10473289%2 e1995%2e10467418

This article was downloaded by: [Bibliothek der TU Muenchen]On: 21 May 2013, At: 12:50Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of the Air & Waste Management AssociationPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/uawm20

In Situ Velocity Measurements from an IndustrialRotary Kiln IncineratorAllen L. Jakway a , Arthur M. Sterling a , Vic A. Cundy b , Charles A. Cook c & Alfred N.Montestruc da Louisiana State University, Baton Rouge, Louisiana, USAb Montana State University, Bozeman, Montana, USAc Nationai Institute of Standards and Technology, Gaithersburg, Maryland, USAd ELM Engineering Company, Slidell, Louisiana, USAPublished online: 05 Mar 2012.

To cite this article: Allen L. Jakway , Arthur M. Sterling , Vic A. Cundy , Charles A. Cook & Alfred N. Montestruc (1995):In Situ Velocity Measurements from an Industrial Rotary Kiln Incinerator, Journal of the Air & Waste ManagementAssociation, 45:11, 877-885

To link to this article: http://dx.doi.org/10.1080/10473289.1995.10467418

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arisingdirectly or indirectly in connection with or arising out of the use of this material.

http://www.tandfonline.com/loi/uawm20

http://dx.doi.org/10.1080/10473289.1995.10467418

http://www.tandfonline.com/page/terms-and-conditions

TECHNICAL PAPER ISSN 1047-3289 /. Air & Waste Manage. Assoc. 45: 877-885

Copyright 1995 Air & Waste Management Association

In Situ Velocity Measurements from an Industrial RotaryKiln Incinerator

Allen L. Jakway and Arthur M. SterlingLouisiana State University, Baton Rouge, Louisiana

Vic A. CundyMontana State University, Bozeman, Montana

Charles A. CookNationai Institute of Standards and Technology, Gaithersburg, Maryland

Alfred N. MontestrucELM Engineering Company, Slidell, Louisiana

ABSTRACTFor the first time, velocities were measured inside a field-scale rotary kiln incinerator. Combustion gas velocities andtemperatures were measured at multiple points across a quad-rant of the kiln near its exit using a bidirectional pressureprobe and suction pyrometer. To accommodate the new bi-directional probe and gain access to the upper portion ofthe kiln, a lighter and stiffer positioning boom was designed.The kiln was directly fired using natural gas in a steady statemode. Results indicate strong vertical stratification of bothvelocity and temperature, with the highest values corre-sponding to the top of the kiln. Access restraints preventedthe lower region of the kiln from being mapped. Horizontalvariations in both temperature and velocity were insignifi-cant. Operating conditions were varied by adjusting theamount of ambient air added to the front of the kiln. In-creasing the flow of ambient air into the front of the kiln

IMPLICATIONSPrevious experimental work has shown that incineratorflows can be highly stratified in both temperature and chemi-cal species. This latest work shows that the exit of a rotarykiln incinerator can also be highly stratified in velocity, andpresents evidence that regions of reverse flow may exist.It is, therefore, important to consider the general velocityfield when interpreting other measurements taken from therotary kiln section of an incinerator. This is particularly im-portant if single-point sampling is used to characterize theincineration process, so that stagnant areas and regionsof reverse flow can be identified. This work presents a de-vice and methodology for measuring velocities in high-tem-perature, particulate-laden turbulent flows.

reduced the measured temperatures as expected, but did hothave as significant an effect on measured velocities. Thequality of the results is examined by performing mass bal-ances across the incinerator and by comparison to an exist-ing numerical model. Both methods indicate that theexperimental results are reasonable.

INTRODUCTIONAt Louisiana State University, an ongoing research programis focused on obtaining a better understanding and charac-terization of the physical and chemical processes associatedwith rotary kiln incineration. In this particular study, tem-peratures (and for the first time, velocities) were mappedover a significant portion of the exit region of a directly-fired, field-scale, rotary kiln incinerator under controlledexperimental conditions. These measurements provide moreinsight into the complex heat transfer and fluid dynamicsoccurring inside the rotary kiln incinerator chamber. Theyalso provide a means to develop and validate numericalmodels of these phenomena. Instrumentation used to ob-tain these data is discussed, the data are presented and dis-cussed, and comparisons with a numerical model areprovided.

BACKGROUNDThis study was performed at the Dow Chemical Companyrotary kiln incinerator located in Plaquemine, Louisiana.This facility has been described by Cundy, et al.1 Access tothis kiln for experimental measurements is through an off-axis view port located at the back of the transition sectionbetween the exit of the rotary kiln incinerator and the en-trance to the afterburner (see Figure 1). The refractory brickis 33 cm thick at this port, thus limiting boom movement.

Volume 45 November 1995 Journal of the Air & Waste Management Association 877

Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3

Jakway, Sterling, Cundy, Cook, and Montestruc

-3.8m

Measurement Locations

Plan view

Measurement Locations

Ash Pit Concrete Apron

Side view

Figure 1 . Schematic of rotary kiln incinerator and sampled area,Dow Chemical Company, Plaquemine, LA facility. Not drawn to scale(NTS).

View port geometry, along with its location relative to thekiln, precludes access to all of the kiln exit; however, a newboom, developed to support the measuring devices, alloweda complete quadrant of the kiln exit to be mapped. Designof this boom will be discussed later.

Three off-axis primary burners are located on the kiln'sfront face, each of which may be fired using a combinationof waste and/or conventional make-up fuel (typically natu-ral gas). A large pack/drum loading chute is also located onthe front face. Two tangentially oriented air nozzles on thekiln front face provide external air to increase turbulenceand promote better mixing. Operation with and withoutthe use of this turbulence-enhancing mixing air is denotedas TA-on (turbulent air on) or TA-off (turbulent air off),respectively.

This particular kiln has been the focus of study at Louisi-ana State University since the mid-1980s.1"9 Gas tempera-tures and compositions have been obtained from inside thekiln, afterburner, and stack. Experiments have been con-ducted under a variety of operating and feed conditions.The significant vertical gradients in temperature and chemi-cal composition at this kiln's exit, characterized by a highly

reactive combustion region (high temperatures and high lev-els of combustion products) in the upper kiln and a lessreactive environment (low temperatures and chemical com-positions close to that of ambient air) in the lower kiln, havebeen discussed at length. The gradients were observed inboth steady waste feed experiments and transient pack feedruns, persisting even when the turbulence-enhancing mix-ing air was added.

Leger et al.10 developed a fully three-dimensional numeri-cal model of the flow field inside this rotary kiln incinera-tor. The model reproduced the experimentally-observedvertical stratification and further predicted the existence ofa recirculation region in the lower area of the kiln exit. Aparametric study using the model showed that the locationand quantity of unmetered air infiltrating the kiln have amajor influence on the flow inside the kiln. Overall, thestudy demonstrated that a relatively simple numerical modelof a rotary kiln incinerator can provide valuable insight intothe process, especially when used in conjunction with ex-perimental data.

Results of these experimental and numerical studies havehelped to provide a better picture of the incineration pro-cess in this rotary kiln incinerator; however, the picture isfar from complete. A velocity map with corresponding tem-peratures at the kiln exit is needed to improve understand-ing of the flow dynamics, and to assist both in interpretatingpast data and further development of the model. This isalso an important step in generalizing results from the kilnunder study to other rotary kiln incinerators.

APPARATUS OVERVIEW

Velocity ProbeAny probe used inside an operating incinerator must besturdy, since field conditions often involve difficult physi-cal layouts and rough handling, not to mention low ve-locities and high temperatures, both of which fluctuate,and an oxidizing and corrosive environment inside thekiln. These limitations ruled out the use of laser opticalmethods, hot wire anemometry, and typical narrow-borePitot tube instruments commonly used to measure veloc-ity. Robustness is the critical criterion for probe design inthis work environment.

McCaffrey and Heskestad11 developed such a probe foruse in flame and fire applications. This robust, bidirectionalprobe is sensitive for use in low velocity flows (as low as0.3 m/s), and is relatively insensitive to the flow orienta-tion. Kent and Schneider12 used the bidirectional probe todetermine velocities in large pool fires. Measured velocitiesfrom their work ranged from an average of 4.6 to 12.6 m/swith temperatures of 460 to 1025 K. The probe was modi-fied for use in the current work by installing an extra tube,as shown in Figure 2, providing a combination of structuralsupport and air cooling for the probe.

878 Journal of the Air & Waste Management Association Volume 45 November 1995

Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3


Ceramic FabricShroud

Differential PressureTap Lines

Cooling-AirSupport Tube

Section A-A

Typical Cooling AirOutlet Vents

Figure 2. Bidirectional velocity probe schematic (NTS).

Velocity measurement using this probe is based on thedifferential between static and stagnation pressures. Underthe Ideal Gas Law and Bernoulli assumptions, the free streamvelocity, V, and the measured pressure differential betweenstagnation and static pressures, AP, can be related as follows:

V =APAP V 2RT AP

MP 0 C (1)

where R is the universal gas constant; T is the absolutetemperature of the gas; the term AP /1 AP gives the correctsign to V; M is the molecular weight of the gas; C is a cali-bration constant; and Po is the static pressure in the kiln.For the experiments presented here, the gas is assumed tobe air and the static pressure is assumed to be atmospheric,since the kiln is operated at only very slight negative pres-sures. For Reynolds Numbers (based on the probe instru-ment head outer diameter) greater than 600 and less than4000, Kent and Schneider12 determined the calibration con-stant, C, to be 1.07. Kent and Schneider12 also found theprobe to be relatively insensitive to its alignment with theflow, producing errors no higher than 10% for angles upto ± 50° (where 0° represents probe alignment parallel toflow streamlines).

Pressure TransducerThe pressure differential, AP, was measured using an MKSInstruments Inc. model 220CO-00001A2BS pressure trans-ducer. This instrument is rated from 0 to 0.5 in. water col-umn (0 to 1.27 cm), with an accuracy of ± 0.005 in. watercolumn (± 0.13 mm), and was factory calibrated.

Suction PyrometerTo minimize radiation-induced error, a suction pyrometerwas used to measure gas temperatures. The pyrometer is a1.59 mm diameter, sheathed and grounded, type K thermo-couple housed in a 9.53 mm OD Monel tube. This pyrom-eter was attached to the boom and placed in close proximityto the velocity probe, but not so close as to affect the free-stream velocity. Gas flow through the pyrometer was pro-vided by an eductor using high pressure air for the driver. Asuction flow rate of approximately 2.3 standard cubic metersper hour (SCMH) was maintained throughout the test pro-gram. Exhaust gas from the eductor was routed back intothe incinerator downstream of the kiln exit.

BoomThe boom supports the velocity probe and suction pyrom-eter in the kiln and protects the associated tubing fromthe kiln's environment. The access port to the kiln is lo-cated 3.8 m downstream from the exit of the rotary kilnincinerator (see Figure 1). Hence, a relatively long and stiffboom was required in order to reach into the incineratorwhile maintaining confidence in the measurement loca-tion. A circulating-water-cooled boom used in previousexperiments19 was considered for use in this study;

Cooling Water Return

Thin StainlessSteel OuterJacket

Aluminum iSeparatorTubes Pressure and Suction

Support Lines Pyrometer

Section A-A

Thicker StainlessSteel OuterJacket

SuctionPyrometer

AirAnnulus

6.35 m

5.08 cmOD

Section B-B

Figure 3. Circulating-water-cooled boom with bidirectional veloc-ity probe tip (NTS).


Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3


Table 1 . System input parameters.

MeasuredParameter

KilnUpper BurnerMiddle BurnerLower BurnerSludge LanceTurbulence Air

AfterburnerBurner ABurner B

TA-offAir1

990+8.224014.5990±7.4—

0

990±3.5710113

Fuel1

0.268—

27910.287——

23825510.384

Steam 2

2211.402112.53710.40—

4.510.070

1SCMH - Standard Cubic Meters /Hour: (1 atm and 21.1 °C for air) (1.022 atm and 21.12kg/hr.

Air'

99011024014.899016.9—

2300

99013.2710+8.3

°C for natural gas).

TA-onFuel1

22510.165—

27910.176——

23825510.280

Steam 2

2210.2002212.73710.35—

4.510.130

however, because of the large droop associated with theprevious boom design and the inability to easily modifythe boom to accommodate the velocity instrument headand associated tubing, this was ruled out. Consequently,we designed a light, stiff, and robust boom cooled withcirculating water. A schematic of this boom is shown inFigure 3. The new boom design incorporated the follow-ing improvements:

• A removable probe-tip plate allows the boom to beused with different probes

• Two concentric aluminum jackets direct water flowand provide stiffness to the boom

• An air annulus between the inner and outer alumi-num water jackets reduces the weight of the boomand isolates the hot return water from the cooler sup-ply water

• The use of aluminum on all parts which are not di-rectly exposed to the kiln environment reduces theweight of the boom

• A thicker-walled tube incorporated along the base ofthe boom provides added support where the bend-ing moment is highestThe velocity probe tip was aligned as close to co-

axial with the expected kiln flow as possible. Tip deflectionwas measured outside the kiln to be 11.4 cm when the boomwas filled with water and cantilevered. This same deflectionwas assumed to occur when the boom was fully insertedinto the kiln, as the return cooling water, in general, washeated only by 17 °C, exiting the boom around 33 °C.

Boom Positioning RackFor these experiments, a positioning rack was designed tosecurely hold the boom during measurement periods andto ensure repeatability of the position inside the kiln. Therack was attached to a handrail near the view port throughwhich the probe was inserted.

EXPERIMENTAL METHODIn the tests reported here, natural gas, air, and steam werefed to the kiln at the rates shown in Table 1. Rates are givenas mean and standard deviations from data recorded everyminute by permanent facility equipment. Kiln rotation ratewas set to 0.25 rpm. No waste was fed to the kiln, nor wasthere a solids bed in the kiln during these experiments.

At the beginning of the experiment, the two pressure linesfrom the pressure transducer were connected together andvented to the atmosphere in order to establish a zero pres-sure differential. This procedure was repeated throughout theexperiment to monitor zero drift in the pressure transducer.Zero drift was corrected for during post-run data analysis.The pressure transmission lines connecting the differentialpressure cell to the velocity probe were periodically purgedwith high-pressure, dry nitrogen to ensure clear and dry lines.A sampling matrix was established prior to the experiment;during the experiment, the probe was positioned as close aspossible to the predetermined locations using the position-ing rack. The actual locations were recorded using locatorson the boom and the boom positioning rack.

Temperature and differential pressure data were recordedfor 90 seconds at most locations. For two locations duringeach operating condition (TA-on and TA-off), the data wererecorded for 240 seconds, corresponding to the time requiredfor one complete kiln revolution. Temperature data were re-corded at 1.0 Hz while differential pressure data were recordedat both 0.3 and 1.0 Hz. After data were recorded at each probelocation, the boom was moved, the zero differential pressurereading was recorded, and position, pressure differential, andtemperature were measured at the new location.

RESULTSMean Velocity and Temperature Data

Figure 4 presents a view of the kiln exit cross section show-ing the locations where measurements were obtained.


Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3


Kiln ExitCross Section

Figure 4. Kiln exit cross section showing measurement locations.

Table 2 lists the location coordinates along with the corre-sponding velocity and temperature data.

Measurement locations were approximately the same forboth TA-on and TA-off operating conditions. Location co-ordinate error was estimated to be ±0.2 m in a sphere sur-rounding the boom tip. Plans were to map a greater portionof the kiln exit; however, slag build-up in the access win-dow hindered boom movement. Still, a complete quadrantof the upper kiln exit was mapped. Temperatures shown areaverage values from the 1 Hz data. Velocities were calcu-lated at 1-second intervals using a factory calibration of thepressure transducer, the recorded zero drift, and Equation 1.A velocity mean and standard deviation were then calcu-lated for each location. Small-scale, turbulent fluctuationswere not measurable, due to the characteristic slow responsetime and insensitivity to flow orientation associated withthe velocity probe. Raw velocity and temperature data are

presented in Figure 5 for the operating condition of TA-onat three different locations along with a typical zero read-ing. In this figure, location 2 illustrates the relatively lowfluctuations about the mean, common in the upper regionof the kiln, while locations 10 and 12 demonstrate the widerfluctuations typically observed close to the center of the kiln.Location 10 presents conditions suggestive of intermittentregions of reverse flow. Difficulties with the velocity instru-mentation were encountered at two locations—location 3during TA-off operation and location 9 during TA-on opera-tion—and therefore, the data associated with these locationsare not reported. Probe-based Reynolds Numbers rangedfrom 1000 in the upper region to 300 near the centerline.Calibration work by Kent and Schneider12 shows a drop inthe calibration constant from -1.07 to -1.02 betweenReynolds Numbers of 600 to 300 along with an increase inuncertainty. However, the resulting error in calculated ve-locities near the centerline, up to 5%, was ignored in thepresent study.

Figures 6 and 7 show the velocity and temperature dataas a function of vertical position, along with fitted lines (tobe discussed later). Strong vertical stratification is evident intemperature and velocity during both TA-on and TA-off op-erating conditions. Velocities and temperatures increase sig-nificantly from kiln centerline to the upper regions of thekiln. In contrast, closely spaced and even overlapping datafor a given vertical position indicate the lack of stratifica-tion in the horizontal direction. An exception to this lack ofhorizontal variation is that, for both operating conditions,the velocity at location 2 is lower than at location 1. Onepossible explanation is that location 2 may be close enoughto the wall for wall effects to reduce the velocity. Proximityto the wall may also account for the generally higher tem-perature measured at location 6 and lower temperature at

Table 2. Measurement locations and measured mean velocity and temperature.

Location

§X(m)

Y(m)

Zf(m)

TA-off

Velocity(m/s)

TA-on TA-off

Temperature

(K)TA-on

123456789101112

0.080.530.350.050.521.260.320.040.521.200.050.32

1.451.471.210.790.800.820.340.00-0.01-0.02-0.17-0.18

9.809.649.649.679.519.389.539.649.499.339.609.53

7.0 ±0.75.8 ±0.7

4.1 ±0.73.7 ±0.84.2 ±0.72.9 ±0.8

2.2± 1.11.5 ± 1.42.0 ± 1.02.0 ± 1.2

7.1 ±0.76.6 ±0.66.0 ±0.74.1 ±0.84.2 ±0.84.5 ±0.83.4 ±0.92.7 ±0.8

1.3± 1.82.1 ± 1.22.2 ±1.0

1340 ± 71328 ± 91313± 101225±121219+ 71266 ± 71139± 14978 ± 13975 ± 141001± 24956 ± 26945 ± 14

1289 ±41274 ±51266 ±61188 ±61188 ±41200 ±81115± 11977 ± 16973 ± 12962 ± 38851 ± 16859 ± 34

§ X, Y, and Z measurements have an approximate error sphere of 0.2 meters.t Distance from front (burner) face of the kiln.** Data omitted because of problems/instabilities in the velocity measurements.


Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3


0 100 200 300 400 500 600 700 800

Elapsed Time (sec.)1300

0 100 200 300 400 500 600 700 800

Figure 5. Sample raw data from three locations: TA-on.

6-1

I 2\

-2

O Experimental DataLinear Data FitExtrapolated Linear Fit

11,12

-1.5 " - i ' -6.51 6' ' ' as1 ' i

Distance Above Kiln Centerline (m)

1.5

1400

13001

1200;

noo:IOOO;

900

8oo;

700

600

O Experimental DataLinear Data FitExtrapolated Linear Fit

5 -1 -0.5 0 0.5 1 1.5

Figure 6. Experimental data with curve fits used in mass flow cal-culations: TA-off.

-2

O Experiment DataLinear Data FitExtrapolated Linear Fit

-1.5 -1 -0.5 0 0.5 1

Distance Above Kiln Centerline (m)

1.5

1400

BOO;

1200

1100

1000

900

800:

yotf

600

Experiment DataLinear Data FitExtrapolated Linear Fit

-1.5 -0.5 0.5 1.5

Figure 7. Experimental data with curve fits used in mass flow cal-culations: TA-on.

location 2. A study of the data recorded for one completerevolution of the kiln indicated no variance in velocity ortemperature with kiln angular location. These results areconsistent with previous observations in this kiln.16-89

When turbulent air is turned on, the amount of metered,ambient air entering the front of the kiln approximatelydoubles. This is the only independent input parameter thatdiffers between the two cases. Thus, it is expected that thetemperatures should be lower and the velocities higher inthe TA-on case. During TA-on conditions, average tempera-tures are indeed lower, but only by approximately 45 °C;velocities show an even smaller difference with one of theaveraged velocities actually lower for the TA-on case. Rea-sons for this small effect on velocities are the following: first,a large amount of unmetered air infiltrates the incinerator(5.5 and 3.2 times the metered air flows for the TA-off andTA-on cases, respectively); second, less unmetered air infil-trates the incinerator during the TA-on conditions; and third,lower gas temperatures result in higher gas densities.

Previous numerical modeling10 for conditions similar tothis experiment suggested the existence of reverse flow atthe kiln exit. During this experiment, short-duration reverseflow was observed near the kiln centerline; however, quan-tification was not obtained. The pressure transducer was ini-tially calibrated to read only positive pressure differences.


Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3


We planned to interchange the pressure taps when reverseflow was encountered, so that a positive pressure differencecould be maintained, thus allowing quantification of anyreverse flow. This swapping was accomplished by changingvalve settings on a five-valve manifold, a process that took15 to 30 seconds. Unfortunately, the reverse flow had ashorter than expected duration (typically 5 seconds or less)and was infrequent. After the experiment, it was determinedthat the instrument remained linear into the near reversepressure range. This information was used to calculate thenegative velocities shown in Figure 5. Combining the trendof decreasing velocity with elevation together with the ap-pearance of short periods of reverse flow at some of the lowersampling points suggests that, had lower regions beensampled, substantial reverse flow may have been detected.

Mass Flow StudyThis paper presents, for the first time, velocities measured in-side an operating incinerator using the bidirectional probeassembly of Figure 3. Since this is the first time the probe hasbeen used in a confined combustion environment, the dataobtained need to be examined and checked for reasonable-ness. This check is accomplished by comparing the mass flowat the exit plane of the incinerator, which is calculated in twoways. The first technique uses the measured velocities andtemperatures to calculate the mass flow, while the secondmethod is based on a mass balance across the kiln. The massbalance method uses the metered inputs along with an esti-mate of unmetered air infiltrating into the system. Since thesecond method does not involve the measured velocities, itprovides an independent check of the experimental data.

Kiln Mass Flow from Experimental Velocity and Temperature

Measurements. To calculate a kiln mass flow rate using themeasured data, the individual point values of temperatureand velocity shown in Table 1, and graphed irt Figures 6 and7, were first transformed into smooth surface functions. Be-cause of the lack of horizontal variation, the surface equa-tions were assumed to vary with "y" only, where "y" is thevertical direction measured from kiln centerline. Utilizinglinear approximations for the y-dependency of temperatureand velocity along with the ideal gas approximation, thelocal mass flux was calculated. By integrating this local massflux over the area of the kiln's upper quadrant, where theexperimental values were taken (29% of the total cross-sec-tional area), the total mass flow rate through this region ofthe kiln was obtained. Results of these calculations indicatemass flow rates of 2.5 kg/s and 2.9 kg/s for the TA-off andTA-on conditions respectively. If the lack of horizontal varia-tion observed in this quadrant of the kiln is true of the otherupper quadrant as well, then the above calculated flows canbe doubled to yield the mass flow in the entire upper 58% ofthe kiln: 5.0 kg/s and 5.7 kg/s for the TA-off and TA-on con-ditions, respectively.

To determine the total mass flow through the rotary kilnincinerator, the linear curve fits to the temperature and ve-locity data are assumed to extend to the bottom of the kiln.These straight-line extrapolations are shown as dashed linesin Figures 6 and 7. (Note that this extrapolation implies theexistence of reverse flow in the lower region of the kiln.)Using this approximation, along with horizontal symmetry,the total net mass flow out of the kiln was calculated to be4.5 kg/s and 5.2 kg/s for the TA-off and TA-on conditions,respectively.

The assumption that the velocity and temperature datatrends extend linearly to the bottom of the kiln is ratherbold. Preliminary numerical modeling suggests, however, thatlinear extrapolation is usually appropriate. If this linear ex-trapolation method is indeed valid, the resulting net massflow rates across the entire exit plane (4.5 kg/s and 5.2 kg/sfor the TA-off and TA-on, respectively) being less than thenet mass flows out the upper half of the exit plane (5.0 kg/sand 5.7 kg/s for the TA-off and TA-on, respectively) indicatesthat there must be recirculation in the lower region of thekiln, and that the majority of the flow out of the kiln takesplace in the upper region, as expected.

Kiln Mass Flow from Mass Balance. The mass flow rate calcu-

lated from the experimental data can be compared to thatfound from a mass balance on the system. To perform a massbalance, all inlet and exit flows need to be quantified. Infor-mation on all metered inlets to the incinerator is availableand has previously been presented in Table 1. However, thereis also a considerable amount of unmetered air infiltratingthe incinerator. This infiltration results from the operationof the incinerator at a slight vacuum (1.1 cm negative watercolumn during TA-off operation and 1.0 cm negative watercolumn during TA-on operation), which is done to preventfugitive emissions. This unmetered air infiltration rate mustbe determined in order to complete the mass balance.

Unmetered air infiltrates this system from the front andrear rotary seals of the kiln, around the perimeter of the sol-ids loading chute at the front face of the kiln, and throughthe pressure relief hatch near the front of the incinerator.Additional air infiltrates the system through various instru-mentation ports and other small openings. Although theamount of unmetered air entering through any one of thesesources is not known, their combined effect can be calcu-lated in two different ways. The first way, termed theOxygen Method, uses the metered flow rates into the incin-erator, the measured dry oxygen concentration at the stack(13.3% for TA-off and 13.5% for TA-on operation), and theassumption that the natural gas is pure methane which re-acts completely to water vapor and carbon dioxide.79

A second way of calculating the unmetered air infiltra-tion rate is to perform a mass balance across the whole incin-erator on a dry basis, and is termed the Mass Balance Method.Performing the mass balance on a dry basis allows use of the


Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3


Table 3. Calculated leak air (SCMH).

TA-Off TA-On

Oxygen Method 21,400Mass Balance Method 21,630

19,89019,550

measured stack flow rate (24,551 SCMH forTA-off and 24,755SCMH for TA-on operation), which was recorded on a drybasis. Again using the assumptions of pure methane com-pletely combusting to water vapor and carbon dioxide, theunmetered air infiltration rate can be determined.79 Resultsof these calculations are shown in Table 3.

This table shows that the two methods of calculatingunmetered infiltrating air compare very favorably. An aver-age is used in subsequent calculations. Comparison of thesedata to those in Table 1 also show that the unmetered airinfiltrating the incinerator can be as much as 5.5 times theamount of metered air fed into the incinerator. This in-leak-age is commonly included when calculating the amount ofmetered air needed to ensure complete combustion in anincinerator.

The problem now reduces to one of proportioning theunmetered infiltration air in order to determine the massflow in the rotary kiln incinerator. Obviously, this is a diffi-cult process requiring a considerable degree of estimation.Leger et al.10 reasoned that 55% of the total unmetered airinfiltrating this system entered at the front face of the kiln.Using this estimate, the mass flow leaving the exit plane ofthe kiln was calculated to be 5.1 kg/s for the TA-off case and5.5 kg/s for the TA-on case.

These mass flow values differ by 13% and 8%, respec-tively, from the values calculated above using the experi-mentally determined temperature and velocity. Thus the dataappear reasonable.

Numerical ModelA second way to examine the reasonableness of the experi-mental data is by comparison to a numerical model. Leger etal.10 constructed a three-dimensional numerical model of thesame rotary kiln incinerator studied in this work. The modelis a finite difference type utilizing the SIMPLEC algorithm.The main weakness of this model is that radiation heat trans-fer is not included. While Leger et al. modeled the same in-cineration facility utilized in the present study, the modelinputs do not exactly match the present experimental con-ditions. For TA-off, the modeled kiln inputs differ from theoperating conditions (previously presented in Table 1) in thefollowing ways: the kiln natural gas flow was 27% lower, themetered air to the afterburner was 8% higher, and theunmetered infiltrating air was 25% lower. However, incin-erator operating conditions for the TA-on case were nearly

identical to those of the present study. Given the similaritybetween the model inputs and the operating conditions ofthis paper, useful comparisons can be made between themodeling and experimental results.

As an expected result of the omission of radiation heattransfer, the temperatures are over-predicted for both theTA-off and TA-on cases. Model predicted velocities at thekiln exit for the TA-off case ranged from a high of 7.7 m/s atthe top, to 0.7 m/s at centerline, to -1.5 m/s at the bottomof the kiln, in an approximately linear fashion. For the TA-on case, velocities were 8.6 m/s at the top, 1.0 m/s atcenterline, and - 1.3 m/s at the bottom of the kiln, againvarying in an approximately linear fashion.

These modeled velocities are close to the measured val-ues, but are slightly higher at the top of the kiln and mar-ginally lower at the centerline of the kiln. The model alsoconfirms the uniformity of the flow field in the horizontaldirection. At a distance one quarter from the top, the maxi-mum deviation from the predicted mean values of velocityand temperature taken horizontally across the kiln are re-spectively 10% and 14% for TA-off, and 13% and 1.22% forTA-on operation. The small size of these deviations predictedby the model improves confidence in the assumption thatthe lack of horizontal variation of velocity and temperaturein the quadrant sampled extends to the other quadrants ofthe kiln exit.

SUMMARYA new device for measuring velocities and temperatures in-side a directly-fired, full-scale, rotary kiln incinerator has beendeveloped, constructed, and tested. Temperatures (and forthe first time velocities) were mapped across an upper quad-rant of a rotary kiln incinerator during steady state burningof natural gas. The experimental results and ensuing analy-sis provide the following conclusions.

• Stratification of both temperature and velocities isevident in the vertical, but not in the horizontaldirection.

• The highest velocities and temperatures were recordedat the top of the kiln.

• The effect of the turbulent air jets on velocities andtemperatures at the exit of the rotary kiln incinera-tor is largely mitigated by the large amount ofunmetered air infiltrating at the front of the kiln.

• Temperature values and data stratification trends gen-erated by this new device agree with previous experi-mental findings on the same incinerator under similaroperating conditions.

• Mass flow rates calculated using the experimental re-sults compare favorably with mass flow rates calcu-lated by a mass balance across the kiln.

• Numerical modeling also produces results that com-pare favorably with those generated by the new ex-perimental device.


Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3


• The need to sample a complete vertical traverse ofthe rotary kiln incinerator to determine the amountof air infiltrating at the front of the kiln and the pos-sible existence of reverse flow at the exit of the kilnis reinforced.

Prior to obtaining the velocity and temperature data re-ported in this paper, there were virtually no means to quan-tify the mass flow inside rotary kilns. This led to uncertaintyabout the distribution of unmetered infiltration air and theflow dynamics at the locations of previous sampling efforts.Because the limited access of this incinerator did not allowa complete mapping of the kiln exit, the amount ofunmetered air entering the kiln is still uncertain. However,the results do help to provide confidence that the estimatedinfiltration air distribution of Leger et al.10is realistic. Fur-ther, the results indicate that, with adequate access, this probeassembly could completely characterize the mass flow fieldof this rotary kiln incinerator or other similar combustiondevices. Limitations include material compatibility with kilnenvironment, pressure transducer limitations, and calibra-tion constant applicability limits (not a theoretical limit,but so far determined only for the Reynolds Number rangeof 300 to 4000 according to Kent and Schneider12). In addi-tion, the relatively good agreement between the measuredand calculated kiln mass flow rates suggests that the experi-mental techniques used and the assumptions imposed (hori-zontal symmetry at the exit region of the kiln, along withlinear velocity and temperature profiles) are reasonable.While this may seem a circular argument, it should be notedthat, primarily due to the complexity of the system, neverbefore has the flow field of an operating rotary kiln incin-erator been quantified. With each piece of new informa-tion, the picture of what takes place inside a rotary kilnincinerator becomes clearer, and the ability to test previousassumptions becomes possible. While the measurementsreported in this paper only covered a portion of the kilnexit region, they have added considerably to our understand-ing of the complicated process of rotary kiln incineration.

ACKNOWLEDGMENTSThe authors gratefully acknowledge the assistance and co-operation of the Louisiana Division of Dow Chemical Com-pany located in Plaquemine, Louisiana. In particular, theassistance of Jonathan Huggins, J.J. Hiemenz, and ChuckLipp is appreciated. Although this research was undertakenwith the cooperation of Dow Chemical Company, it doesnot necessarily reflect the views of the Company and there-fore no endorsement should be inferred. Fellowship awardsfrom the State of Louisiana Board of Regents for authorsAllen Jakway and Charles Cook are gratefully acknowledged.The support offered by Louis J. Thibodeaux, Director of theHazardous Substance Research Center South and Southwest,and David Constant, Director of the Hazardous Waste Re-search Center, both of the Louisiana State University, is

appreciated. The capable assistance of Rodger Conway alongwith that of Jodi Roszell is also acknowledged.

REFERENCES1. Cundy, V.A.; Lester, T. W.; Sterling, A.M.; Morse, J.S.; Montestruc, A.N.;

Leger, C.B.; Acharya, S. "Rotary kiln incineration I. An in depth study-liquid injection," JAPCA 1989, 39:1, 63 - 75.

2. Cundy, V.A.; Lester, T.W.; Sterling, A.M.; Montestruc, A.N.; Morse, J.S.;Leger, C.B.; Acharya, S. "Rotary kiln incineration III. An in depth study—kiln exit/afterburner/stack train and kiln exit pattern factor measure-ments during liquid CC14 processing," JAPCA 1989, 39.7 944-952.

3. Cundy, V.A.; Lester, T.W.; Sterling, A.M.; Montestruc, A.N.; Morse, J.S.;Leger, C.B.; Acharya, S. "Rotary kiln incineration IV. An in depthstudy—kiln exit, transition and afterburner sampling during liquidCC14 processing," fAPCA 1989,39:8, 1073-1085.

4. Lester, T.W.; Cundy, V.A.; Sterling, A.M.; Montestruc, A.N.; Leger, C.B.;Acharya, S. "Dynamics of rotary kiln incineration of toluene/sorbentpacks," Combustion Science and Technology 1990, 74, 67-82.

5. Leger, C.B.; Cundy, V.A.; Sterling, A.M.; Montestruc, A.M.; Jakway, A.L.;Owens, W.D. "Field-scale rotary kiln incineration: Oxygen responses atthe kiln afterburner and stack," Remediation, Summer 1991, 275-291.

6. Cundy, V.A.; Sterling, A.M.; Lester, T.W.; Jakway, A.L.; Leger, C.B.; Lu,C; Montestruc, A.M.; Conway, R.B. "Incineration of xylene/sorbentpacks. A study of conditions at the exit of a full-scale industrial incin-erator," Environmental Science and Technology 1991, 25:2, 223-232.

7. Cook, C.A.; Cundy, V.A.; Sterling, A.M.; Lu, C; Montestruc, A.N.; Leger,C.B.; Jakway, A.L. "Estimating dichloromethane evolution rates froma sorbent bed in a field-scale rotary kiln incinerator," Combustion Sci-ence and Technology 1992, 85, 217-241.

8. Leger, Christopher B.; Cundy, Vic A.; Sterling, Arthur M.; Montestruc, AlfredN.; Jakway, Allen L; Owens, Warren D. "Field-scale rotary kiln incin-eration of batch loaded toluene/sorbent. I. Data analysis and bedmotion considerations, "Journal ofHazardous MateriaLs 1993,34,1-29.

9. Leger, Christopher B.; Cook, Charles A.; Cundy, Vic A.; Sterling, ArthurM.; Montestruc, Alfred N.; Jakway, Allen L.; Owens, Warren D. "Field-scale rotary kiln incineration of batch loaded toluene/sorbent. II. Massbalances, evolution rates, and bed motion comparisons," Journal ofHazardous Materials 1993, 34, 31-50.

10. Leger, C.B.; Cundy, V.A.; Sterling, A.M. "A three-dimensional detailednumerical model of a field-scale rotary kiln incineratbr," Environmen-tal Science and Technology 1993, 27, 677-690.

11. McCaffrey, B.J.; Heskestad, G. "A robust bidirectional low-velocity probefor flame and fire application," Combustion and Flame 1976,26,125-127.

12. Kent, L.A.; Schneider, M.E. "The design and application of bi-direc-tional velocity probes for measurement in large pool fires," ISA Trans-actions 1987,26:4, 25-32.

About the Authors

Dr. Jakway recently completed his doctorate in Mechani-cal Engineering, and Dr. Arthur M. Sterling (correspond-ing author) is Chairman of the Chemical EngineeringDepartment at Louisiana State University, BatonRouge, LA 70803, U.S.A. Dr. Cundy is Head of theMechanical Engineering Department at Montana StateUniversity in Bozeman, Montana and was formerlyChairman of the Mechanical Engineering Departmentat Louisiana State University in Baton Rouge. Dr. Cookis a National Research Council Postdoctoral ResearchAssociate at the National Institute of Standards andTechnology in Gaithersburg, Maryland and formerPostdoctoral Research Associate at Louisiana StateUniversity in Baton Rouge. Mr. Montestruc, PE, is withELM Engineering Company in Slidell, Louisiana.


Dow

nloa

ded

by [

Bib

lioth

ek d

er T

U M

uenc

hen]

at 1

2:50

21

May

201

3

Date post:	18-Dec-2014
Category:	Design
Upload:	muhammad-awais
View:	32 times
Download:	0 times