Design and Development of Sand Bed Filter

15B. S. Pathak et al. / International Energy Journal 8 (2007) 15-20

Note: Some tar values are indicative since definitions are not specified

Fixed bed

cocurrent gasifier

Fixed bed countercurrent

gasifier

CFB gasifier

Particles, mg N-1m-3

100 - 8000 100 - 3000 8000 – 100,000

Tar, mg N-1m-3

10 - 6000 10,000 – 150,000 2000 – 30,000

Mukunda et al. developed sand bed filter to remove theparticulates collected by the cooling water in spray tower.The filter is separated in to coarse and fine sections. Thecoarse filter is filled with sand of 0.5 – 2 mm size particlesand the fine sand bed filled with 0.2 – 0.6 mm size sand. Thesize of the filter area is so chosen that the gas velocitiesthrough the filter bed do not exceed 0.1 m s-1. This lowvelocity coupled with the tortuous path causes the removalof a large part of the dust from the gas. The results indicatedthat the hot-end tar is about 100 mg m-3 and comes down to20 +/- 10 mg m-3 at the end of the fine filter (cold end). Theparticulates level also comes down to 50 mg m-3 at the coldend from about 700 mg m-3 at the hot end [2]. Hasler et al.showed that a 90% particle removal is easier than a 90% tarremoval. Based on experimental data for the removal ofparticles and tar, none of the investigated gas cleaningsystems can safely meet the gas quality requirements for

Table 1. Gas quality of raw producer gas from atmosphericair blown biomass gasifier [1]

IC engine Gas turbine Particles, mg N-1m-3 < 50 < 30 Particle size, μm < 10 < 5 Tar, mg N-1m-3 < 100 Alkali metals, mg N-1m-3 0.24

Table 2. Gas quality requirements for power generators [1]

www.serd.ait.ac.th/reric

Design and Development of Sand Bed Filter for UpgradingProducer Gas to IC Engine Quality Fuel

B. S. Pathak*, D. V. Kapatel*, P. R. Bhoi*1, A.M. Sharma* and D. K. Vyas*

Abstract - Engine quality producer gas must be almost free of solid particulate matters (SPM) and organic contaminants(Tar) to minimize engine wear and maintenance. Except for the catalytic tar crackers, none of the gas cleaning systemscommercially available can securely meet a tar removal exceeding 90 % and hence new concepts for tar removal arerequired. This paper presents a design and development work of sand filter for upgrading producer gas to IC enginequality fuel. The developed sand filter was tested for its performance with SPRERI’S 20 kWe down draft circular throattype gasifier with engine set up. The experimental investigations show that the percentage reduction in tar and particulatematters is above 90 %. The total amount of tar and particulate matters was 319 mg N-1m-3 and 53mg N-1m-3 before andafter filter respectively.

Keywords - IC engine application, Producer gas cleaning, Regenerative sand bed filter, Tar and particulate matters.

*Thermo-chemical Conversion Division, Sardar PatelRenewable Energy Research Institute, Vallabh Vidyanagar - 388120, Gujarat.

1Corresponding author:Tel.: +91-2692-231332 / 235011 Fax: +91-2692-237982E-mail: [email protected]

1. INTRODUCTION

The present study was carr ied out under qualityimprovement programme of thermo-chemical conversiondivision of SPRERI. The organics produced under thermalor partial-oxidation regimes (gasification) of any organicmaterial are called tar and are generally assumed to be largelyaromatic [1]. The producer gas must be free from tar andparticulate matters and it should be cooled up to the ambienttemperature for IC engine application.

The gas quality requirement for the power generatorsis very strict. However, the postulated gas quality given inthe literature should be interpreted carefully as it dependson the type of engine used and the methods of contaminantsampling and analysis procedures used during experiment.Gas cleaning systems are expected to reduce particles andtar content from the raw producer gas to the postulatedlevels. Typical values of particulates and tar content aregiven in Table 1 [1]. Table shows that the amount of tar ismuch higher in countercurrent than cocurrent gasifiers.Therefore in the present investigation cocurrent throat typegasifier with IC engine for power generation was used toevaluate newly developed sand filter. Gas qualityrequirements for the power generation are very strict (seeTable 2)[1]. Numerous methods of contaminant samplingand analysis procedures are in use, which may lead to results,which are not strictly comparable. Large differences can beexpected, especially, for the sampling and analysis of tar. Inthe present study, the field type tar and particulate samplingunit designed by IIT Bombay was used for filter evaluation.

16 B. S. Pathak et al. / International Energy Journal 8 (2007) 15-20

Wood shaves Coarse sand Fine sandRaw gas collector Clean gas collector

Raw producergas inlet

Clean producergas outlet

Wire mesh

D r a i n t a p

and particulate mattersSampling tap for tar

Pressure tap

satisfactory IC engine applications. The sand bed filter andwash tower have already been successfully tested in fixedbed biomass gasifier coupled to IC engines. The resultsshow that particle collection is less critical than tarseparation. The highest particle separation efficiency wasobserved in sand bed filters, rotational atomizers and in wetelectrostatics precipitators (ESP). The highest tar reductionwas also found in high temperature catalytic tar crackers, inventuri scrubbers and in sand bed filters. [3]. The gas qualityfor successful IC engine operation has been postulated asbelow 50 mg N-1m-3 for the particle and less than100 mg N-1m-3 for the tar [4]. Wet packed bed scrubber, spraytowers or venturi scrubbers are often used for tar andparticulates removal from producer gas. Thesecommercialized cleaning devices make use of water for gascleaning and cooling. For large-scale gasifier systems, waterrequirement would be in huge amount. In addition, there areseveral environmental regulations restricting disposal ofpolluted wastewater in the environment. The tar is presentin vapor form at the gasifier exit where the temperature ofthe producer gas ranges from 350 ºC to 450 ºC. But as thegas cools to less than 150 ºC, the tar condenses and solidifies.For IC Engine application, the gas needs to be cooled toaround 40 ºC i.e. ambient temperature. As stated earliercooling of producer gas to this temperature will condensetar which affect the engine operation [5]. The aim of thepresent work is to design, develop and evaluate performanceof sand bed filter.

2. EXPERIMENTAL METHODS

Design of Sand Bed Filter

The design philosophy behind the sand filter is as follows:1. The sand is neutral and non-reactive material.2. It is inexpensive and easily available.3. It is easily available in different grain size grades.4. It can withstand high gas temperature5. It is easy to clean and recycle i.e. regenerative

material.The sand bed filter was designed, developed and tested

with the existing 20 kWe down draft throat type gasifier withIC engine set up for power generation. The area of sand bedfilter was designed for superficial velocity of 0.1 m s-1 [2]and gas flow rate of 60 Nm3h-1 i.e. the designed flow rate ofthe gasifier. The sand bed filter rectangular was fabricatedfrom 3 mm thick mild steel sheet and wire meshes fabricatedfrom stainless steel 304 were used to separate filter bed.

SPRERI had conducted an experimental study on effectof bed height on pressure drop across the bed on dry packedbed filter. Experimental investigations suggest a linerrelationship between bed height and the pressure dropacross the bed. This in turn, increases power consumptionof the blower. Literature and our experience shows that mostof the gas contaminants (SPM and tar) were deposited inthe initial 20 to 30 mm layers of bed height. Therefore, to beon safe side the bed height of each bed in the present filterwas taken as 85 mm.

The filter consists of following five compartments:First compartment : Raw producer gas collectionSecond compartment : Sun dried wood shavesThird compartment : Φ 0.60 mm coarse sandFourth compartment : Φ 0.15 - Φ 0.35 mm fine sandFifth compartment : Clean producer gas collection

The second, third and fourth compartment were filledwith wood shaves, coarse sand and fine sand respectivelyand each bed having bed height of 85 mm. The sun driedwood shave was used to remove the moisture from theproducer gas. The coarse sand was used to remove coarseparticulates and tar while the fine sand was used to removefine particulates and tar. Wire meshes having size equivalentto that of respective sand size was used to separate thebeds. The three beds of wood shaves, coarse sand and finesand can be removed by opening flange provided at thebottom of the filter. The filter has facility to measure pressuredrop at each bed and also across the filter. Two samplingport were also provided on the top of the filter in the raw gascollector and clean gas collector for measurement of tar andparticulate in raw and clean gas. The schematic diagram ofthe regenerative sand bed filter is shown in Fig. 1.

Fig. 1. Schematic diagram of regenerative sand bed filter toclean producer gas.

Fig. 2. Comparative photographic view of wood shaves, coarsesand and fine sand before and after test.


Water pond

20 kWe

circulatedpump

Grate

towerSpray

Water inlet

Sand bed filter

IC Engine

WaterWater outlet

Gasifier

The properties of sun dried wood shaves, coarse andfine sand were measured at thermo-chemical conversionlaboratory of SPRERI. The density of wood shaves was68 kg m-3 at 9.5 moisture content (w.b.). The density andvoid fraction of fine sand (φ 0.15 – 0.35 ìm) is 1399 kg m-3

and 43 % respectively. The density and void fraction ofcoarse sand (φ 0.6 ìm) is 1450 kg m-3 and 46 % respectively.The comparative photographic view of wood shaves, coarsesand and fine sand before and after test is shown in Fig. 2.

Experimental Setup

The instrumented experimental set up was developedto evaluate the performance of the filter in terms of tar andparticulates removal efficiency. The schematic diagram ofthe experimental set up is shown in Fig. 3. The experimentalset up consists of 20 kWe down draft gasifier, spray tower,sand bed filter and IC engine. The details of gasifier andspray tower were the same as described in Patel et al [6].

System Operation and Measurements

The gasifier was charged with charcoal up to thenozzles and then with woody biomass, is ignited by startingblower and using lighting torch at the ignition port. After15 minutes a good quality of producer gas is generatedwhich is flared at the burner. Then the producer gas wasintroduced to IC engine through the sand filter. The gasflow rate of producer gas of 60 Nm3h-1 was maintainedconstant throughout the experimental investigations. Theperformance of the sand filter was evaluated by measuringthe following parameters.

1. Tar and particulate matters2. Producer gas temperature at inlet and outlet of the

filter.

3. Pressure drop across the filter.4. Water temperature at inlet and outlet of spray tower.Chromel-Alumel (K type) thermocouples and a digital

multichannel temperature indicator were used to measurevarious temperatures. Water-filled U tube manometers wereused to measure the pressure drop across the filter and atthe orifice meter. The field type tar and particulatesmeasurement set up was used to measure tar and particulatesof raw and clean producer gas. This unit measures total tarand particulate (TTP) contents of producer gas. It has thepiston cylinder assembly, Non-return suction valve, andfilter – paper holder assembly, stroke counter and pedal andbase. A Whatman glass micro fiber filter, grade: GF/C cat.No 1822 047 having 47 mm diameter was used in the fieldtype tar sampler.

3. RESULTS AND DISCUSSIONS

The system was extensively tested to evaluate filterperformance. The performance parameters like pressuredrop across the filter, temperature of the producer gas atthe inlet and outlet of the filter and inlet and outlettemperature of the spray tower, tar and particulate mattersin the producer gas before and after the filter were measured.The readings were taken at an interval of 15 minutes. Theexperimental results of tar and particulate matters measuredby using field type tar sampler are shown in Table 3. Table 3shows that the cleaning efficiency of the sand filter obtainedin the range of 83 – 97 %. Table 3-shows significant reductionin collection efficiency with time because the filter area wasreduced due to deposition of tar and particulate in the filtermedia. To avoid this problem water shower can be providedon the top of each bed and collect the contaminated waterfrom the bottom of the filter, which continuously removesthe tar and particulates from the filter media.

Fig. 3. Schematic diagram of experimental setup of the sand bed filter.

B. S. Pathak et al./ International Energy Journal 8 (2007) 15-20 17


However, this techniques leads to another two problemsfirst, requirement of large quantity of water and second,disposal of contaminated water to the environment.

Figure 4 shows the variation of raw gas temperature,combustion zone temperature and drying zone temperatureof 20 kWe down draft gasifier with time, which shows thatthe variation of temperature is almost constant, whichindicates the steady and smooth operation of gasifier.

Weight of Filter Paper, mg Tar and Particulate Sr. No.

Time Sampling No. of

Stroke Final

Before Drying

A

Final After

Drying B

Initial C

Diff. A-C

Diff. B-C

Before Drying, mg/m3

After Drying, mg/m3

Cleaning Efficiency,

%

1 3.00 PM Filter Inlet 40 0.185 0.123 0.094 0.091 0.029 4840 1542 97 2 3.00 PM Filter Outlet 40 0.096 0.095 0.094 0.002 0.001 106 53 ---- 3 4.15 PM Filter Inlet 40 0.109 0.100 0.094 0.015 0.006 798 319 83 4 4.15 PM Filter Outlet 40 0.096 0.095 0.094 0.002 0.001 106 53 ----

Table 3. Tar and particulate matters measurement using field type tar sampler

0

200

400

600

800

1000

1200

3:15

PM

4:15

PM

4:45

PM

5:45

PM

6:15

PM

7:00

PM

12:1

5 PM

1:15

PM

2:15

PM

3:15

PM

4:00

PM

4:45

PM

10:4

8 AM

Time, h

Tem

pera

ture

, deg

C

Raw gas temperature

Combustion zone

Experiment 1 Experiment 2

Fig. 4. Variation of raw gas temperature, combustion zonetemperature and drying zone temperature with time.

Figure 5 shows the variation of gas temperature at filterinlet and outlet with time. It is observed that the slightvariation in the gas temperature at filter inlet and outletwhich is due to variation in raw gas temperature. It is alsoobserved that there is small temperature drop of producergas in the filter.

0

10

20

30

40

50

60

70

3:15

PM

4:15

PM

4:45

PM

5:45

PM

6:15

PM

7:00

PM

12:1

5 P

M

1:15

PM

2:15

PM

3:15

PM

4:00

PM

4:45

PM

10:4

8 A

M

Time, h

Tem

pera

ture

, deg

C

Filter inlet temperatureFilter outlet temperature


Fig. 5. Variation of gas temperature at inlet and outlet offilter with time.

01020304050607080

3:15

PM

4:15

PM

4:45

PM

5:45

PM

6:15

PM

7:00

PM

12:1

5 PM

1:15

PM

2:15

PM

3:15

PM

4:00

PM

4:45

PM

10:4

8 AM

Time, h

Pres

sure

dro

p, m

m H

2OPressure drop across the filter


Fig. 6. Variation of pressure drop across the filter withtime.

0

10

20

30

40

50

60

70

3:15

PM

4:15

PM

4:45

PM

5:45

PM

6:15

PM

7:00

PM

12:1

5 PM

1:15

PM

2:15

PM

3:15

PM

4:00

PM

4:45

PM

10:4

8 AM

Time, h

Tem

pera

ture

, deg

C

Outlet water temperature of sprayer

Inlet water temperature of sprayer


Fig. 7. Variation of water temperature at inlet and outlet ofspray tower with time.

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

3:15

PM

4:15

PM

4:45

PM

5:45

PM

6:15

PM

7:00

PM

12:1

5 P

M

1:15

PM

2:15

PM

3:15

PM

4:00

PM

4:45

PM

10:4

8 A

M

Time, h

Sup

erfic

ial v

eloc

ity, m

s-1

Superficial velocity


Fig. 8. Variation of superficial velocity with time.


Fig. 6 shows the variation of pressure drop across thefilter. It was observed that the pressure drop across thefilter is increased with time, which shows blockage ofcleaning area of the filter, which shows deposition of tarand particulates in to the filter medium. It was also observedthat the pressure drop was low in the bed of wood shavesand coarse sand bed while pressure drop was high in thebed of fine sand. Fig. 7 shows the variation of watertemperature at inlet and outlet of water spray tower. It isobserved that the inlet temperature of water to spray towerremains constant while the outlet temperature decreasesas raw gas outlet temperature decreases and increases asraw gas outlet temperature increases. Fig. 8 shows variationof superficial velocity with time. It is shown that during thefirst experiment superficial velocity is maintained constantas per designed value and in the second experiment it variesdue to variation in flow rate of producer gas as superficialvelocity is function of producer gas flow rate.

Fig. 2 shows that the coloure of coarse and fine sandpresent in the third and final compartment has changed toblack colour from yellow, which proves that the tar andparticulate matter have been trapped. It is also observedthat the fine sand of size of φ 0.15 – φ 0.35 (fine sand) aremore darker in colour than the sand of φ 0.60 mm (coarsesand) which shows that the fine sand are more effective forcleaning than coarse sand. The cooling water washes partof the tar and the rest is deposited in the sand bed filter.Experiments have shown that the moisture carries over fromthe cooler causes slight wetting of the sand bed. It wasobserved that the percentage reduction of moisture contentby the filter was around 10 – 18 %.

4. CONCLUSIONS

1. The developed sand filter worked satisfactorily andthe tar and particulates removal efficiency was in therange of 83 – 97 %.

2. The pressure drop across the filter was increased withthe duration of test, which indicates that the moisture,tar and particulates being filtered out from the producergas.

ACKNOWLEDGEMENT

The fund and infrastructure facilities for the research workprovided by the management of Sardar Patel RenewableEnergy Research Institute is gratefully acknowledged.

REFERENCES

[1] Milne, T.A.; Abatzoglou, N.and Evans, R J. 1998.Biomass Gasifier "Tar" Their Nature, Formation andConversion. The Biomass Energy Foundation Press.1810 Smith Rd. Golden, CO 80401.

[2] Mukunda, H. S.; Dasappa, S.; Paul, P. J.; Rajan, N.K.S.and Shrinivasa, U. 1994. Gasifiers and combustors forbiomass - technology and field studies. Energy forSustainable Development 1(3).

[3] Hasler, P and Nussabaumer, Th. 1999. Gas cleaning forIC engine applications from fixed bed biomassgasification. Biomass and Bioenergy. 16 : 385 - 395.

[4] Knoef, H. A. M. and Koele, H.J. 2000. Survey of tarmeasurement protocols. Biomass and Bioenergy. 18:55 - 59.

[5] Patel, J.B.; Jena, U.; Kapatel, D.V. and Patel, V.B. 2005.Recycling of wastewater generated during producergas conditioning. Proceeding of 39th ISAE Convention& Symposium, ANGRAU Hydrabad, India. 9-11 March.

[6] Patel, S.R.; Swamy, S.M. and Rao, C.S. 1993.Development and performance of a 20 kW gasifiersystem using woody agricultural residues as feedstokes for power generation. Fourth national meet onbiomass gasification and combustion. InterlinePublishing, Banglore, India, 187 - 197.

Performance Test Techniques for EJECTOR VENTURI SCRUBBER*

L. S. HARRIS, Manager, Research and Development Department, R. HARTENBAUM, Research Engineer, Schutte and KoertingCompany, Cornwells Heights, Pennsylvania

I he techniques presented weredeveloped for the purpose of testingand evaluating the performance of theEjector Venturi Scrubber System. Atypical six-inch size system is shownin Fig. 1. The system is comprised ofan ejector venturi scrubber and ascrubber separator. The separator isattached to the discharge end of theventuri unit.

The ejector venturi scrubber8 is a wettype of contactor, which utilizes the jetaction of the contacting fluid to pumpthe gas through the system. The per-formance characteristics of interest forsuch a system are as follows:

1 Pumping capacity and correspond-ing self induced draft.

2 Liquid carry-over in dischargestack.

8 Solid particulate collection effi-ciency.

The tests and evaluation were per-formed on a four-inch scrubber system.The test scrubber and separator weretransparent models made to the samedimensions as our stock units, and canbe seen at the right in Fig. 2.

Pumping Capacity

A typical set of pumping performancecurves of the ejector venturi scrubber atvarious nozzle pressures is shown inFig. 3. Velocities in the suction ductvary from 139 ft/min at the low capac-ities up to a high of approximately 3300ft/min.

Air flow was measured in a four-inchdiameter duct by two means; a spe-cially calibrated Biram's type anemom-eter, for duct velocities from 0 to 2000ft/min, and a 5/ie inch diameter pitottube for duct velocities higher than 2000ft/min.

The pitot tube was located 10 diam-eters from the upstream opening of thestraight inlet duct and three diametersfrom the downstream entrance to thescrubber, as illustrated in Fig. 4.These dimensions are greater thanthose specified by the National Associa-tion of Fan Manufacturers, Inc.,1 for

* Presented at the 54th Annual Meetingof APCA, Commodore Hotel, June 11-15,1961, New York, New York.

determining performance of fans andblowers by means of pitot tube traverse.The aforementioned specifies seven anda half diameters upstream and two and ahalf diameters downstream of straightduct. Air flow into the unit was con-trolled by two methods: (1) by a num-ber of various size orifice plates locatedat the inlet of the duct when the airflow was determined by pitot tube,and (2) by a segmented circular typeshutter located 13 diameters from theinlet of the duct when using the ane-mometer, as illustrated in Fig. 4. Airflows determined by horizontal andvertical traverse pitot tube readings,at velocities in the range of 0 to 1000ft/min, were found to be subject tolarge errors, anywhere from 10% togreater than 100%. The lower thevelocity the greater the errors and un-certainties. The errors and uncer-tainties incurred were due to (1)manometer reading accuracy in deter-mining velocity head pressure in rangeof 0.00074 to 0.063 inch of water(139 ft/min to 1000 ft/min), (2) minorair flow fluctuations of the order of 0.01inch water, (3) manometer zero ad-justment and zero reading inaccur-acies, and (4) the existence of negative(minus) velocity head readings near theduct wall at mean air flows less than1000 ft/min, indicating internal eddyingand flow reversals near duct wall.

To obtain air flow readings withinfive percent accuracy in the range of 0 to2000 ft/min, the specially calibratedfour-inch anemometer was mounted intothe entrance of the four inlet duct.The anemometer was calibrated for totalair flow versus dial reading, by suckingair through the anemometer and HEI(Heat Exchange Institute) standardflow nozzles.2'7 The anemometer wassealed into the entrance of a duct thesame size as used in our scrubber testand the HEI flow nozzle was mountedseven and a half diameters downstream.The calibration was achieved by criticalpressure drop across the nozzle, whichwas created by using an S&K two stagesteam jet vacuum pump. Five HEInozzles XU to z/i inch were used to coverthe range in which we were interested.The anemometer was not used for

determining flow above 2000 ft/min,due to mechanical speed limitationsrecommended by the manufacturer.

During the first series of tests per-formed in the range above 2000 ft/min,the flow was determined by four hori-zonal traverse and four vertical traversepitot tube readings. Since both ver-tical and horizontal traverse readingswere found to be almost identical, withingage reading accuracjr, remainder of airflow readings were determined byhorizontal traverse pitot readings only.

The ejector venturi scrubber's corre-sponding draft was measured immedi-

Fig. 1. Typical six-inch figure 4014 Schutfe andKoerting Company ejector venturi scrubber sys-tem.

May 1 962 / Volume 1 2, No. 5 227

2398

Fig. 2. Solid particulate scrubber test stand.

20 40 60 80 100 120^ 140 160 180

CAPACITY- ACFMFig. 3. Air handling capacity for S&K four-inch 4010 ejector venturi scrubber.

ately after the scrubber inlet flange,by means of a well-type manometerreading inches of water.

Liquid Carry-Over

The term "liquid carry-over" refersto droplets of contacting fluid in thedischarged gas leaving the scrubbersystem. Obviously, in scrubber appli-cations, it is desired to have minimumliquid carry-over. A typical liquidcarry-over performance curve for theejector venturi scrubber system is shownin Fig. 5. Since we were working witha small unit, total sampling by impac-tion was found to be the most prac-tical method. The results obtainedwere reproducible within ±10%. Themethod used consisted of ordinarysix-inch diameter blotter paper rigidlymounted in a support one inch from theend of the scrubber discharge, shown inFig. 6. The circumferential openingwas equal to the flow area and was wideenough so that its effect on pumpingcapacity was nil.

The liquid carry-over droplets werecaught by inertial impaction and heldfast by being absorbed by the blotter.Due to the relatively small amount ofliquid carry-over, the sampling timeeven under the worst conditions was aslong as five minutes before the blotterbecame saturated. The blotter supportand sampling blotter were weighed inthe vicinity of the test, immediatelybefore and after each test. The blotterand support weighed approximately 600grams and the amount of liquid samplecollected varied between one to sixgrams. At the high efficiency points,it took as long as one hour to collect theone gram of liquid carry-over, and thedischarge gas was dry to the "touch"of the hand. A triple beam balancehaving an accuracy of 0.1 gram wasused. This resulted in a 20% uncer-tainty in the lowest readings, which wasfelt to be reasonable when consideringall the factors involved.

The above technique for sampling anddetermining the liquid carry-over wasnot the first method tried. Attemptswere made to obtain representativesamples by an isokinetic probe in thedischarge stack with the droplets beingcaught in a series of drying tubes filledwith glass wool. Results obtained werenot consistently reproducible within10% and the errors were as high as 50and 100%.

Several different size glass samplingprobes were tried, x/4- to

1/2-inch diam-eter. Close examination revealed thatthe small liquid droplets were collectingon the walls of the probe, coalescing,and then falling back down into thestack. After witnessing the afore-mentioned, further attempts to use anytype of probe sampling for liquid carry-over were abandoned. Prior to usingblotter paper as the means of collecting

228 Journal of the Air Pollution Control Association

4"DIA. DUCT

PITOT TUBE PORT.

ANEMOMETER FITTED INTOENTRANCE OF DUCT(REMOVED WHEN USING PITOT TUBE)

PRESSURE TAPFOR DRAFTREADING

CIRCULAR ORIFICESINSTALLED OVER OPENNINGOF DUCT WHEN USINGPITOT TUBE

SEGMENTED CIRCULAR-TYPESHUTTER (REMOVED FROMLINE WHEN USING PITOTTUBE)

AIROUT

WATER OUT

Fig. 4. Air flow fest set up for S&K four-inch ejector venturi scrubber.

the water drops, fiber glass filter padswere tried. This was unsatisfactorydue to the fiber glass pad becomingweak and falling apart. To avoid this,the fiber glass pad was backed up byblotter pad. This produced satisfactoryresults which were reproducible within10%. As a matter of interest, we ranidentical tests with the blotter paperalone and it produced the same resultsas the fiber glass pad backed withblotter paper. Since the fiber glass padwas not needed, it was eliminated andblotter paper was used alone.

Solid Particle CollectionEfficiency t

The typical performance charac-teristic being sought is the grade or

fractional efficiency curve, particle sizeversus collection efficiency, as illus-trated in Fig. 7. The obtaining of thisperformance characteristic involved thefollowing work: (1) developing amethod of generating the desired solidparticle aerosol, (2) determining amethod of obtaining a representativesample of solid particulate in both intakeand discharge ducts, and (3) developinga method of conveniently determiningsolid particle size and distribution.

Aerosol Generator

The method used to generate thedesired solid particulate aerosol in therange of one to 100 microns consisted ofusing an S&K Which air jet exhauster,3

a constant speed turntable type feeder

100

90 -

80 -

70

eO[-f—I

LJDC

COCO 5 0

DCQ_

4 0LLJ

20

10

4" E.V. SCRUBBER WITH S S KSCRUBBER SEPARATOR

4 " E.V. SCRUBBER WITH

GRAVITY SEPARATOR

SCHUTTE a KOERTING CO.CORNWELLS - HEIGHTS

BUCKS CO., PENNA.

O.I 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 L4 1.5

CARRY OVER- GALOFWATER/1000 FT3 AIRx 10*

Fig. 5. Carry-over from a four-inch S&K scrubbing system.

May 1 962 / Volume 1 2, No. 5

similar to the Harvard No. 4 unitdeveloped by M. W. First, et al.,4

and an inexpensive pneumatic vibratorfor the hopper, as illustrated in Fig. 8.The air jet exhauster sucks the particu-late off of the turntable and mixes itwith the air jet and discharges themixture at high velocity into the en-trance of the four-inch scrubber suctionduct. Aerosol concentrations as high asthree grams/cu ft were generated in thismanner. The entire aerosol generatorwas enclosed in a box-like container withopenings provided for suction air to thescrubber and for an exhaust ventingsystem, as illustrated in Fig. 2. Thisarrangement permitted personnel towork in the immediate area without theuse of gas masks for protection againstbreathing the aerosol.

Prior to using the above method,unsuccessful attempts were made to usea similar type of aerosol generator notrequiring the complexity of a constantspeed turntable feeder. The systemconsisted of a V2-inch S&K air jetexhauster, a duPont Dry Dye EductionProbe5 and a barrel or drum vibratedby a pneumatic vibrator. One end ofthe duPont dry dye probe is attachedto the suction port of the air jet ex-hauster while the other end rests in thebarrel of solid particulate. Althoughthis arrangement works well for duPontin introducing dry dye into gasolinewhen using a liquid jet exhauster, it didnot generate a steady enough aerosol forour purposes. The bulk densities of thesolid particulate we used, fly ash andglass beads, were in the range of 90 to300 lb/cu ft, and were very muchhigher than the dry dye used by duPont.It is suspected that this is the reasonwhy the duPont probe did not worksatisfactorily in this application. As amatter of general interest, we have usedthe duPont Probe successfully in appli-cations where it was desired to uniformlysuck or feed other solids directly from abarrel, such as Celite No. 379 (diato-maceous earth), having a bulk densityof eight lb/cu ft.

Duct Sampling Technique

The arrangement of the samplingports and equipment is shown in Fig. 2.A y4-inch diameter null balance sam-pling probe6 was used for both the inletand discharge ducts. The inlet ductsampling port was located 10 diametersfrom the entrance of the duct and threediameters from the entrance to thescrubber. The discharge duct samplingport was located in a straight duct sevenand a half diameters after the separator.A Millipore plastic field monitor, contain-ing an aerosol type membrane filter, wasattached to the end of the probe asshown at right-hand center portion ofthe picture. A 3/4-inch S&K air jetexhauster was used to draw the samplethrough the probe, Millipore filter, and

229

WATER IN

PRESSUREGAGE

AIR IN

BLOTTER IMPACT SURFACE

BLOTTER SUPPORT SWINGSSIDEWISE IN OR OUT OFCONTACT WITH DISCHARGE GAS

PRESSURE TAPFOR DRAFT/READING

I" VERTICAL OPENING

WATER OUT

Fig. 6. Liquid carry-over test set up S&K ejector venturi scrubber system.

collected. Prior to using the trans-parent Millipore field monitors, metalfilter holders were used. With metalfilter holders, the membrane filter hadto be removed to obtain the opticaldensity readings. This method pre-sented the possibility of losing some ofthe sample.

Evaluation of Particle SizeThe samples collected were evaluated

in three ways: (1) weight, (2) size anddistribution, and (8) optical density.

The weight of the sample collectedwas determined by standard laboratoryprocedures using an analytical balancehaving a sensitivity of 0.1 mg. Theinlet samples collected were in the rangeof 0.2 to 0.4 gram and the correspond-ing sampling time varied between two to10 minutes. The discharge samplecollected was in the range of 0 to 40milligrams and the sampling time was aslong as 24 minutes.

The particle size and number wasdetermined by use of a vertical projec-tion, light field microscope, shown inFig. 9. This microscope was madeespecially for dust counting work. Itsmajor advantage over other types ofdust counting microscope equipment isthat it can be used in a manner similar

diy gas meter. The null balance condi-tion was indicated on an inclinedmanometer and was controlled by avalve on the suction line to the air jetexhauster. The samples were takenwith the probe mounted at the pointof average velocity. This was done inorder to keep the amount of workinvolved to evaluate a test run to a

IOO

f-2 90

or —

>58OzUJ

oU_U.UJ_ 7 0

O 60O

minimum.The Millipore plastic field monitors

were chosen in favor of other methodsbecause of the handling convenience,versatility, and low cost in our applica-tion. The transparent monitors per-mitted us to watch the buildup of par-ticulate on the filter, and stop thesampling when a sufficient quantity was

SCHUTTE a KOERTIN6 CO.CORNWELUS- HEIGHTS

BUCKS CO..PENNA.

35 405 10 15 20 25 30FART1CLE SIZE-MICRONS

Fig. 7. Fractional efficiency curve for a four-inch figure 4014 S&K ejector venturi scrubbing system.

230

Fig. 8. Aerosol generator.

Journal of the Air Pollution Control Association

Fig. 9. Leitz vertical projection dust countingmicroscope.

to nn ordinary photographic onlargor.This feature permits us to make fullsize eight-inch diameter prints of theprojected view, as shown in Fig. 10.Larger or smaller prints could be madeif desired by adjusting the projectiondistance. This photographic procedurepermitted us to easily make permanentrecords and do the dust counting at anyconvenient time and under much lessfatiguing conditions.

The highest degree of magnificationwe used was 1000 X, and the totalmagnification due to projection was1200X, when working with an eightinch view. At this degree of magnifica-tion, the 100 watt light source was

sufficient to permit viewing in a normallylighted room.

Two methods were used for preparingthe specimens for microscopic examina-tion. For light loading, the membraneniters were used directly by mountingon a glass slide and rendering themtransparent with immersion oil. Forheavy loading, the membrane filter wasdissolved in acetone and small portionplaced in a counting cell.

The distribution was determined byphotographing five areas of each micro-scope specimen and then counting andsizing up to a total of approximately500 particles. The number of particlesthat should be counted in any particularsample is dependent on the width of itsdistribution and the statistical level ofconfidence desired.

The optical density was determinedfor each sample by light comparisonreading of the membrane filter beforeand after sampling. The apparatusused consisted of a photoelectric cell, ameter calibrated in candlepower, and alight source whose intensity was variedby a variable transformer. The trans-parent Millipore field monitors permitus to take optical density readings ofthe membrane filter while mounted inthe monitor. A correlation betweendust loading and optical density wasobtained but due to the limited amountof data collected to date we do not feelwarranted publishing these results.7

REFERENCES

1. "NAFM Standards, Definitions, Terms,and Test Codes for Centrifugal, Axial,and Propeller Fans." Bulletin #110

Fig. 10. Typical microscopic record at 5 2 5 X .

National Association of Fan Manu-facturers, Inc., 19.

2. H. S. Bean, R. M. Johnson, and T. H.Blakeslee. "Small Nozzles at LowValues of the Diameter Ratio." ASMETransactions, 76: 833-871 (August1954).

3. Schutte and Koerting Company, "AirJet Syphons," Bulletin 2A, October1954, 3 and 15.

4. M. W. First, et al., Air CleaningStudies, U. S. Atomic Energy Com-mission, Progress Report (NYO-1581)(June 30, 1951).

5. E. I. duPont de Nemours & Company,"duPont Petroleum Dyes," BulletinA-2723, 11-56-10, p 12.

6. Western Precipitation Corporation,"Methods for Determination of Veloc-ity, Volume, Dust, and Mist Contentof Gases," Bulletin WP-50, p 24.

7. Heat Exchange Institute, "Standardsfor Steam Jet Ejectors," Third Edition,18-28(1956).

8. R. L. Webb, The Draft ProducingVenturi Scrubber, Proceedings of APCASemi-Annual Technical Meeting, 35-41(December 1956).

Index to Advertisers

Academic Press, Inc Inside Front Cover

Buffalo Forge Company Outside Back Cover

Melvin F. Hall, Inc., Buffalo

W. L Faith 256

Hemeon Associates 255

Lauren B. Hitchcock Associates 259

William T. Ingram 257

Jackson & Moreland, Inc 216

Resources Research, Inc. 216

Staplex Company 215

R. E. McGuire Associates, New York

Travelers Research Center, Inc 216

E. K. von Brand 216

Western Precipitation 222Hixson & Jorgensen, Inc., Los Angeles

John Wood Company Inside Back CoverRay Ellis Advertising, Inc.,Chatham, New Jersey

May 1962 / Volume 1 2, No. 5 231

Indian Journal of Science and Technology Vol. 3 No. 1 (Jan 2010) ISSN: 0974- 6846

Research article Biomass energy Sivakumar & Krishna Mohan Indian Society for Education and Environment (iSee) http://www.indjst.org Indian J.Sci.Technol.

58

Performance analysis of downdraft gasifier for agriwaste biomass materials

K. Sivakumar and N. Krishna Mohan

Dept. of Mechanical Engineering, Annamalai University, Annamalai Nagar, Chidambaram-608 002, Tamil Nadu, India

[email protected]

Abstract In this study, combustible gas production from shell biomass materials such as coconut shell, groundnut shell and rice husk was experimentally investigated at 8000C using gasification technique by a downdraft gasifier. The combustible gases H2, CO, CH4, CO2, and N2 were formed. The calorific value of the producer gas for various shell materials was found. The calorific value of coconut shell is 23.01% higher than ground nut shell and 45.45% higher than rice husk. Based on performance analysis, it is concluded that coconut shell is the best suitable fuel for gasifier compared to the other two biomass shell materials.

Keywords: Biomass, gasification, coconut shell, groundnut shell, rice husk, energy from waste. Introduction

Biomass is recognized as one of the major potential sources for energy production. Biomass resources viz. forestry residues, energy crops, manufacturing food waste, coconut shell, bagasse from sugarcane processes and food processing residue all have been used for energy generation.

Biomass gasification, the complete conversion of biomass to a gaseous fuel by heating it with a gasification medium such as air, oxygen or steam is fast becoming the most promising process for electricity generation. Biomass gasification products have been demonstrated for the generation of electricity via boilers and steam turbines or internal combustion engines, Fired in gas turbines and even proven suitable for some types of fuel cell.

Combustible gases viz. H2, CO, CH4, CO2, and N2

were formed in the reactor (Adnan Midilli et al., 2001) reported the cashew nut as biomass fuel).

Shell biomass materials maintains net calorific values between 6000 K.cal/kg and 3000K.cal/kg. This is almost half that of natural gas and fuel oil. This coupled with the low mass density of biomass means it has a low energy density compared to fossil fuels and therefore shell bio mass materials are suitable for small scale plants (Savitri Garivait & Ulonwan Chaiyo, 2006). There are various biomass shell materials are available in tropical countries. Where in the following three bio mass materials which are commonly available and reliable were taken for performance analysis (Singh et al., 2006): Biomass fuels

1) Coconut shell was used as the feedstock with size range 0.75-1.0 mm. The proximate and ultimate analysis of biomass as follows: moisture 10.53%,fixed carbon 13.10%,volatile matter 57.96% and ash 18.4%.The

ultimate analysis was C 50.2 %,H 5.30%, N 0.0% and O 43.4%.

2) Ground nut shell has great potential for commercial use. It is used as fuel filler in cattle feed, hard particleboard, cork substitute, activated carbon etc. The Proximate analysis reveals that moisture 8.76%, fixed carbon 15.50%, volatile matter 54.96 and Ash 20.3%. The ultimate analysis shows that C 48.3 %, H5.70%, N 0.8 % and O 39.4%.

3) Rice husk is recognized as an important source of energy, particularly in developing countries where it economies largely based on agriculture and forestry. Agriculture waste is one form of biomass which is readily available but is largely not utilized in energy recovery schemes. Rice husks are an agriculture waste produce as a by-product of the rice milling industry. The world wide annual production of rice husks is estimated to be about 100 million tones, 90% of which generated in developing countries. The proximate analysis and ultimate analysis of biomass as follows: moisture 7.9%, volatile matter 59.5%, fixed carbon 19.9% and ash 17.1%. The ultimate analysis was C 38.9%, H 5.1%,N 0.6%,O 32.0% (Van der Drift et al., 2001). Experimental setup and procedure

The prototype small-scale biomass gasifier system comprised an agri-waste feed hopper and feed auger, air blower, cyclonic gasifier, external cyclone, cooler/condenser and engine (Fig.1). Equipment The biomass gasification system developed within this project was a swirling flow gasifier. The concept of this entrained flow gasifier used air to entrain shell biomass in a turbulent vortex within the reactor, which incorporated two stages of separation to remove the char and ash produced in the process. The intense continuous



59

Fig.1. Schematic diagram of biomass gasifier

reaction enabled gasification of high volumes of biomass in the compact reactor.

The system used for this work was sized to gasify up

to 20kg/hr of shell biomass under atmospheric pressure conditions; This gives a total energy input to the gasifier of around 20kW. The system was tested over a range fuel input rates from 12kg/hr up to 15kg/hr. For the final testing with engine operation the gasifier was turned down to half its peak capacity of 20kg/hr, results shown in this report reflect the performance of the gasifier in turn down operation (Philippe Mathieu & Raphael Dubuisson, 2002). The final prototype gasifier was manufactured in 316-grade stainless steel and proved suitable when tested under continuous high temperature 800°C operation conditions. Khater et al. (1992) discussed behavior of downdraft gasifier and maximum temperature range, Pattaraporn Chaiprasert and Tharapong Vitidsant (2009) showed that increasing carbon conversion to gas from 44.13-78.43% where as tar was decreased from 19.55-1.4% at temperature of 800ºC. Heat-up gas burner

A burner system was designed to supply and ignite a supplementary gas to fire through the gasifier to heat it up to temperatures to initiate gasification reaction. Specification temperatures were up to NOT for the reaction initiation. The design also ensured no interruption of the gasification reaction and not be fouled by biomass particles. Hopper/feeder system The biomass hopper and feed system have a small hopper with agitator provision. The hopper was small, however provided shell biomass loading of 15kg of feedstock, which under most test conditions was suitable but needed refilling and agitation was required for longer

tests. Limited and controlled amount of air for partial combustion enters through the air nozzles (Fig.2).

Cooler/condenser

An array of 'U' tube design condensers were used to cool the Shell gas and condense any water and tars from the gas. The cooler was designed to reduce shell gas from temperatures of 7000C - 8000C down to 300C - 400C, so that the maximum volumetric energy content of the shell gas is achieved. The design factored in minimum pressure drop across the system. Dogru et al. (2002) concluded that, the pressure drop is directly proportional to the air flow

rate for all bio mass. Filter

The cool gas was filtered in various filter systems such as in fine filters and cloth bag filters to evaluate the performance of the upstream separation systems. Experimental procedure

In the down draft gasifier, both the fuel and the gas flow downwards through the reactor enabling the pyrolysis gases to pass through a throated hot bed of char which is supported by a grate. This results in cracking of most of the tars in to non condensable gases

and water. Furthermore, air is admitted to the fuel bed through air intake nozzles causing pyrolysis to charcoal and volatiles that partially burn as they are produced. The gaseous products of this flaming pyrolytic combustion then consume the charcoal produced during the pyrolysis and are reduced to fuel gas. In this way, tar vapors are typically lowered to 0.1% of the total feed. Whereas, in draft or cross flow gasifier, tar levels are higher than 0.1%. Paulo R. Wander and Carlos (2004)

reported the various level of tar gasification. Avdhesh and Sharma (2009) revealed the characteristic parameters of gasifier. This gas is used for operation of, internal combustion engine. Results and discussion

The gasification experiments were performed for three agriwaste biomass materials. The volumetric flow rate of wet product gas was measured with a gas flow meter. Drying, pyrolysis and oxidation zone temperature of the gasifier and the water scrubber and box filter outlet temperature were also monitored with the aid of R and K

Fig.2. Down draft biomass gasifier plant



60

type thermocouples. The amounts of tar-dust and condensate in the product gas were determined before and after cleaning the product gas. The combustible gases from the shell biomass were obtained as H2, CO, CO2, CH4, and N2 and their amounts in the dry product gas were estimated in units of kg h-1 and as volume by volume percentage.

The calorific value and total energy of the produced gases were determined and evaluated based on the dry product gas volumetric flow at STP hot cold and raw gas efficiencies of the dry product gas were estimated in order to investigate its usage in the CHP engines. During the experiment, pressure drops were measured as 0.10- 0.80 mm Hg at the gasifier outlet, 0.20 -1.80 mm Hg at the water scrubber outlet and 2.15 -4.60 mm Hg at the filter box outlet. It was found that they were small at the wet product gas flow rate of around 8.145 N m 3 h-1. Experiment was carried out at 800⁰C for coconut shell, ground nut shell and rice husk. The producer gas analysed and the results were plotted in the graph. The percentage of composition of H2, CO, CH4, CO2, and N2 are shown in Fig.3. The calorific values of the producer gas for these three materials were founded out (Table 1). Conclusion

Combustible gases can be produced from coconut shell, ground nut shell and rice husk; they were utilized as a feedstock in a down draft gasifier. The calorific value of the coconut shell is more than groundnut shell and rice husk. The coconut shell has more carbon content and producer gas.

The hydrogen amount in the producer gas for ground nut shell is more than coconut shell and rice husk. But carbon monoxide is 17.55% and 21.22% higher than ground nut shell and Rice husk in Coconut shell.. Also 6.15% and 38.71% methane are highly present in coconut shell than the other two shells. The coconut shell have more carbon content and also in producer gas.

Coconut shell is one of the waste biomass and easily available material. It is the best alternative energy source shell material. References 1. Adnan Midilli, Murat Dogru, Colin R, Howarth Mike,

Ling J and Teoman Ayhan (2001) Combustible gas production from sewage sludge with a down draft gasifier. Energy Conservation & Management; 42,157-172.

2. Avdhesh K and Sharma R (2009) Experimental study on 75 KW the downdraft (biomass) gasifier system. Renewable Energy. 34, 1726-1733.

3. Dogru M, Howarth CR, Akay G, Keskinler B and Malik AA (2002) Gasification of hazelnut shells in a down draft gasifier. Energy. 27, 415-427.

4. Khater EMH, El-Ibiary NN, Khattab IA and Hamad MA (1992) Gasification of rice hulls. Biomass & Bioenergy. 3, 329-333.

5. Pattarporn Chairprasert and Tharapong Vitidsant (2009) Promotion of coconut shell gasification by steam reforming on nickle-dolomite. Am. J. Appl. Sci. 6(2), 332-336.

6. Paulo R. Wander and Carlos R (2004) Assessment of a small saw dust gasification unit. Bio mass & Bio Energy. 27, 467-476.

7. Philippe Mathew and Raphael Dubuisson (2002) Performance analysis of a biomass gasfier. Energy Conversion & Management. 43, 1291-1299.

8. Savitri Garivait and Ulonwan Chaiyo (2006) Physical and chemical properties of Thai biomass fuels from agricultural residues. 2nd Intl. Conf. on Sustainable Energy & Environment.

9. Singh RN, Jena U, Patel JB and Sharma AM (2006) Feasibility study of cashew nut shells as an open core gasifier feedstock. Renewable Energy. 31,481-487.

10. Van der Drift A, Van Doorm J and Vermeulen JW (2001) Ten residual biomass fuels circulating fluidized-bed gasification. Biomass & Bio-energy. 20, 45-56.

Table 1. Calorific value of various biomass materials

Shell Materials Calorific value Kcal/Kg

Coconut Shell 5500 Groundnut Shell 4229 Rice Husk 3000

Table 1. Calorific value of various biomass materials

Shell Materials Calorific value Kcal/Kg

Coconut Shell 5500 Groundnut Shell 4229 Rice Husk 3000

Fig.3. Gas composition of various biomass materials

SUPERFICIAL VELOCITY - THE KEY TO DOWNDRAFT GASIFICATION1 T. B. Reeda, R. Waltb, S. Ellisc, A. Dasd, S. Deutche

a The Biomass Energy Foundation, 1810 Smith Rd., Golden, CO 80401; b Community Power Corporation, Aurora, CO, c Colorado School of Mines, Golden, CO, d Original Sources, Boulder, CO, e The National Renewable Energy Laboratory, Golden, CO.

ABSTRACT

The “superficial velocity” (hearth load) of a gasifier is the most important measure of its performance, controlling gas production rate, gas energy content, fuel consumption rate, power output, and char and tar production rate.

The superficial velocity, SV, of a gasifier is defined as:

SV = Gas Production Rate/Cross Sectional Area = (m3/s)/(m2/s) = m/s

It is easily estimated or measured by measuring gas production rate or fuel throughput and gasifier dimensions. It controls the rate at which air, then gas, passes down through a gasifier. This in turn exercises a primary effect on heat transfer around each particle during flaming pyrolysis of the volatiles, combustion of the tars and gasification of the charcoal.

A low SV causes relatively slow pyrolysis conditions at around 600°C, and produces high yields of charcoal - 20-30%, large quantities of unburned tars, and a gas with high hydrocarbon content and high tar (volatile) content. A high SV causes very fast pyrolysis, producing less than 10% char-ash at 1050 C and hot gases at 1200-1400 C in the flaming pyrolysis zone. These gases then react with the remaining char-ash to yield tars typically less than 1000 ppm, 5-7% char-ash and a producer gas with less energy.

These relationships have been investigated in a velocity controlled inverted downdraft gasifier with a 7.5 cm diameter. As the superficial velocity was varied from 0.05 m/s to 0.26m/s , the gas production rate increased from 102 to 679 cm3/s, charcoal production decreased from 13.0% to 4.7% and tar in the gas decreased from 8330 to 300 mg/kg (ppm).

At low Superficial Velocities (and low Biot numbers), the particles are heated slowly to pyrolysis temperature and remain essentially isothermal. At high superficial velocities the outside of the particle can be incandescent (> 800°C) while the center is still at room temperature. This permits the escaping gases to react with the charcoal, thus reducing the charcoal yield and increasing the gas yield. We call this phenomenon “simultaneous pyrolysis and gasification”, SPG and believe that it is the fundamental reason why the Superficial Velocity controls all other aspects of gasification.

In producing heat the “tars” in producer gas are a useful fuel, providing no cold surfaces intervene. The “inverted downdraft” stoves provide potentially simple, clean cooking for developing countries. At low SVs they also produce charcoal. In producing electric power, tars are detrimental to engine operation, and so high SVs must be maintained to minimize tar and char production. *Presented at 4th Biomass Conference of the Americas; Oakland, CA, 8/29/99

SUPERFICIAL VELOCITY

Introduction

Clean cooking and distributed power are the major problems of half the world population. Gasification of biomass can help provide solutions to both these problems; unfortunately very little fundamental research has been done on gasification and most new designs are made “by guess and by golly”. Over a million vehicles used gasified biomass successfully during World War II. Since then there have been many attempts to gasify biomass, most unsuccessful.

Biomass gasifiers can be divided into three main categories depending on the source of heat for pyrolysis: Charcoal burning gasifiers (updraft, counter-flow …); tar burning gasifiers (downdraft, inverted downdraft, crossdraft, co-flow, open top, topless, …); and fluidized bed (with some char, some tar burning and many varieties). Due to the very high volatile content of biomass, the tar burning gasifiers are preferred when it is necessary to produce a very clean gas for power generation. This paper deals primarily with tar burning gasifiers, but many of the principles can be applied to other forms.

The “superficial velocity”, SV, of a gasifier is the most fundamental measure of the expected behavior of a gasifier and controls most of the other aspects of gasifier operation as shown here:

Gas Production Rate Char Production

Fuel Consumption Rate

Gas Energy Content Tar Production

It is as important to gasifiers as engine rpm to an engine designer or metabolism rate to a physiologist.

It is defined as:

SV = Gas Production Rate/Cross Sectional Area = (Nm3/s)/(m2/s) = m/s

(It is sometimes defined as the “Hearth load”, Bh measured in Nm3/cm2hr. Bh = 0.36 SV.)

It is independent of gasifier size, and so permits comparison of gasifiers of very different dimensions. It is easily estimated or measured from air/oxygen fuel throughput and gasifier dimensions. It controls the rate at which air, then gas, passes through a gasifier. This in turn exercises a primary effect on heat transfer around each particle during flaming pyrolysis of the volatiles, on combustion of the tars and the on the degree of gasification of the charcoal. While the SV is the most important measure of performance for design, it is not generally recognized or discussed.

1. HISTORICAL In the book, “Gengas”1 G. V. Nordenswan says:

“The concept of hearth load plays a very important role in dimensioning a wood gas generator hearth. The hearth load is the quantity of prepared gas, … divided by the smallest

passage area of the hearth. Thus the hearth load is dimensionally a velocity…”. He goes on to say that the practical range for SV is 0.8m/s (where tars are quite high) to 2.5 m/s (above which charcoal dusting is unacceptable). These numbers only apply to constricted hearth (Imbert) gasifiers. Table 5-2 in Ref. 2 gives design parameters for various Imbert gasifiers ranging from 4 Nm3/hr to 230 Nm3/hr.

More recent gasifiers would have other values of SV as shown in Table I.

Table I – Reported Superficial Velocities in various gasifiers2

GASIFIER TYPE

PYROL. ZONE SV m/s

CHAR ZONE SV m/s

Imbert 0.63 2.5

Biomass Corp. 0.24 0.95

SERI Air 0.28 0.28

SERI Oxygen 0.24 0.24

Syn-Gas air 1.71. 1.71.

Syn-Gas oxygen 1.07 1.07

Buck Rogers (Chern)

0.23 0.23

Buck Rogers (Wallawender)

0.13 0.13

The values in Table I are taken from literature reports and generally represent the highest values reported there. The Stratified Downdraft gasifier has the same values of SV in the pyrolysis and char zones since it uses a constant diameter gasification tube. Note that the values differ widely from those recommended for Imbert gasifiers.

2. EXPERIMENTAL

4.1 Apparatus The apparatus shown in Figure 1 has been built to measure the superficial velocity over a

wide range in batch inverted downdraft gasification. (It can be configured with minor modifications to operate in the conventional downdraft mode.) The apparatus is well insulated with pressed fiber refractory (riser sleeve) to minimize heat loss. The air supply is measured with a calibrated flowmeter and the pressure drop, temperature, and gas composition, are measured during the runs.

In use the cylinder is filled with fuel to the gas outlet. The runs reported here were made on mixed +1/4, -1/2 inch hardwood chips with a moisture content of 6.1% wet basis. The top

chips are saturated with alcohol and ignited and the cap is put on. The temperature and pressure drop are monitored as a function of time.

The gas from the gasifier passes through the collection train shown in Figure 2. The outlet pipe is heated with a torch to prevent tar condensation at startup. The gas then passes through a condenser in an ice bath to cool the gas and remove most of the moisture and tar. At the end of the run the moisture recovered was measured and the condenser was rinsed with acetone to collect the majority of the tar. Residual moisture and tar can be removed after the condenser and total gas flow can be measured with a gas meter at the exit and the gas can be flared.

Tars were also measured using a filter to draw aliquot portions directly from the gasifier during the runs. These results and new methods will be reported later. After the run the amount of charcoal remaining was measured.

While the experiments were performed on an “inverted downdraft” gasifier, it is believed that the results will be similar for conventional and new downdraft gasifiers. Further experiments are planned.

3. RESULTS The results of these measurements at four SVs are shown in Fig.

3 and reported in Table II

4.1 Flaming Pyrolysis Tar burning gasifiers consume the tars by a process we call

“flaming pyrolysis.2 The combustion of pyrolysis products in air in sufficient supply (as in a match) is termed “flaming combustion”. We apply the term “flaming pyrolysis” to the combustion of the same volatiles in an inadequate supply of air so that the products are largely CO and H2 rather than CO2 and H2O.

Fig. 2 - Gas Train

Cap

Grate

Air In

Temperature

UnburnedFuel

Char fromPyrolysis

Pressure

Gas Sample,Pressure

Gas Out

4 " Pipe

PyrolysisZone

Insulation

30 cm

10 cm5 cm

40 cm

Figure 1 – Gasifier for measuring effects of superficial velocity change.

G a s i f ie r

T e m p e r a tu r e

G a s S a m p le ,P r e s s u r e

P r e s s u r eT a p

G a s O u t

0 C Ic e B a thC o m p r e s s e d

A ir , F lo w m e t e r

F la r e ,E S P ,

F lo w m e t e r

4.2 Charcoal and Tar Yield It is seen in Table II and Figures 4 and 5 that charcoal and tar yields depend strongly on

the superficial velocity. The reduction in charcoal and tar with increasing SV is due to the increased temperature during pyrolysis shown in Figure 3 and Table II. Tars were measured both by total condensate in the condenser and by use of an absolute filter method.3

4.3 Air/Fuel ratio It is obvious in Table II that many other variables depend on SV. In Table II it can be seen

that the air/fuel ratio increases from 1.44 to 5.21 as more and more of the charcoal is gasified and the process approaches complete combustion.

Table II – Results of operation of gasifier at 0.05, 0.19, 0.26 and 0.44 m/sec SV Run # 5 6 8 9MEASURED VALUES Air Flow - l/m 4.1 18.8 26.0 46.4Initial Fuel (g) 89.1 91.1 82.2 84.2Final Charcoal (g) 11.6 6.8 3.8 0Condensate (g) 16.7 13.8 11.8 13.8

Tar + Particulate – (g) 2.044 1.463 0.44 0.189

Average Pressure Drop – iwc 0.025 0.11 0.15 0.33Time of Run – min 33.5 14.4 9.5 8Pyrolysis Temp 770 993 1033 1045Pyrolysis time – min 1.0 1.0 0.6 0.6

DERIVED VALUES Superficial Velocity - m/s 0.052 0.187 0.260 0.437Gas Produced – kg (m3) 0.206 0.423 0.387 0.548Gas production rate – cm3/sec 102 489 679 1141Tar, Particulate in Gas - mg/m3 9941 3460 1138 346Tar – mg/m3 -Filter Method 8330 1330 300 10Fuel Velocity - cm/min 1.20 2.08 3.16 3.75Air/Fuel Ratio 1.44 3.80 3.47 5.21Charcoal Yield - % 13.0% 7.5% 4.7% 0.0%

4.4 Gas Quality The gases have not been analyzed.

However, it is expected that the energy content will be very high at low superficial velocities (pyrolytic gas) and progress to lower values as more and more of the charcoal is consumed.

4.5 Reaction velocity The flaming pyrolysis gasification

reaction in solids is analogous to the combustion of a gas and proceeds at a finite rate. In steady state gasifiers the fuel is consumed at the same rate as the reaction progresses, so fuel flow velocity is identical to reaction velocity. In the batch gasifier studied here the gasification reaction proceeds through the solid bed of fuel and the time required to pyrolyse all the fuel gives the reaction velocities of 1.2 to 3.8 cm/min.

4.6 Gasification-Combustion Gasification can be viewed as a partial combustion reaction, and the declining energy

content of the gas and the increasing air fuel ratio suggests that there may be a continuum

between pure pyrolytic gasification, producing gas and charcoal, and pure combustion at high enough air/fuel ratios.

Run 9 produced no charcoal and was probably close to complete combustion. Gas measurements are in progress and will be reported separately.

4.7 Time for Pyrolysis The time required for pyrolysis to charcoal plus volatiles depends on the heat transfer rate

and the amount of charcoal produced.

The time for pyrolysis shown in Table II is taken as the time necessary for the first rapid temperature rise shown in Figure 3. It is remarkably short, 0.6 to 1 minute for these chips. We were surprised that there was not a steeper rise time at higher superficial velocities. We believe this can be explained by the less complete pyrolysis at lower velocities.

Char Yield vs Superficial Velocity

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

0.00 0.10 0.20 0.30 0.40 0.50

Superficial Velocity - m/s

Cha

r yie

ld -

%

Tar vs Superficial Velocity

0

2000

4000

6000

8000

10000

12000

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

Superficial Velocity - m/s

Tar,

Part

icul

ate

mg/

m3

Tar+Particulate in Condensor

Tar only, Filter method

Pyrolysis Time vs Superficial Velocity

0

200

400

600

800

1000

1200

0 5 10 15 20 25 30Time - min

Tem

pera

ture

- C

SV= 0.26

SV=0.44

SV=0.19 SV=0.05

Figure 3 – Temperature Vs time for 4 sample runs

Figure 4- Charcoal yield Vs SV Figure 5 – Charcoal yield Vs SV

4. SIMULTANEOUS PYROLYSIS AND GASIFICATION, SPG In an earlier paper4 we described the rapid pyrolysis of single particle birch dowels 3/8-3/4

in diameter in a high temperature gas flame. We noticed that the charcoal yield was 3.1-11.7% rather than the 20-25% yield characteristic of slow pyrolysis. We also measured a heat for pyrolysis of 2.9-3.4 kJ/g, much higher than the 1.8-3 kJ/g required for slow pyrolysis.

We now believe that this is due to “simultaneous pyrolysis and gasification”, SPG. In slow pyrolysis the particle remains close to isothermal and the surface temperature does not exceed 400-700°C, not high enough for char gasification reactions. However, with fast pyrolysis at higher heat transfer rates, the particle surface temperature can exceed 1000 °C while pyrolysis is continuing at the cool core. In this case the exiting gases and volatiles must pass through the surface charcoal and can react to gasify the charcoal and crack the volatiles. Since these are endothermic reactions, it is not surprising that the heat for pyrolysis is so high. We believe SPG is the root cause of reduced tar and charcoal at high superficial velocities. It also explains the success of downdraft gasification in eliminating tars.

5. IMPLICATIONS FOR GASIFIERS FOR POWER AND STOVES We said at the beginning of this paper that clean cooking and power were the primary needs of half the world. A great deal of progress is being made today in gasification for both stoves and power. Much of this work is reported on the Internet. In particular there are continuing forums on both stoves and gasification at www.crest.org.

4.1 Gasifier Stoves Recently there has been a great deal of interest in cooking for developing countries. Wood

cooking is inherently inefficient and dirty, but wood-gas can be burned like propane or natural gas to give fast, clean cooking. The inverted downdraft gasifier is a batch process in which a vessel of biomass is ignited at the top and produces a gas as the pyrolysis reaction proceeds to the bottom. This gas can then be mixed with air and burned much more cleanly than in conventional wood cooking. This can be operated with natural draft 5 or forced draft 6. The apparatus described here is useful for characterizing biomass gasification at these very low SVs.

4.2 Power Gasifiers The operation, cost and complexity of power gasifiers is very dependent on the tar

produced. It is obvious from Figures 4 & 5 that one should operate a gasifier at as high a superficial velocity as possible, consistent with producing a combustible gas, corresponding to 0.26m/s on Figure 3 but not as high as 0.44m/s for these wood chips. We plan to operate this gasifier with different fuels & moisture contents to characterize the optimal velocities for producing gas. Power gasifiers are usually operated in the downdraft mode, but can also be

operated tar-burning updraft.7 The apparatus described here is useful for investigating the various factors affecting power gasifiers at high SV.

6. ACKNOWLEDGEMENTS The authors would like to thank their many friends at the Colorado School of Mines and

the National Renewable Energy Laboratory who cooperated to help make this paper possible. The work was funded by the Community Power Corporation and the Biomass Energy Foundation.

REFERENCES 1. “Generator Gas: The Swedish Experience with Wood-gas 1939-1945” 3rd Edition, Reed,

T. B. and Jantzen, D., Eds., The Biomass Energy Foundation Press, 1998.

2. Reed, T. B. and Das, A., “Handbook of Biomass Gasifier Engine Systems”, The Biomass Energy Foundation Press, 1988.

3. Das, A., “Contaminant Testing for Gasifier Engine Systems” 2nd edition, Original Sources, The Biomass Energy Foundation Press Press, 1999.

4. Reed, T. B. and S. Gaur, “The High Heat of Fast Pyrolysis for Large Particles”, in Developments in Thermochemical Biomass Conversion, Ed. AA. V. Bridgwater, Blackie Academic Press, 1996.

5. Reed, T. B. and Larson, R., “A wood-Gas Stove for Developing Countries”, in Developments in Thermochemical Biomass Conversion, Ed. A. V. Bridgwater, Blackie Academic Press, 1996.

6. Reed, T. B. and Walt, R., “The “Turbo” Wood-Gas Stove”, paper, this conference.

7. Brand, Stephen, “Multi-Fuel Gasification for Energy Production in Rural Areas”, in Proceedings, 2nd Biomass Conference of the Americas, Aug. 21, Portland Ore., the National Renewable Energy Laboratory, NREL/CP-200-8098, 1995.

Design and Development of Sand Bed Filter

Documents

Transcript of Design and Development of Sand Bed Filter