systems ranged from 0.9 to46 mm/s for the PPV of continuous vibration at12 Hz. The authors considered these values toolow and could probably be increased.
3.5.2 Task interference
The efficiency of personnel carrying out delicatetasks requiring a high degree of skill may also beaffected by the presence of vibration at the workplace. The vibration may interfere directly with thetask itself or may produce an annoying distraction.The 8ritish Standards Institution providesinformation on peak velocities in buildings belowwhich comments or complaints are rare (BSI,1g84). For critical working areas such as hospital
operating theatres and some precision laboratoriesthe guide value for vertical peak amplitude is inthe region of the threshold for perception of about0.3 mm/s.
3.6 DISCUSSION OF VIBRATION
NUISANCE STUDIES
Methods for the prediction of the averagenuisance and the percentage of residents likely tobe disturbed by traffic vibration at sites whereairborne vibration predominates have beendescribed. Both the vibration survey and juryexperiment show that the median vibrationnuisance score can be predicted by a number ofdifferent acoustic measures. This is likely to resultfrom the fact that many of the different measuresof noise levels are themselves highly correlatedwith each other. Therefore, although the dB(A)weighting attenuates the contributions from lowfrequencies (eg the 63 Hz third octave level isattenuated by nearly 30 dB) it can still act as aproxy for the low frequency sound which is likelyto largely condition residents’ judgments ofvibration nuisance. For example, the lowfrequency linear noise level Leq (40–125 Hz) waswell associated with L1o dB(A) in the survey(r= 0.91 ). Consequently, for ease of prediction theLIO dB(A) index may be preferred since it is widelyused and can itself be predicted from trafficparameters, road surface texture and sitegeometry (Department of Transport and WelshOffice, 1988). Section 3.3.2 shows that medianvibration nuisance can also be predicted withsimilar precision by a composite measure basedthe 18 hour HGV traffic flow and the distancefrom the front facade to the carriageway.
The suitability of these prediction methods in a
on
range of circumstances must be considered if theyare to be widely used in environmentalassessment. The results of the survey wereobtained from a study of 50 residential sites. In allcases the sites had simple geometries in that therewere no intervening barriers to noise propagationsuch as other buildings, screening barriers, ornatural features such as large earth mounds.Therefore extrapolation of the results to morecomplex situations should be carried out with thepossible limitations clearly in mind. Consequently,where low frequency sound is attenuated bybarriers of various types, predictions based on the
13
equation developed in Section 3.3.2 involvingsimply lorry flow and distance from road todwelling may overpredict the likely disturbancesince no account is taken of screening effects.Predictions of annoyance based on dB(A) levelswhich in turn have been predicted using theDepartment’s calculation method which takescreening effects into account (Department ofTransport and Welsh Office, 1988) may lead tounderprediction since low frequencies may beattenuated by a smaller amount than higherfrequencies (Hothersall et al., 1989). For thesereasons, predictions for these more complexsituations should be made cautiously. Greaterconfidence can be placed on predictions where thesite geometries lie within the range of variablescovered in the survey.
Section 3.4.4 describes a method for makingpredictions to determine whether ground-bornevibrations are likely to be above an establishedthreshold of perception at foundation level.However, it is necessary to determine if aparticular threshold is applicable to traffic vibrationand the Ievt?l at which the vibrations might beexpected to become unacceptable. Early studiesby Reiher and Meister demonstrated that forsinusoidal vibration the threshold of perception inthe vertical direction was 0.3 mm/s (Steffens,1974) and, recently, similar results have beenobtained for frequencies near the wheel-hopfrequency (Parsons and Griffin, 1988). In the latterstudy the effects of short duration sinusoidalvibration were examined and this is particularlyrelevant since the time histories of ground-bornevibration from HGVS traveling over a surfacedefect often reveal just one or two major peaksper axle. The threshold for these short durationevents of one or two major cycles was 1.7 timesthe value for continuous vibration at the samefrequency (16 Hz). In a further test, subjects wereasked to adjust the vibration level until theyconsidered it would be just unacceptable if itoccurred in their own home. It was tentativelyconcluded that vibration may becomeunacceptable when the threshold is exceeded by afactor of two, and consequently it appears thatvibrations due to ground-borne traffic vibrationmay become unacceptable above a level of1 mm/s. It should be noted that this value wasderived from the average response of a relativelysmall sample of subjects. Clearly there will besome residents who are more sensitive tovibration effects and who will find lower levelsunacceptable.
A further consideration is the extent to whichvibration levels at the foundations relate to levelson living room and bedroom floors. A number ofstudies have shown that ground-borne vibrationlevels on ground floors are similar to those at thefoundations but that amplification often occurs athigher levels in buildings (Watts, 1987; Watts,
1988c; Watts, 1989b). For examPle, it is possiblethat peak vertical particle velocities in the middleof upper floors may be several times that at thefoundations. Potentially this could lead to greaterannoyance, although a person lying down, forexample, may not be exposed to these higherlevels because of the attenuation afforded by themattress and springs of the bed. On the availableevidence it is not possible to give precise guidanceon the level at the foundations above whichcomplaints from occupants can be expected.However if the levels are significantly above0.3 mm/s then some degree of disturbance willprobably occur while if levels are well in excess of1.0 mm/s then this may prove unacceptable andcomplaints may be made.
It appears that under unfavorable conditionstraffic vibrations have the potential to degrade theperformance of sensitive equipment and interferewith delicate tasks. It is difficult to give generalguidance on the levels of vibration and theirfrequencies which will produce these effects sincethe response of a particular system will dependnot only on design details but may also bedetermined by the mounting conditions. Forexample, the problem can occur because lightlydamped, finely balanced, movements in someprecision measuring equipment exhibit sharpresonance peaks when excited at certainfrequencies. A small change in the design detailsmay shift the resonant frequency outside therange of ground-borne vibration and, as aconsequence, the detrimental effects may not beobserved. For this reason the effects of vibrationare likely to vary greatly even for equipment of asimilar type. In some cases equipmentmanufacturers may provide sufficient details aboutvibration sensitivity to allow the likely impact oftraffic vibration to be reasonably assessed. Incases of doubt it may be necessary to expose theequipment and operator to a range of frequenciestypical of that produced by traffic vibration, andestablish the vibration levels which produce adetrimental effect. These levels can then becompared with levels expected from traffic at theparticular location.
4 VIBRATION DAMAGE
There is concern about the effects of trafficvibration on buildings close to heavily traffickedroads. The survey of vibration nuisance describedin Section 3.2 has shown that over half therespondents were bothered by traffic vibrationbecause they felt that traffic vibration coulddamage their homes; nearly 20 per cent allegedvibration damage had already occurred. In additionthe Civic Trust, consulting engineers andacademics who are involved in the preservation ofhistoric buildings and monuments have expressed
14
concern at the effects of traffic vibration on thesesensitive buildings (Civic Trust, 1970; Crockett,1966 and 1973; Bata, 1971). Increases inallowable vehicle weights may have heightenedconcerns and anxieties (Armitage, 1980). Despitethese concerns there was little evidence tosupport or reject these beliefs and it wasnecessary to study the problem using a variety oftechniques and involving a wide range of buildingsand soil types so generalizations could be madewith some degree of confidence.
4.1 POSSIBLE DAMAGE MECHANISMS
There are four mechanisms that may result invibration damage in buildings. Three can affect thestructure directly and the fourth may act indirectlyby modifying the underlying soil which in turn mayaffect the structure.
4.1.1 Direct effects
If vibration levels are high enough the stressesimposed by shear and compressional waves cancause failure of building components. Much workhas been carried out by the USA Bureau of Mineswhere the effects of vibration from blasting havebeen extensively studied. Peak particle velocity ofthe hard structure of the building near foundationlevel is the measure most frequently used since itcan be related to the stresses imposed on thestructure by the propagating waves (New, 1988).Studies such as these have shown no conclusiveevidence of significant vibration damage below aPPV of approximately 10 mm/s (House, 1973;Nelson and Watts, 1988), whereas measurementsat the foundations of buildings adjacent to heavilytrafficked roads have shown PPVS up to only3.5 mm/s (Watts, 1988b).
Although these peak levels from traffic are wellbelow vibration levels that have been shown toproduce damage, it is not inconceivable that directdamage may occur at lower levels. A smalladditional stress imposed by traffic vibration mightpossibly add to a much greater static stressresulting in damage. Such a ‘trigger’ mechanismcould perhaps cause premature failure in a buildingcomponent already weakened by other causes. Amore widespread concern is the possibility offatigue damage occurring as a result of longperiods of exposure to low levels of vibration.Buildings close to heavily trafficked roads may beexposed to many thousands of stress cycles eachday so that the vibration dose over many yearscould be considerable.
4.1.2 Indirect effects
It is known that granular soils such as sand canbe induced to change volume if subjected tovibration. This phenomena has been studied underlaboratory conditions (Linger, 1963) and thetendency is for the soil to densify as the particles
move closer together under the action of vibration.Such assisted densification could lead tosettlement and structural damage if it occurredunder building foundations. The risk of seriousdamage would be particularly high if differentialsettlement occurred due to the relatively highexposure of the front foundations compared withthe rear where vibration levels would beattenuated to some degree. Buildings probably atgreatest risk are those constructed without properfoundations on loose or low density sands or softsoils. Such vibration assisted settlement has beensuggested as a cause of tilting of the walls ofchurches and cathedrals towards the nearestheavily trafficked roads (Crockett, 1973).
4.2 DESIGN OF EXPERIMENTS
There are a number of difficulties in attempting tostudy the possible effects of traffic vibration onbuildings. An important consideration is the scaleof the likely effect when compared with thoseproduced by other causes of damage. A cursoryinspection of buildings adjacent to busy roadsshow that they are not obviously deterioratingfaster than similar buildings further away. Trafficvibration in city centres is subjecting buildings tohundreds of millions of stress cycles every yearwithout any obvious widespread damaging effects.It is therefore likely that, if they occur at all, theeffects of traffic are probably relatively small,taking many years to have any measurable effect.The research method must therefore be capable ofseparating any small damage caused by trafficvibration from damage due to natural ageing andweathering of materials and settlement that mighttake place on loose or soft soils. In addition, majoralterations and additions to buildings may, overseveral years, have a significant impact onstructural integrity.
There were four types of study that were carriedout in order to quantify the possible effects ofvibration.
(i) A fatigue study was carried out on anunoccupied dwelling using simulated trafficvibration. This separated the effects of vibrationand ageing and allowed virtually unambiguousevidence to be obtained on the isolated effects ofvibration since the site conditions remainedconstant except for the addition of vibration. Itwas possible to examine the effects of bothairborne and ground-borne vibration.
(ii) A pairwise comparison of occupied buildingsfronting heavily trafficked roads with similarbuildings in the neighborhood but in quiet areasaway from traffic was carried out at three sites.The objective was to determine if any excessdamage was detectable in houses most exposedto traffic vibration.
15
(iii) Acaseby case examination of heritagebuildings was carried out with the assistance ofstructural engineers from the Historic Buildingsand Monuments Commission for England (EnglishHeritage). Buildings showing signs of distress andexposed to relatively high levels of traffic vibrationwere identified and a structural survey was carriedout in order to identify the probable causes of theobserved damage.
(iv) A review was made of both published andunpublished studies dealing with the effects oftraffic vibration on heritage buildings. In somecases an attempt was made to check the likelyaccuracy of the original data and the conclusionsthat had been drawn.
4.3 FATIGUE STUDY
This study, which was conducted under contractby Travers Morgan Planning, involved theexposure of a conventional two storey building tosimulated traffic vibration. The intention was toexpose the house to the equivalent of many yearsof heavy traffic using simulated sources ofairborne anti ground-borne vibration. Throughoutthe exposure period the building was carefullymonitored to determine the precise nature of anydamage or settlement. A full description of the
study can be found in Hood and Marshall (1987)and Watts (1988c).
4.3.1 Site description
The test building was a recently vacated pair ofsemi-detached houses built on medium densitysands. The sand was loose down to a depth ofapproximately 1.5 m and below this level it waslightly cemented. The houses were built at thebeginning of the century and were constructed ofbrickwork in lime mortar. Test foundation stripswere also constructed so the effects of ground-borne vibration on foundations under differentstatic loads could be investigated. The generallayout of the site is shown in Figure 11.
4.3.2 Simulation of vibration
Airborne traffic vibration was simulated using fourlarge loudspeakers mounted in the wall of a highsided refrigerated lorry parked adjacent to thehouse facade (Figure 11). A computer was usedto generate a low frequency waveform whichwhen suitably amplified produced a peak linearsound level of 110 dB at the facade. Initially thesystem was used to generate a broad band noisecharacteristic of heavy goods vehicles but theresulting vibration was lower than expected. Inorder to create higher vibration levels thefrequency was adjusted to produce resonance inthe window adjacent to the loudspeaker system.
Test foundation strips
‘%Arbo::::tensometerssource
Fig. 11 Site layout for fatigue test of a pair of
Test foundationstrips
semi-detached houses
16
-1.0 3.16 10.0 31.6 100 316 1000Frequencv (Hz)
F ig.12 Third octave sound levels in gap betweenspeaker array and facade - house fatigue test
Figure 12 shows that the principal frequencies ofthe simulated sound lay in the 25 Hz third octaveband. By carrying out this adjustment thegenerated vibration levels were at or above thehighest levels likely to be produced by vehiclespassing close to a building. During the experimentit was estimated that the simulation produced anexposure to noise equivalent to the passage ofapproximately 500000 HGVS.
Ground-borne traffic vibration was simulated usinga geophysical vibrator located 2 m from a sidewall of the building (Figure 11 ). Levels wereadjusted so that the vertical PPV at foundationlevel adjacent to the vibrator was in the range2.5–3.0 mm/s. This is close to the extreme end ofthe range of peak velocities that have beenrecorded in buildings close to significant roadsurface irregularities during the passage of heavyvehicles (see Section 3.4.4). The frequency wasadjusted to approximately 13 Hz, which is withinthe range of frequencies produced by HGVS, andthis input produced a relatively large response inthe structure. Figure 13(a) and (b) show the time
Peak velocitv(a) Time historv 2.6 mm/s
3, I
>-3 ~.~o 0.5 1.0 1.5 2.0 2.5
Time (see)
(b) FrequencyspectrumPeakVelocitvat 12.9 Hz
z2-~
o 10 20 30 40 50 60 70 80 90 100Frequencv (Hz)
Fig. 13 Vertical particle velocity at foundationsadjacent to vibrator - house fatigue test
history and frequency content of a typical pulse.The house was exposed to 880000 such pulseswhich was estimated to have simulated the effectof over 3.5 million HGVS axles passing over alarge surface irregularity near the house.
4.3.3 Monitoring techniques
During the exposure period, which lastedapproximately three months, the building andsurrounding soils were carefully monitored usinginstruments capable of resolving movements ofthe order of 0.1 mm/s. Electrolevels andextensometers were employed to measure soilmovements and Ievelling stations were installed at36 locations on the structure to determine anyfoundation settlement. In addition, high resolutionMoire photography was used to indicate ifdifferential movement of the building facade hadoccurred. {l) Forty existing cracks in variouslocations were monitored for movement with aDemec gauge. At various stages throughout theexperiment an inspection of the building was madeand any further cracking was recorded. Vibrationmeasurements were also made throughout thehouse so the response of the house could bedetermined and damage mechanisms identified.
4.3.4 Results
The whole building responded relatively strongly toground-borne vibration and generally the highestlevels were recorded in suspended wooden floorsand ceilings on the first floor. These vibrationswere very noticeable and would have beenunacceptable to most occupants. Acousticallyinduced vibrations were only perceptible in roomsclose to the noise source and were generallybelow those produced by the vibrator despite thehigh levels of low frequency noise produced bythe simulator. It was concluded that the exposureto airborne vibration did not produce anyobservable damage.
Ground-borne vibration produced no detectablesettlement of the house. The accuracy of thelevels were + / – 0.3 mm. It was anticipated thatthe sands under the foundations would densifyunder the action of vibration since the measured insitu density indicated a potential settlement of upto 20 mm was possible. Tests on the soil after theexposure period revealed that it was likely that nodensification had occurred and it was concludedthat the vibration levels being generated wereinsufficient to compact this particular soil futiher.Moire photography showed that, apart from onesmall area close to the vibrator, there was nodifferential movement within the house facade at
(1I The front facade of the building was coveredwith paper on which was printed rows of dots.Movements of the house would produce Moirefringes between pairs of photographs of thepapered walls.
17
any time during the experiment. The amount ofmovement detected in the one area where it wasobservable was only 0.4 mm which is justsignificant since measurement error using thistechnique was estimated to be + / – 0.2 mm.
No structural or trigger damage was found butfatigue damage occured in some plaster finishes.The most significant cracking occurred in the endwall of the house facing the vibrator and inceilings close to the chimneys. The amount ofdamage was very slight and probably would havegone unnoticed in a normally decorated house.The absence of trigger damage may have beendue to the extensive cracking in the plaster whichexisted prior to exposure. Many of these crackswere very large and some allowed movementswithin the plaster in response to the vibration(Watts, 1988d) which may have preventeddamaging stress concentrations. Of the 40 crackswhich were monitored for movement, only fiveshowed significant changes of 0.1 mm or moreduring the exposure period.
Settlements of between 1 and 14 mm occurred inthe test foundation strips. This was not due todensification of the underlying soil but was causedby a migration of soil particles which resulted in asmall amount of rotation of the strips. It wasconsidered that rotation of a facade of the mainbuilding might have occurred if it had been poorlytied to the rest of the structure. [n addition,settlement might have taken place if the layer ofloose sand under the foundations had beendeeper. It is also possible that settlement mighthave occurred if the soil type had been differenteg saturated sand, soft clay or peat. During theexperiment none of these effects did in fact occur.
4.4 PAIRWISE COMPARISONS
The objective of this study was to determine ifexcess damage occurs in occupied houses whichhave been exposed to relatively high levels oftraffic vibration over a considerable period of timewhen com~)ared with similar houses that have notbeen exposed to significant vibration. A fulldescription of the study, which was carried outunder contract by Travers Morgan Planning, canbe found in Muskett and Hood (1989).
4.4.1 Description of sites
Worst case conditions were sought so that theeffects of these vibrations might more readily bedetected. Since differential settlement did notoccur on sands in the simulation test describedabove, it was considered worthwhile toinvestigate the possibility of this effect occurringin soft ground such as saturated alluvial deposits.Therefore suitable rows of houses were sought inareas where there were known to be generallysoft soils. Sites were found in King’s Lynn,Bridgwater and Cardiff and at each site similarrows of houses built on comparable soils butexposed to very different levels of vibration wereidentified.
Table 8 lists the number of light and heavyvehicles at each exposed site. At SaddlebowRoad, in King’s Lynn, the buildings were within10 m of the main road which carried a substantialflow of heavy vehicles. The control site at BeloeCrescent is a CUI de sac and levels of traffic werein consequence very much lower. The houses atboth exposed and control sites form part of acouncil estate and originally were all of verysimilar design. All the houses were constructedbetween 1927 and 1947 and were of traditionalconstruction being two-storey with brick loadbearing walls and with mass concrete stripfoundations.
The houses at the exposed site in Bridgwater(Bristol Road) were approximately 5 m from thecarriageway. This road forms a major radial routeinto Bridgwater from the north and the flow oflorry traffic is relatively high. The traffic at thecontrol site in Devonshire Street was very low andparked cars lined the road. The houses at bothexposed and control sites were similar and theywere built in terraces c 1890. The houses were oftraditional construction comprising brick loadbearing walls.
In Cardiff the exposed buildings in Penarth Roadwere within 10 m of the carriageway whichcarried a very high volume of traffic. The controlproperties were in a CUI de sac (Chester Place).The buildings at control and exposed sites weresimilar, being two-storey terraced properties in
TABLE 8
Paired comparison study–traffic flow at exposed sites between 7:00–1 9:00 hours
Site
Saddlebow Road in King’s LynnBristol Road in BridgwaterPEnarth Ro{~d in Cardiff
Direction
SouthboundSouthboundEastbound
Light vehicles*
31656482
10233
* Light vehicles–cars and goods vehicles <1.5 tonnes.
Heavy vehicles**
432901
1 012
Percentage ofheavy vehicles
12.012.2
9.0
* * Heavy vehicles–goods vehicles >1.5 tonnes, buses and coaches.
18
TABLE 9
Paired comparison study–peak vertical particle velocity at foundations and L1o linear noise level
Site
King’s Lynn
Bridgwater
Cardiff
Saddlebow Road (exposed)Beloe Crescent (control)
Bristol Road (exposed)Devonshire Street (control)
Penarth Road (exposed)Chester Place (control)
Peak vertical velocity(mm/s)
L1o* linear level(dB)
0.96<0.10
1.16<0.08
0.420.25
87.583.5
89.577.6
89.082.0
* LTOlinear level is the unweighed level which is exceeded for 10 per cent of the recording period.
TABLE 10
Paired comparison study–number of houses inspected for damage
Exposed row Control row
External External External ExternalSite and internal only Total and internal only Total ‘
King’s Lynn 6 — 6 3 — 3Bridgwater 6 15 21 1 12 13Cardiff 3 7 10 4 7 11
stone or brick. It was estimated that thesedwellings were built at the turn of the century.
4.4.2 Exposure to vibration and noise
Vibration measurements were made at thefoundations of the properties fronting the heavilytrafficked roads and also at the control sites sothat the peak levels of vibration could beestablished and a check could be made thatsignificant differences in exposure did existbetween the two groups of houses. Noisemeasurements were also made at both exposedand control sites during similar times of the day socomparisons could be made. As expected, bothvibration and noise exposures were very muchgreater at exposed sites than at control sites (seeTable 9). Vibration levels at the exposed siteswere all relatively high, exceeding the level ofperception (0.3 mm/s) at the foundation of thebuildings.
4.4.3 Comparisons of building damage
The comparisons of building damage in exposedand control properties was a difficult taskrequiring skilled professional judgement. Althoughnumerical indices for damage assessment havebeen developed (BRE, 1966) these are normally
only applicable to large scale damage involvingsignificant cracking. A further problem whenmaking assessments in occupied properties is thatminor damage is often repaired and covered byredecoration work. More serious damage occursmore rarely but it is easier to identify especially ifit involves cracking or differential settlement ofexternal brick or stone work.
Table 10 shows the numbers of houses inspectedat exposed and control sites at each of the threesites. In many cases it proved impossible to obtainpermission to carry out internal inspections and sothe survey relied heavily on the comparisons ofexternal defects. To complete the surveys ofdamage, the vertical alignment of the frontfacades of houses was measured with a theodoliteto check if there was evidence of excess tiltingtowards the major roads at exposed sites. Thismight possibly be expected if vibration settlementoccurs since vibration levels will tend to be greaterat the fronts of buildings because these facadesare closer to the source of vibration.
Soil investigations were also carried out todetermine reasons for any possible tilting ofproperties. Trial pits were dug close to selectedproperties at both exposed and control propertiesat the three sites. This allowed the soil types andshear strengths to be determined at and belowfoundation level.
19
4.4.4 Results
Results showed that at all sites there were nosignificant differences in the amount of damage inexposed and control properties. It is possible thatsome minor damage effects mav not have beenfound because of the problems of identifying thislevel of damage in normallv decorated andoccupied homes. There were also considerabledifferences in the conditions of individual houseswithin anv one group of buildings due to thevariation in the level of maintenance. This addedfurther to the difficulties in detecting anv vibrationeffects which might possibly be present.
With the exception of King’s LVnn, there were nosignificant differences in the degree of tiltbetween exposed and control properties. AnalVsisof the results at King’s Lynn revealed that therewas a statistically significant tilt towards the roadat the exposed site (p< 0.02) but that at thecontrol site there was no significant trend in anvdirection. The soil investigations revealed thepresence of verv soft ground at the front of aproperty at the exposed site and enquiriesestablished that this was on the line of an olddrainage ditch. It was further established that theditch had been filled during construction of thehouses. It was concluded that the presence of thesoft ground under the foundations of the exposedfacade was a probable reason for the tilt of thesebuildings towards the road, so it could not beascribed to the effect of traffic vibration.
The buildings examined were tvpical urban terracedwellings built on relatively soft ground near theturn of the century and were probably exposed togenerallv high levels of traffic vibration over manvVears. Although these conditions can reasonablebe considered to be conducive to inducingvibration effects there was no evidence thatdamage had been caused bv heavv traffic.
4.5 HERITAGE BUILDINGS
In order to assess the possible contribution oftraffic vibration to damage in older properties,eight buildings showing signs of distress andexposed to relatively high traffic vibrations wereidentified and examined. Vibration, noise andcrack movements were monitored and soilconditions at each site were examined and trafficflow levels recorded. English Heritage, which hasresponsibilities for historic buildings andmonuments of national importance, collaborated inthis research bv carrving out structural survevs ofthe’ buildings and bv providing written reports ofthe observations. The main objectives of thisstudy were: —
(i)
(ii)
(iii)
to quantifv the vibration exposure and theresponse of the buildings.
to determine the condition of the buildingsand identifv damage.
to attempt to determine the main causes ofanv observed damage.
The studv was carried out in two phases. In thefirst phase four brick built grade 2 listed buildingswere examined (Watts, 1988b) while a muchwider range of buildings in terms of age, size, andtype of construction was studied in the secondphase (Watts, 1989b).
4.5.1 Description of buildings
The sites were selected to represent the ‘worstcase’ conditions (ie where it was considered thatbv virtue of the combination of high vibrationlevels, soil and building conditions, there wassome potential risk of vibration damage occurring).Sites were found within a few metres of roadscarrying HGV traffic and generally this was at ahigh level producing perceptible ground-bornevibrations at foundation level. Soil conditionsincluded wind blown sand deposits and soft soilssuch as saturated peat and alluvium since it wasconsidered that these soils had the greatestpotential to cause settlement as soil particlesbeneath the foundations could possiblv densifv ormigrate when vibrated.
Table 11 provides descriptions of these buildingsand includes details of approximate age andsignificant alterations that had been made to thestructure of the buildings. There was a wide rangein the sizes, ages and types of construction ofthese buildings. The oldest, and also the largest,was the 15th centurv parish church in the centreof Louth in Lincolnshire. The building was over60 m long and the spire is one of the tallest in thecountrv. The most recent building studied was alarge house originally built as an inn at thebeginning of the century but which has since beenconverted to a farmhouse. At this site it waspossible to investigate the likelihood of triggerdamage since during the study the Iorrv trafficincreased significantly over a short period of timedue to the opening of a gas pipeline store within100 m of the propertv. BV studying the buildingbefore and after the increase in Iorrv traffic anynew damage resulting from traffic vibration couldbe identified.
4.5.2 Measurement and survey methods
To determine the exposure of the buildings tovibration, peak amplitude and rms values ofparticle velocitv and acceleration were measuredat various parts of the building. Crack movementswere also recorded to supplement this information.Noise measurements were made at the facades of
20
TABLE 11
Case studies of heritage buildings–description of buildings
Sitenumber
1
2
3
4
5
6
7
8
Description
Shrimpers cottage interrace
Detached house
Cottage
Large house
Georgian town house
Large parish church
Cottage
Large farmhouse
Numberof stories
2
2
213
2
4
—
2
2
Construction
Brick
Brick
Brick
Brick
Stone
Stone
Timber-framed withbrick extension
Stone and brick
buildings to quantify the level of low frequencynoise. In addition surveys were made of the trafficand soil conditions at each site. Table 12 givesdetails of the site conditions including the flows oflight and heavy vehicles and the peak verticalparticle velocities at the most exposed foundationsrecorded during the day. The farmhouse (site 8)was examined before and after the large increasein lorry traffic and it can be seen that there is acorresponding increase in the peak particlevelocity recorded in the after period.
Building surveys were carried out by structuralengineers at the same time or shortly aftermeasurements were made. The engineersidentified damage both internally and externallyand recorded defects on building plans. Recordeddamage included cracks in plaster finishes, brickand stone work and distortions of walls andceilings. Significant cracks were monitored using aDemec gauge. The structural engineers had a wideexperience of common types of damage in otherbuildings of similar type to the ones examined inthis study and this provided a useful referencewhen they assessed the likely effects of trafficvibration.
Significantalterations
None
None
Ground floorconverted to shop
Ground ftoorconverted to publicbar
Ground fkorconverted to shop
Spire added in 16th C
Porch and bell toteadded in 19th C
Mid 20th C extension
Built as an innconverted todwelling
Approximateage
Late 18th C
Mid 19th C
Early 19th C
Early 18th C
Mid 18th C
Early 15th C
15th C
Early 20th C
4.5.3 Results
It was found that ground-borne vibrationgenerated by passing HGVS was the mostsignificant source of vibration at all sites.However, it was demonstrated that other activitiessuch as stamming doors, jumping on upper floors,and in the case of the church playing the organ,could produce similar or greater levels of vibration,although the frequency of such events wouldnormally be much lower than the number ofvibration events produced by traffic. Peak levels ofvertical particle velocity at foundations wereabove the level of perception established by Reiherand Meister of 0.3 mm/s (see Section 3.6) at attsites except the farmhouse. Maximum vibrationamplitudes were greater on upper ftoors and wattsat the fronts of the buildings than at foundationlevel. The damage surveys identified a range ofdefects in the buildings ranging from cracks inplaster finishes to more substantial structuraldamage resulting from foundation settlement. Briefdetails of some of the more significant casestudies are given below. In all cases it wasconcluded that the main causes of the damageobserved was not traffic vibration.
21
TABLE 12
Case studies of heritage buildings–description of site conditions
Sitenumber
1
2
3
4
5
6
7
8
Location
Botanic Road,Southport
Monmouth Street(A38) Bridgwater
Wisbech Road(A47) Thorney
Double StreetSpalding
Widcombe Parade(A36) Bath
Upgate (Al 6)Louth
A435 Norton
C-road nr Honiton
Soil description
Blown sand
Alluvium
Gravelly sand overlvingpeat
Alluvium
SiltV sand, soft clayIaVers beneath
Sand, gravel and Marlclav
Sand, gravel and clay
SiltV clav overlving softrock
Total flow between10–19:00
Light *
4288
11 916
4008
2480
8472
4572
5664
HeavV**
* Light–cars and goods vehicles < 1.5 tonnes.* * Heavy–goods vehicles a 1.5 tonnes, buses and coaches.
Shrimpers cottage in Southport, MerseVside
The side wall of this building formed one side of awaggon porch which led to the rear of theproperty and it was severelv cracked (see Plate 1and 2). Close to this wall ran an old sewer pipeand it was thought that the trench above mayhave been poorlv backfilled and so causedsettlement of the foundations resulting in thestructural cracks. A tapered brick course near thefoundations indicated that this distortion of thebuilding had occurred manv Vears ago, so it wasnot related to the effects of recent heavy traffic.
349
1 600
1 192
548
980
652
1 212
Before pipe store opened372 I 48
After pipe store opened593 118
Cottage at ThorneV, Cambridgeshire
This building was generallv in a poor conditionhaving very marked distortions in floors and walls.The low wall fronting the A47 was leaningtowards the road and was out of plumb bv170 mm at one point (see Plate 3). This was
Peak verticalparticle velocity
recorded between10–19:00
(mm/s)
0.92
0.80
3.52
1.37
0.92
0.33
0.46
0.11
0.16
thought to be due to a deep layer of peat underthe footings shrinking over the Vears and causingthe settlement problems. That this process hadoccurred was indicated bv the fact that settlementhad taken place in buildings in the same terracenot exposed to high levels of traffic vibration.Another reason for the poor state of the buildingwas the fact that structural supports had beenremoved to provide an open plan area on theground floor. This is likely to have weakened thestructure considerable and produced over a periodof time some of the distortions of brickwork andfloors on upper levels in the building. It is possiblethat traffic vibrations had accelerated the naturalprocess of settlement but there is no evidencethat this is the case.
Parish church at Louth, Lincolnshire
There was much concern that traffic vibration wascausing damage to stained glass windows at the
22
Plate 1 Entrance to Waggon Porch in Shrimpers Plate 2 Side wall of Waggon Porch showingcottage showing vertical crack extensive cracking
Plate 3 Out of plumb wall adjacent to A47 atThorney
23
Plate4 15th century parish church adjacent to Plate 5 15th century timber framed buildingA16 at Louth adjacent to A435 at Norton near Evesham
east end of this large building since somewindows were heard to buzz and rattle as vehicles
passed by. This anxiety was understandable since
this end of the church is within two metres of the
heavily trafficked Al 6 (Plate 4). Since the churchwas very large it was feasible to comparewindows near the west end, which were notexposed to such high vibration levels, with thewindows at the east end. Detailed examinationshowed that there were no differences in thecondition of the windows which could reasonablybe explained by exposures to traffic vibration.
Timber-framed cottage near Evesham,Warwickshire
This cottage was situated close to the heavilytrafficked A435 and Plate 5 shows a pronounceddistortion of the gable wall adjacent to thecarriage way. Distortions of this type are notuncommon in this type of building and a similarbuilding was found in Evesham which had evenmore severe distortion and yet was well removedfrom any main road. The cottage can be identifiedin a photograph published in a touring guide in1954 and the distortion can be seen, although it isnot possible to quantify the degree of tilt with anydegree of accuracy. However it is evident that thedistortion is not new and cannot plausibly beattributed to modern heavy traffic.
In these and all other cases it was concluded thatthe main causes of the damage observed werelikely to have been site factors rather thanexposure to traffic vibration.
4.6 REVIEW OF DATA ON VIBRATION
DAMAGE IN HERITAGE BUILDINGS
This study was carried out under contract by BrianMorton and Partners. They approached over fivehundred individuals and organisations throughoutthe world for relevant information on damage toheritage buildings that could possibly be attributedto exposure to traffic vibration. These sourcesincluded architects and surveyors who haveresponsibilities for cathedrals and churches,professional engineers and a range oforganisations such as civic amenity societies whopromote the conservation of these buildings. Theinformation collected consisted mainly ofbibliographic references and case histories ofbuildings of national importance. Replies werereceived to many of the letters and, in the caseswhere records were available and measurementshad been taken, in-depth assessments were made.This involved visits to selected sites and wherepossible discussions were held with those whohad been involved in damage assessment andmeasurement. In the case of Lincoln cathedral, the
24
data available was substantial since a number ofindependent measurements of both vibration andsettlement had been made during a period of over20 years. The evidence for the assertion that oldcathedrals and churches close to heavily traffickedroads show signs of excessive movement whichcan be attributed to the action of traffic vibration(Crockett, 1973) was also examined.
4.6.1 Results
From the very large amount of informationreceived there were only a relatively small numberof studies where damage, including excess tilting,was claimed to have been definitely caused byexposure to traffic vibration. From the availableevidence these claims were not substantiatedexcept in two cases. The first case is the fatiguestudy described in Section 4.3 where hairlinecracking of plaster resulted from high levels ofsimulated traffic vibration. In the second case,damage to the flashings of the roof of TowerBridge, London, was considered to have beencaused by heavy vehicles using the bridge. It islikely that resonances were excited in the flexiblebridge structure which resulted in high amplitudevibrations in the bridge towers. This mechanism isplausible but it is clear that the conditions that ledto this damage are very unusual and that thiseffect could not occur adjacent to normal roads.There were many studies where vibration wasthought to be a possible cause of damage eitherdirectly or in combination with other factors butno evidence was available to confirm or denythese impressions. Wherever evidence wasavailable it suggested that traffic vibration was notthe main cause of the damage observed.
4.7 DISCUSSION OF VIBRATION
DAMAGE STUDIES
This report has described a number of differentstudies which attempted to determine the possibleeffects of traffic vibration on buildings. Thefatigue test on a pair of semi-detached houses,where carefully controlled simulated trafficvibrations were generated, provided firm evidencethat the direct effects of traffic vibration on abuilding in generally good condition is likely to bevery small indeed. Despite the high levels ofvibration, which would have proved intolerable tomost occupants, there was only a small degree ofdamage. This consisted of hairline cracking ofsome plaster finishes which would probably havegone unnoticed in a normally decorated house. Itwas concluded that ground-borne vibration andnot airborne vibration was responsible for thisdamage.
Some indirect effects of vibration were alsoexpected since the building was on a mediumdensity sand and it was thought that the ground-
borne vibration would cause some densification ofthe sands under the foundations. The building wasvery carefully monitored for any movements andyet despite this there was no evidence ofsettlement occurring. However the individual testfoundation strips did settle by varying amounts upto 14 mm due to soil migration: these stripsmoved by rotation and so there is a possibilitythat had the facade nearest the vibrator been lesswell tied to the rest of the building, or the sandhad not been cemented at 1 m below thefoundations, some movement might have takenplace.
The pairwise comparison of buildings exposed torelatively high and low levels of traffic vibrationwere made at sites on generally soft alluvial typesoils where it was considered that conditions weresuitable for observing the effects of trafficvibration. The dwellings were built near the turn ofthe century and so it is likely that the buildingshad been exposed to substantial traffic vibrationover a period of many years. There was noevidence of excess structural damage at theexposed sites and a check on the verticalalignments of the front facades showed nosignificant differences except at the King’s Lynnsite. The exposed buildings at this site, unlike thecontrol buildings, were tilting significantly towardsthe main road but this was considered to be dueto the presence of an old badly filled drainageditch at the front of the properties. This containedvery soft soil and fill material and was thought tobe the most probable cause of the foundationmovements.
The case studies of heritage buildings did notproduce any firm evidence that the buildingdefects had been caused by traffic vibration. Itwas more likely that poor ground conditions and illadvised alterations were the major causes ofsignificant damage in many cases. In addition, thereview of the available information from sourcesworldwide did not produce any substantiveevidence of damage to buildings caused by trafficrunning on normal roads.
In the studies described considerable efforts weremade to examine buildings under ‘worst case’conditions. These have failed to show anysignificant effect of traffic vibration on ordinarydomestic dwellings or heritage buildings. Thus theevidence does not support the assertion thattraffic is responsible for major damage. However,at sites exposed to very high levels of ground-borne vibration for a substantial period, someminor damage to plaster finishes could occur. Therisk of damage would obviously be greater if theplaster work was in a fragile condition. Additionally,there may possibly be soil conditions that could besusceptible to settlement produced by high levelsof traffic-induced ground-borne vibration. A furtherpossibility is that traffic vibration could exacerbate
25
damage effects due to other causes. In all thestudies no evidence of these effects were found inbuildings, but some newly constructed foundationstrips on uncompacted sand did settle up to14 mm.
5 METHODS TO AMELIORATETHE EFFECTS OF TRAFFICVIBRATION
There are a number of strategies that can beadopted to reduce the nuisance caused bv trafficvibration. The type of action will depend onwhether the problems are produced bv airborne orground-borne vibration. A reasonable initialapproach would be to establish bv simpleobservation, prediction or bv suitablemeasurements, the contribution from each source.A site visit during a period when HGVS or busesare frequently passing the site is most useful. If itis established that the main manifestations ofvibration are windows or doors that rattle andbuzz in rooms fronting the main road then it isIikelv that airborne vibration is a problem. Incontrast, ground-borne vibration often produced ashort duration impulsive vibration which is mostreadilv detected in the middle of upper floors. Thisshould be distinguished from a longer duration,higher frequencv, trembling of the floor which mavsometimes be produced bv high levels of lowfrequencv noise. A ground-borne vibration problemis most acute where the building is within a fewmetres of a significant road surface irregularitysuch as a poorly backfilled trench or sunkencover. Section 3.2.3 describes a method forpredicting the Iikelv level of nuisance due toairborne vibration and Section 3.4.4 develops apredictive equation which determines whetherground-borne vibrations are Iikelv to be intrusive.Measurements of peak particle veiocitv at thefront foundations will indicate that ground-bornevibration is a problem if levels are significantlyabove 0.3 mm/s and the dominant frequencv is inthe range 8–25 Hz. Possible methods to reducethe nuisance are discussed below, grouped bvtvpe of vibration.
5.1 AIRBORNE VIBRATION
Low frequencv noise readilv affects light, flexiblestructures such as doors and windows particularlyif thev are loose fitting and therefore have adegree of freedom of movement. Some older sashwindows are particularly prone to rattle and buzz.In cases where this appears to be the main sourceof annovance, an obvious solution is to wedge theoffending frame or glass to prevent movement.However it can sometimes prove hard to locatethe exact source of this parasitic noise. This isparticularly difficult when the noise occurs onlv
occasionally. Opening windows in modern thermaldouble glazed units are bv comparison usuallv wellfitting and should be less prone to producingannoving vibrations. Low frequencv noise isattenuated to a greater extent when passingthrough such double panes of glass and anadditional reduction of 3 to 5 dB can be expectedfrom these replacement windows. Doublewindows fitted as part of the remedial packageprovided under the Noise Insulation Regulationsare also Iikelv to reduce nuisance although the lowfrequencv noise reduction is generallv less. Thesereductions are clearlv not large and can obviouslvbe degraded if windows are opened. It should benoted that low frequencv noise can be perceiveddirectlv and can sometimes lead to annoyingmuffled sensations in the ears and perceptiblechest vibrations.
Figure 14 shows the trend in the percentage ofresidents ‘bothered bv vibration’ with noiseexposure level. The TRRL data was taken as partof the 50 site survev where most dwellings hadsingle glazing. This data is compared with resultsfrom a Building Research Establishment survev(UtleV et al, 1986) where all dwellings were fittedwith double windows which had been installed aspart of the insulation package available under theNoise Insulation Regulations. The TRRL survev
80I I I I I I*** I I
.+.**
.40
.“””/’/+” A-
,<..
/ .“A
,/ ..*
# .**●** A A
.**
i
I I 1 1 I I I I 1 I68 70 72 74 76 78 80 82
External noise level Llo (18 hour) dB(A)
— BR E data. Houses along dual carriagewaysfitted with double windowsY=I.9X - 105.6; r=O.83
==--= --= TRRL data, Houses along dual carriageways,mainlv single glazing: Y=4.71X - 287.4; r=O.90
‘-- TRRL data. Houses at all sites, mainlv singleglazing: Y=2.36X - 116.0; r=O.64
Fig. 14 Percentage of respondents bothered byvibrationSource: Utley et al 1988
26
included a wider range of road types than the BREsurvey and therefore the results from dual-carriageway sites, which are likely to be similar tothe BRE sites, have been considered separately.Whichever of the sets of data from the TRRLsurvey is used, there appears to be a considerablereduction in the number of people bothered byvibration in dwellings where the insulation packagehas been installed. At a L1o ( 18-hour) dB(A) levelof between 72 and 74 dB(A) at the facade, thepercentages bothered reduces from approximately55 to 30 per cent.
Two main approaches can be adopted to reducethe problem at source. Firstly, it may be possibleto reduce the number of heavy vehicles atproblem sites by various traffic managementschemes. Section 3.3.2 describes an equationwhich allowed the prediction of the medianvibration nuisance score from the number of HGVSpassing the site in an 18 hour day and thedistance from the front facade of the building tothe edge of the carriageway. The dependence onlorry flow is logarithmic so a substantial reductionin flow would be needed to effect a noticeableimprovement. Secondly, a long term approach, butone which would produce general improvements,is to reduce the low frequency emissions fromvehicles. Such controls would, however, requirethe development of new test procedures and,additionally, the setting of limit values wouldrequire new regulations to be developed andagreed internationally. It should be noted thatnoise barriers designed to screen traffic noise maybe of limited use in reducing airborne vibrationproblems since there is typically little attenuationof the low frequency sound which is responsiblefor perceptible vibration effects.
5.2 GROUND-BORNE VIBRATION
There are a number of approaches that can beadopted to reduce the exposure to ground-bornetraffic vibration. In this case it is easier to makereductions at source rather than attempt toattenuate the transmission of these vibrations intobuildings.
5.2.1 Reductions at source
An obvious remedy is to ensure that a smoothroad surface is maintained where dwellings andsensitive work areas are close to the road sinceirregularities of the order of 20 mm in the surfaceprofile can produce perceptible vibrations inbuildings located within a few metres of thecarriageway. On soft soils such as peat andalluvium there is a greater need to ensure asmooth surface because of the greater responseof the ground.
Generally, peak particle velocities increase withspeed for all vehicles and therefore there should
be some advantage in reducing the maximumpermissible speeds of HGVS past sensitive sites.An estimate of the likely reduction can beobtained from the trend with speed shown in thepredictive equation developed in Section 3.4.4.Reducing the maximum speed of these vehiclesfrom say 80 km/h to 48 km/h should decreasepeak particle velocity by 40 per cent. This maybring substantial relief if the resulting peakvelocities fall below the perception threshold.
Decreasing the load carried on a particular HGVdoes not necessarily reduce the peak particlevelocity and in some cases an empty lorry canproduce higher levels than when fully laden(Watts, 1988a). However, smaller vehicles dotend to induce smaller vibrations as can be seen inFigure 7 which shows a clear trend of rising PPVwith increases in gross vehicle weight. A typicalgross vehicle weight restriction on public roads is7.5 tonnes, and provided maximum speeds do notrise when the weight limit is reduced, theexpected peak velocity, based on the regressionline, would be reduced substantially.
A measure that may have benefits in the longterm is the design and regulation of HGVsuspensions to reduce the generation of vibration.Tests on HGV vehicle suspension systems haveshown that different systems loaded to similaraxle weights produce some differences in the peaklevels of vibration (Watts, 1988a) and so thereappears to be some scope for improvement.
5.2.2 Attenuation methods
One possible technique to control ground vibrationis to construct a trench between the source andaffected buildings. The trench acts as a barrier tovibration and can therefore reduce transmissionthrough the soil. To obtain maximum performancethe impedance of the fill material used in thetrench should be as low as possible so energy isreflected rather than transmitted towards theaffected building. A possible fill material isexpanded polystyrene which has a low impedance,is inexpensive and is also strong enough towithstand soil pressures within the trench.However, in a study by Hood and Marshall (1 987)a polystyrene filled 3 m deep trench produced onlya 15 per cent reduction in vibration at a distanceof approximately 40 m from the trench, althoughhigher reductions were recorded close to thebarrier. Other experiments, involving trenches andsheet piling, have shown varying degrees ofsuccess (Barkan, 1966; Richard et al, 1970; Liu etal, 1974). Since surface ground waves producesignificant disturbances down to about a third of awavelength, which for some ground waves maybe greater than 10 m, it is clear that to achievesignificant attenuation of low frequencies verydeep trenches would probably be needed. Theseresults suggest that it is unlikely that such a
27
technique would be commercially viable forscreening conventional dwellings from trafficvibration. The technique may, however, offer amore attractive solution at sensitive locationswhere other forms of building isolation prove to beeither too expensive or fail to achieve anacceptable degree of control.
A further method of attenuation is to isolate theaffected bllilding by decreasing the naturalfrequency of the building to below that of thevibration source. In this way the transmission ofthe vibration into the building can be reduced. Thishas apparently been carried out successfully byintroducing rubber mounts into the foundations
(Crockett, 1985; Grootenhuis, 1979). Again this isan expensive solution for existing buildings andcould probt]bly only be justified for very sensitivelocations. Costs would probably be much lowerwhere the system formed part of the design of anew building.
If the main problem is the detrimental effect ofvibration on the performance of sensitiveequipment it should be possible to provide localpassive isolation by mounting the apparatus onsuitable isolators. Isolation systems have utilised arange of materials and techniques including cork,rubber, helical springs and pneumatic devices(Ferahian arid ‘Ward, 1970). Measurements haveshown that in domestic buildings at least,vibrations on ground and basement floors tend tobe significantly lower than on suspended floors athigher levels in the building and therefore it maybe possible to reduce the problem simply byrelocating the equipment to these lower floors.
6 CONCLUSIONS
This report describes a number of studies of theeffects of traffic-induced vibrations on people andbuildings. Methods are described which allow theprediction of nuisance from airborne and ground-borne vibration and possible methods to amelioratethe environmental impact. The results should alsobe of assistance in the planning process in thatproper weight can be given to the likelyenvironmental impact of traffic vibrations onbuildings. The main conclusions are as follows: –
6.1 VIBRATION LEVELS
1. Overall, fewer people are bothered by vibrationfrom traffic than by traffic noise. However, theproportion of residents seriously bothered byvibration (8Yo) is similar to the percentageseriously bothered by noise (9Y0).
2. A majority of residents interviewed in a surveyon traffic vibration said they were bothered by
vibration because they thought it could causedamage to their properties.
3. Where vibration nuisance is caused mainly byairborne vibration, it is the low frequency contentof the noise which causes the problem.Nevertheless, standard acoustic indices, whichcover the whole spectrum of noise, were found tobe significantly correlated to the average level ofvibration nuisance at residential sites. Theseindices were generally better predictors ofdisturbance than were measures of windowvibration, traffic flow or road roughness. Howevera composite measure of heavy vehicle flow anddistance of the affected building from thecarriageway was as good a predictor of vibrationnuisance as was the best acoustic measure. TheL1o (18 hour) dB(A). index was among the bestcorrelated acoustic measures and since it is inwidespread use it would be suitable for predictionpurposes.
4. Ground-borne vibration affects only a smallproportion of residents. However, the peak levelsof these vibrations at building foundations can berelatively high, especially where the underlying soilis soft and houses are close to significant roadsurface irregularities. Peak levels up to ten timesthe level of perception have been recorded and inthese situations serious nuisance and anxietiesabout building damage are likely to arise.
5. A prediction equation has been developedwhich enables the likely maximum peak verticalparticle velocity at the foundations due to ground-borne vibration to be estimated. The parameters ofimportance include the maximum speed of HGVS,road surface profile, and ground conditions. Thetype of underlying soil was found to greatlyinfluence the level of vibration and the largestresponse was recorded on soft soils such asalluvium and peat deposits.
6. Sensitive equipment and critical work areascan be affected by very low levels of vibrationclose to the level of perception. General guidanceon tolerable levels for satisfactory equipmentperformance is of limited value since theinstrument design and mounting conditions arecritically important in determining the degradationof performance under vibration. Manufacturersmay sometimes provide guidance on the maximumpermissible levels of vibration.
6.2 EFFECTS ON BUILDINGS
A number of studies have been carried out todetermine the possible effects of traffic vibrationon a range of building types. In addition, aworldwide search of possible sources ofinformation has been made. The conclusions ofthese studies are as follows: —
28
7. In a fatigue damage study where a recentlyvacated house was exposed to relatively highsimulated airborne and ground-borne trafficvibration over a prolonged period only non-structural damage was found to have occurred.This was caused by ground-borne vibration andwas limited to a small amount of fine plastercracking. It is unlikely that this damage wouldhave been recognized in a normally decoratedhouse. There was no evidence that exposure toairborne vibration had caused even minor damage.There was no settlement of the house foundationsdue to the action of vibration. Some movementhad been expected since the test house was builton medium density sand and there was thepotential for some densification of soils beneaththe foundations.
8. In studies of occupied buildings on relativelysoft soils, where the degree of damage wascompared in groups of similar houses adjacent toand remote from heavily trafficked roads, it wasfound that there was no significant difference inthe condition of the two groups of buildings. Thiswas despite the fact that ‘worst case’ conditionscould reasonably be considered to have beenstudied.
9. Case studies of eight heritage buildings ofwidely different ages, size, and type ofconstruction exposed to relatively high levels oftraffic vibration revealed that there was noevidence that traffic vibration has caused theobserved damage. The defects could moreplausibly be explained by site factors other thantraffic vibration.
10. A worldwide search for sources of relevantinformation on vibration damage in heritagebuildings did not reveal any evidence that damagehad been caused by exposure to traffic vibration.
These findings on traffic vibration damagetherefore lead clearly to the overall conclusionsthat there is no evidence to support the assertionthat traffic vibration has a significant damagingeffect on buildings. There is evidence that a smallamount of superficial damage could be producedby sustained exposure to very high levels ofground-borne vibration, but it is likely that actionwould be taken to limit vibration levels if thesecircumstances ever arose in domestic propertiessince the level of nuisance would probably beintolerable to most occupants.
6.3 ALLEVIATION OF TRAFFIC
VIBRATION
Methods to ameliorate the effects of trafficvibration depend on whether the problem is largelydue to airborne or ground-borne vibration. Thefollowing conclusions can be drawn from thevarious studies: —
11. Where the vibration is largely airborne,window rattle can give rise to nuisance and betterfitting windows may improve the situation. Doublewindows and double glazed windows are likely toreduce vibration nuisance. A reduction in thenumber of HGVS passing the site is likely todecrease the level of disturbance, but, because ofthe logarithmic dependence of average nuisanceon lorry flow, the reduction would have to belarge to produce a significant decrease indisturbance.
12. If ground-borne vibrations are the majorconcern, then there are a number of remedialmeasures that can be taken. The simplestapproach would be to reduce the problem atsource. This can be achieved, for example, byremoving significant surface irregularities in theroad surface near the affected properties. Otherapproaches are to reduce the speed of HGVS nearthe properties, re-route the HGVS to less sensitiveroads and introduce a limit on gross vehicleweights of about 7.5 tonnes. The attenuation ofground-borne vibrations by filled trenches and theisolation of buildings by resilient mounts arefurther conceivable solutions, but these are likelyto be expensive. If sensitive equipment isadversely affected then it should be possible toprovide local passive isolation by the provision ofsuitable mounts that will substantially reduce theeffects of vibration.
7
The
ACKNOWLEDGEMENTS
work described in this report was carried outin the Vehicles and Environment Division ofVehicles Group of TRRL, and by Travers MorganPlanning and Brian Morton and Partners undercontract to TRRL.
8 REFERENCES
ARM ITAGE A (1980). Reports on the enquiry intolorries, people and the environment, HMSO.London.
BARKAN D D (1 962). Dynamics of bases andfoundations. McGraw-Hill. London.
BATA M (1971). Effects on buildings of vibrationscaused by traffic. Building Science, Vol. 1,pp 221 –246. Pergamon press, London.
BRITISH STANDARDS INSTITUTION (1 984).Evaluation of human exposure to vibration inbuildings ( 1 Hz to 80 Hz). British Standard6S 6472. London.
29
BUILDING RESEARCH STATION (1 966). Crackingin buildings. BRE Digest 75. Building ResearchStation, Garston, Hertfordshire.
CIVIC TRUST (1 970). Heavy Lorries. Civic Trust,London.
CROCKETT J H A (1 966). Some practical aspectsof vibration in civil engineering. pp 267–272.Butterworths, London.
CROCKETT J H A (1973). The effects of trafficon older buildings. Paper presented at the PTRCcourse on Environmental Factors in Road Design4–8th June. Planning and Transport Research andComputation Ltd, London.
CROCKETT J H A (1985). Soil vibration insulationfor buildings and structures. Plastics and RubberProcessing and Applications, VOI 5,3,pp 267–275.
DEPARTMENT OF TRANSPORT and WELSHOFFICE (1988). Calculation of road traffic noise,HMSO, London.
FERAHIAN R H and WARD H S (1 970). Vibrationenvironment in laboratory buildings. Division ofBuilding Research, Technical Paper 329. NationalResearch Council of Canada, Ottawa.
FIELDS J M and HALL F L (1 987). Communityeffects of noise. Transportation Noise ReferenceBook (Ed. P. Nelson), Butterworths, London.
GROOTENHUIS P (1979). Resilient mounting oflarge structures. Chapter 13 in Elastomers: criteriafor engineering design. Applied Science Publishers,Barking, Essex.
HOLMBERG R, EKMAN G and SANDSTROM H(1983). Comments on present criteria for vibrationsensitive electronic equipment. Report DS 1983:3.Swedish Detonic Research Foundation. Stockholm.
HOOD R A and MARSHALL C P (1 987). Theeffects of simulated traffic vibration on a dwellinghouse. Department of Transport TRRL ReportCR44: Transport and Road Research Laboratory,Crowthorne.
HOTHERSALL D C, CHANDLER-WILDE S N andHAJMIRZAE M N (1 989). Numerical modelling ofnoise propagation over barriers. Paper presented atAcoustics ’89 at the University of Oxford. Inproceedings of the Institute of Acoustics VOI 2,part 5, pp 207–213, Edinburgh.
HOUSE M E (1973). Traffic-induced vibrations inbuildings. The Highway Engineer, February 1973,The Institution of Highway Engineers, London.
INSTRUMENT SOCIETY OF AMERICA (1 975).Recommended environments for standardslaboratories. Report ISA-RP52. 1, Research TrianglePark, North Carolina.
JORDAN P G and YOUNG J C (1 980).Developments in the calibration and use of thebump-integrator for ride assessment. Departmentof Transport TRRL Report SR 604. Transport andRoad Research Laboratory, Crowthorne.
LINGER D A (1963). Effects of vibration on soilproperties. Highway Research Record no 22.Highway Research Board, Washington DC.
LIU T K, KINNER E B and YEGIAN M K (1974).Ground vibrations. Sound and Vibration, Vol. 8,no. 10, pp 26–32.
MARTIN D J (1978). Low frequency traffic noiseand building vibration. Department of TransportTRRL Report SR429: Transport and Road ResearchLaboratory, Crowthorne.
MARTIN D J, NELSON P M and HILL R C (1978).Measurement and analysis of traffic inducedvibration in buildings. Department of TransportTRRL Report SR402: Transport and Road ResearchLaboratory, Crowthorne.
MORTON-WILLIAMS J, HEDGES B andFERNANDO E (1978). Road traffic and theenvironment. SCPR Report P.390. Social andCommunity Planning Research. London.
MUSKETT C and HOOD R (1 989). Vibrationdamage to occupied buildings. Department ofTransport Contractors Report: Transport and RoadResearch Laboratory, Crowthorne. –to bepublished;
NELSON P M and WATTS G R (1 988). Trafficvibration and building damage. In Traffic vibrationand building damage —TRRL papers presented atAcoustics ‘87–Ed G R Watts. Transport and RoadResearch Laboratory, Crowthorne.
NEW B M (1986). Ground vibration caused by civilengineering works. Department of Transport TRRLReport RR 53. Transport and Road ResearchLaboratory, Crowthorne.
NIE N, HULL C, JENKINS J, STEINBRENNER K andBENT D (1975). Statistical package for the socialsciences (SPSS). Chapter 20. McGraw-Hill. NewYork.
PARSONS K C and GRIFFIN M J (1 988). Wholebody vibration perception thresholds. Journal ofSound and Vibration. VOI 121, no 1, pp 237–258.
30
RICHARD R E, HALL J R and WOODS R D (1 970).Vibration of soils and foundations. Prentice-HallInc. Englewood Cliffs, New Jersey.
RUDDER F F (1978). Engineering guidelines for theanalysis of traffic-induced vibration. FederalHighway Administration Report FHWA-RD-78-I 66.Federal Highway Administration, Washington DC.
SORENSEN A and MAYVEN M (1 982). TheDynatest 8000 falling weight deflectometer testsystem. Symposium on bearing capacity of roadsand airfields. Tapir, Trondheim.
STEFFENS R J (1 974). Structural vibration anddamage. PI 5. Building Research EstablishmentReport. HMSO.
UTLEY W A, BULLER I B, KEIGHLEY E C andSARGENT J W (1986). The effectiveness andacceptability of measures for insulating dwellingsagainst traffic noise. Journal of Sound andVibration. VOI 109, no 1, pp 1–18.
WATTS G R (1 984). Vibration nuisance from roadtraffic –results of a 50 site survey. Department ofTransport TRRL Report LR 1119: Transport andRoad Research Laboratory, Crowthorne.
WATTS G R (1 985a). Subjective responses totraffic noise and noise induced vibration effects indwellings. Paper presented at Internoise ’85 inMunich. Proceedings of Internoise ’85. Availablefrom International Institute of Noise ControlEngineering, Dortmund, West Germany.
WATTS G R (1985b). Traffic induced vibrations indwellings–results of a jury experiment. Paperpresented at Acoustics ’85 at the University ofYork. Proceedings of the Institute of Acoustics,VOI 17, part 2, pp 398–407. Edinburgh.
WATTS G R (1 987). Traffic-induced ground-bornevibrations in dwellings. Deparrmenr of TransportTRRL Reporr RR1 02. Transport and Road ResearchLaboratory, Crowthorne.
WATTS G R (1 988a). Ground-borne vibrationsgenerated by HGVS –effects of speed, load, androad profile. Paper presented at Acoustics ’88 atthe University of Cambridge. Proceedings of theInstitute of Acoustics, VOI 10, part 2,pp 541–545. Edinburgh.
WATTS G R (1988b). Case studies of the effectsof traffic induced vibrations on heritage buildings.Deparrmenr of Transporr TRRL Reporr RR 156.Transport and Road Research Laboratory,Crowthorne.
WATTS G R ( 1988c). Traffic vibration andbuilding damage –TRRL papers presented atAcoustics ’87. Department of Transport TRRLReport RR I 46: Transport and Road ResearchLaboratory, Crowthorne.
WATTS G R (1988d). Response of buildingelements to simulated traffic vibration. Paper 4 inDeparrmenr of Transporr TRRL Report RR 146:Transport and Road Research Laboratory,Crowthorne.
WATTS G R (1 989a). Ground-borne vibrationsgenerated by HGVS –effects of vehicle, road andground parameters. Paper presented at Acoustics’89 at the University of Oxford. In Proceedings ofthe Institute of Acoustics, VOI 11, part 5,pp 99–1 09. Edinburgh.
WATTS G R (1989b). The effects of trafficvibration on heritage buildings—further casestudies. Depatimenr of Transport TRRL ReporrRR207: Transport and Road Research Laboratory,Crowthorne.
WHIFFIN A G and LEONARD D R (1971). Asurvey of traffic-induced vibrations. Department ofTransporr TRRL Reporr LR4 18: Transport andRoad Research Laboratory, Crowthorne.
WOOTON L R (1975). Protecting buildings againstground-borne vibration. Ground Engineering. VOI 8,no 4, pp 35–36. London.
Printed in the United Kngdom fo, HMSO
0d8222792 9/90 450 G476 10170
31
