2001-01-0517

Combined Silencers and Urea-SCR Systems for Heavy-dutyDiesel Vehicles for OEM and Retrofit Markets

Pär Gabrielsson, Ioannis Gekas and Peter SchoubyeHaldor Topsøe A/S

Søren Æ. Mikkelsen and Svend FrederiksenSilentor A/S

Copyright © 2001 Society of Automotive Engineers, Inc

Reprinted with permission from SAE International. This paper may not be photocopied, e-mailed or otherwise reproduced without permission of SAE International.

ABSTRACT

Selective Catalytic Reduction (SCR) with NH3 or urea isone of the most effective methods for removal of NOx inexhaust from HD diesel engines with potential forachieving more than 90% NOx-reduction measured inthe European transient or a US HD FTP test cycle. Thepresent paper describes the following two systems;

• One OEM UREA-SCR SILENCER, comprising asilencer with built-in catalyst. The system was testedon a Scania DC1205 320 kW diesel engine, whichwas calibrated for the Euro II emission standard. Thetest results showed that it is possible to reduce morethan 85% of the NOx emission with an insignificantNH3 slip in the ETC transient test cycle. Thepressure-drop of the system was measured at 80%of that of the engine’s original silencer and thesilencing performance was improved for lowfrequencies below 125 Hz.

• One RETROFIT UREA-SCR SYSTEM for HDengines, comprising a silencer with built-in catalyst,an electronic urea injection control system, ureainjection and a urea tank. The importance of injectionstrategy and the relation between urea/NOx ratio,ammonia slip and obtained NOx-removal waselucidated by transient tests with a 10 liter 190 kWVolvo HD excavator engine on test rig. NOx-removals of 77 to 88% were demonstrated in ETCand the European ISO 8178C1 off road test cycleswith essentially no ammonia slip. 99.7% NOxreduction was obtained with a high NH3-slip of about100 ppm, illustrating the potential of the urea SCRtechnology. The silencer contained 29 liter catalyst.

INTRODUCTION

Exhaust after-treatment on diesel vehicles will beintroduced and become mandatory in the coming years.

In the European Union legislators have decided on verystringent emission standards regarding both dieselparticulate and NOx emissions for heavy-duty dieselvehicles. The new standards, Euro IV and Euro V, will beintroduced year 2005, respectively 2008. A new type ofemission standard called EEV, EnhancedEnvironmentally friendly Vehicles, will also be introducedin the EU, aiming to promote new technologies able tomeet more stringent future emission standards prior totheir introduction. The driving force will be a tax reductionfor such vehicles if the Euro V emissions standards arefulfilled before year 2008.

The Euro IV standard can probably be met by usingexhaust gas re-circulation and particulate traps. TheEuro V and EEV standards will require use of particulatetrap and Urea-SCR system. Alternatively the Euro Vemission standard may be met by a fuel optimizedengine with high NOx emissions in combination with aUrea-SCR system and without PM-trap [1].

EPA and California ARB are proposing emissionstandards for year 2007, which, as in Europe, willnecessitate exhaust after-treatment like Urea-SCRcatalysts and particulate filters in diesel vehicles.

SCR with urea or ammonia is one promising method forreduction of NOx onboard diesel vehicles. The use ofSCR was already investigated during the 1980´ties andin the beginning of the 1990´ties. Held et al [2] suggestedthe use of urea to replace the more harmful ammonia.Walker et al [3] proposed a technology for injectioncontrol using look up tables and Andersson et al [4]proposed a more mathematical approach using kineticcalculations for control of the reducing agent. Themathematical approach by Andersson et al was alsoproved in transient tests on HD diesel engines in US HDFTP test cycle. Work by Havenith et al [5] confirmed thepotential for the method onboard diesel vehicles andlately the durability of a Urea-SCR system has beenshown by Fritz et al [6].

GOAL FOR UREA SCR OEM SILENCER - The goal forthe OEM silencer project was to demonstrate what couldbe achieved in a Urea-SCR silencer regarding NOxreduction, pressure drop, sound attenuation and spaceon Heavy-duty diesel vehicles.

The engine system used for this work was a 12-liter, 320kW HD diesel engine.

GOAL FOR UREA SCR RETROFIT SYSTEM - Theretrofit Urea-SCR system was developed to meet thedemand for lowered NOx emissions from the existingfleet of diesel vehicles. The goal of the development hasbeen to make a Urea-SCR system, which should workwith all types of vehicles, without making any emissionmeasurements or mapping of the specific engine orvehicle.

UREA-SCR RETROFIT SYSTEM DESCRIPTION – Forthe reasons described above, the Urea-SCR retrofitsystem is based on three different types of sensorsenabling an electronic control unit, ECU, to calculate thepossible NOx conversion of a given engine operation.The sensors enabling the calculations are an engine inletair massflow sensor, NOx-sensor upstream the catalystand temperature sensors upstream and downstream thecatalyst. The NOx sensor also measures the air fuel ratiomaking it possible to calculate the fuel flow [7], which isused to calculate the exhaust massflow. All calculationsare carried out in real time. The Retrofit Urea-SCRsystem is schematically described in Figure 1 below.

Pressurisedair

SCR catalyst

Air Temp in

Mass air flowNOx sensor

Engine intakeairUrea

tank

Urea-SCRECU CAN port

Temp out

Injection valve control

Figure 1. Urea-SCR retrofit system

SILENCER DESIGN FOR RETROFIT SYSTEM - TheSCR catalyst as well as the up-front flow distributing andmixing radial diffuser are contained within a compact unitwhich has been designed to provide a maximum of noiseattenuation, combined with low overall flow resistance.Special attention was hereby devoted to attenuation atlower sound frequencies. While the catalyst itself tosome extent serves to attenuate medium and higherfrequencies, it contributes rather little to attenuation atlow frequencies. Here, good noise reduction is insteadachieved by the use of patented (pending), advancedsilencer technology, which will become public in thefuture. The outer shape of the Retrofit Urea-SCRsilencer is seen in Figure 2. The dimensions of thesilencer are Ø 450 mm and length 700mm.

The urea solution is atomized with air and injected 1000mm upstream of a diffuser, which accomplishes a goodflow and urea distribution over the catalyst. The diffuseris a patented (pending) construction, which improves thesilencing performance of the reactor [8]. The flowdistribution over the catalyst is controlled by CFDcalculations. Figure 3 shows calculated flow vectors overthe diffusing element and in the reactor where thecatalyst is situated. Two sets of zones of circulating floware formed behind the diffuser and contribute to achievea good urea distribution before the catalyst and also agood flow distribution over the catalyst.

Figure 2. Urea-SCR Retrofit silencer with dimensions

Ø 450 mm and length 700 mm.

Catalyst

Figure 3. CFD model of the exhaust gas flow at the inletof the silencer.

OEM SILENCER DESIGN –The design targets for thecatalytic silencer design were:

• A catalytic-silencer which could replace the originalsilencer using the same space.

• Having the same or better sound attenuationcompared to the original silencer.

• No increase in pressure drop compared to theoriginal silencer.

• Evenly distributed exhaust gas flow over the catalyst.• Short distance between injection of urea solution and

the catalyst in order to avoid precipitation.• Good distribution of the urea solution over the

catalyst.

The concept was firstly evaluated using CFD calculationsand a much effort was put into the design of the inlet ofcatalyst. The design goal for the flow mal-distribution wasa maximum deviation of 10% at maximum flow. Table 1shows the flow distribution of different engine loads; idle,part and full load. The maximum maldistribution wascalculated to be 8.9% at full load and at a gas flow of10.6 m/s.

The outer shell of the silencer is seen in Figure 4. Thedimensions of the silencer are 550 x 550 x 850 mm .

Figure 4. OEM Urea-SCR silencer.

Table 1. Calculated malflow distribution at catalyst inletin different flow situations.

Engineoperation

Flow (m/s) Span (m/s) Mal

Idle 0.9 0.02 2.5%

Part load 3.9 0.2 4.0%

Full load 10.6 0.9 8.9%

EXPERIMENTAL

ENGINE USED FOR UREA-SCR RETROFIT SYSTEMTEST- The engine used was a Volvo TD100 KAE, a 9.6litre turbo charged, direct injected excavator dieselengine with inter-cooler. The engine was certifiedaccording to the ISO 8178 C1 emission standards giving8 g NOx/kWh, 0.28 g PM/kWh, 0.44 g HC /kWh and 1.3g CO /kWh. Its rated power was 190 kW at 1700 rpmand rated torque of 1360 Nm at 1000 rpm. The engine isseen in Figure 5.

Figure 5. VOLVO TD100 KAE diesel engine

ENGINE USED FOR UREA-SCR OEM SILENCERTEST - The engine used was a Scania DSC 12 05 L01,see Figure 6, with a maximum power of 420 Hp at 1700-1800 rpm and maximum torque of 1950 Nm at 1050-1400 rpm. The engine was calibrated for the Euro IIemission standard and certified by ECE-R49 test cycle.

FUEL QUALITY - The diesel fuel used was SwedishEnvironmental Class 1 having less than 10 ppm sulphurand less than 5 vol-% aromatics.

Figure 6. Scania DSC 12 05 L01 diesel engine

UREA-SCR SYSTEM TESTING - The development ofthe Retrofit and OEM systems was carried out atcompany test facilities. Verifying tests of NOx reductionefficiency of the complete Urea-SCR Retrofit systemwere also conducted at MTC´s (an independent testinstitute) certified test facilities in Sweden.

COMPANY TEST FACILITIES - The test facilities usedduring the testing of OEM Urea-Silencer and thedevelopment of the retrofit Urea-SCR system are shownin Figure 7.

Fuel

Engine

MK 1

MK 2Fuel meter

Catalyst Massflow

StandardNOxCOHCCO2

FTIRNH3

HNCON2O…

Urea tankUrea meter

Dynamo-

meter

Mini dilution+

TEOM 1105PM analyser

Torque

Rpm

Temp.

Exhaust

Figure 7. Schematic of test cell

The engines were in all cases connected to an electricdynamometer enabling the engines to be operated intransient test cycles.

A NOx analyser, Ecco Physics 700 EL ht chemi-luminescense measured NOx, a Thermo-FID measuredHC, a BINOS 100 IR measured CO and CO2 and anOxynos 100 measured for O2. A multi-componentGASMET FTIR analyser was used in parallel with theother instruments and measured for NO, NO2, NH3, N2O,CO and CO2.

All described analyses could be made up-stream anddown-stream the catalytic silencer and all gasmeasurements were made in raw exhaust.

The massflow in the exhaust was measured by means ofa Mass ProBar™, which uses the annubar principle andcalculates the massflow by compensating for thetemperature variations in the exhaust.

Urea and fuel flow was measured using corioliusmassflow meters from Micro Motion.

ETC TEST CYCLE - The transient measurements weremade in the newly proposed European transient test-cycle, ETC, that recently has been adopted in the EU forcertifying engines with after-treatment devices.

The test cycle consists of three different, 600 secondslong parts, see Table 2.

Table 2.European ETC transient test cycle

Traffic type Urbanstreets

Ruralroads

Motorway

Duration (sec) 0-600 601-1200 1201-1800

ISO 8178 C1 TEST CYCLE – This test cycle is used fortesting off road engines like construction wheel loaders,bulldozers, off-road trucks and excavators. The torque,engine speed and weighting factors for the cycle aregiven in Table 3.

Table 3.Weighting Factors of the ISO 8178 C1 Test Cycles

Torque, %

100 75 50 10 100 75 50 0

Speed

Rated speed Intermediate speed Idle

Weighting factors (%)

15 15 15 10 10 10 10 15

UREA-SCR CATALYST - The catalyst wasmanufactured as a corrugated ceramic structure. Thecatalytic material is based on metal oxides like Vanadiumand Tungsten. This corrugated catalyst combines a highmechanical strength like the one of wash-coatedcordierite with the high low temperature activity ofextruded catalysts in which all channel wall material iscatalytically active, as in the described catalyst. A pictureof the catalyst is seen in Figure 8.

UREA-SCR CATALYST IN RETROFIT SYSTEM - Thecatalyst system consists of a monolithic SCR catalystsolely. The size of the SCR catalyst was a Ø 387 x L 245mm block (28,8 liters) with a cell density of approximately170 cpsi. The volume corresponded to 3 litre of catalystper engine cylinder litre.

UREA-SCR CATALYST IN OEM SYSTEM - Thecatalysts system consists of a corrugated monolithicSCR catalyst solely. The SCR catalyst had a volume of32 liter and a cell density of approximately 170 cpsi. Thevolume corresponded to 2.7 litre of catalyst per enginecylinder litre.

Figure 8. Corrugated ceramic Urea-SCR catalyst.

REDUCING AGENT - A 32.5% urea-water solution witha freezing point of –12°C was used as reducing agentand supplied by Norsk Hydro to the project.

RESULTS; OEM SILENCER

FLOW DISTRIBUTION THROUGH CATALYST - Theexhaust flow distribution inside the catalyst wasmeasured by applying a set of thermocouples inside thecatalyst and by increasing the inlet temperature to thecatalyst by increasing engine load. The heating rate wasused as a measure of the flow passing through thedifferent parts of the catalyst where thermocouples weremounted. Figure 9 shows that there is an evendistribution of the flow in one radial direction. However, aslightly lower flow is seen along the outer wall of thecatalyst.

UREA DISTRIBUTION THROUGH CATALYST - Theurea distribution was measured by traversing a gas-sampling probe over the outlet geometry of the catalyst.Urea was injected, while the measurements were carriedout. The reduced outlet NOx concentration at a specificposition behind the catalyst was used as measure of theamount of urea distributed to the same position.Measurements at 15 different positions gave a picture ofthe urea distribution over the catalyst. The average

concentration from 15 different positions measured was634 ppm NOx with a standard deviation of 52 ppm. Thelowest NOx level measured was 550 and the highest 740ppm. The concentration downstream the silencer, in theexhaust pipe, was measured to 661 ppm. These resultsshow that the urea distribution was sufficiently distributedover the catalyst.

0

100

200

300

400

500

600

700

400 500 600 700 800 900 1000

Time (s)

Massflow (kg_h)

Outer wall (°C)

Between (°C)

Centre (°C)

Figure 9. Temperature response at one side of thecatalyst on a change in massflow and inlet temperature.

Figure 10 shows the flow/temperature response in fourdifferent radial directions. This figure shows that thereare minor differences in the flow rate between the fourpoints within the catalyst.

0

100

200

300

400

500

600

700

0 100 200 300 400 500 600Time (s)

Mass flow (kg/h)T1 (°C)T2 (°C)T3 (°C)T4 (°C)

Figure 10. Temperature response at four different sidesof the catalyst on a change in massflow and inlettemperature.

STATIONARY EXPERIMENTS AND CATALYSTACTIVITY MAPPING - The urea injection system usedwas based on a NOx conversion map, in which torqueand engine speed were input variables. In order to createthis map the possible NOx conversion without NH3 slipas a function of torque and engine speed was tested.The mapping was carried out for four different enginespeeds, 600, 1000, 1400 and 1800 rpm.

The torque was varied within each of the four differentengine speeds, and leading to different exhausttemperatures.

The mapping work was extensive since the urea injectionhad to be fine-tuned to find the point of breakthrough ofNH3 downstream of the catalyst for different each enginespeeds and torque levels. One example of such anexperiment is seen in Figure 10. The points withmaximum NOx conversion and without NH3 slip werechosen and used to create a 3D urea injection map, seeFigure 11, where the x- and y-axis represent enginespeed and torque, and the z-axis represents the amountof urea required. This map was then used as a base forthe urea-injection even though some parts of it wasmodified in different ways based on the experience fromthe transient testing.

0

20

40

60

80

100

500 600 700 800 900 1000NH3 injection (ppm)

NH

3 sl

ip (p

pm)

0%

20%

40%

60%

80%

100%N

Ox

conv

ersi

on (%

)

NH3-out ppm

Nox conversion %

Figure 11. NOx reduction and NH3 slip as a function ofinjected NH3 at 270°C.

600

900

1200

1500

1800

2100

500

14000

1000

2000

3000

4000

5000

Urea injection

(g/h)

Engine speed (rpm)

Engine torque (Nm)

Figure 12. Urea injection map as a function of torqueand engine speed.

TRANSIENT EXPERIMENTS USING A UREAINJECTION MAP - The transient experiments werecarried out during the 30 minutes long EuropeanTransient Test cycle, ETC.

All measurements of the exhaust gas composition weremade in raw exhaust. The conversion efficiencies werecalculated by comparing emissions from differenttransient tests where measurements were made

upstream the catalyst to tests with measurements madedownstream the catalyst.

The mass emission of each component was calculatedfrom measured concentrations and measured totalexhaust flow. Corrections were made to compensate fortime lag in the exhaust gas sample system.

One first example of the conversion efficiency of such asystem is shown in Figure 13, were the NOx conversionwas measured to be 77%. The measured NH3 slip wasless than 10% of the NOx on the outlet based on molarflows.

00,0020,0040,0060,0080,01

0,0120,014

1

147

293

439

585

731

877

1023

1169

1315

1461

1607

1753

Time (s)M

olar

flow

(mol

/s)

NOx in (mol(s)NOx out (mol/s)NH3 out (mol/s)

Figure 13. Inlet and outlet NOx and NH3 molar flows in atransient test cycle. Total NOx conversion was 77%,NOx on the inlet 9.09 mole, NOx on outlet 2.01 mole andNH3 slip 0.13 mole. In total 669 g urea was injectedduring the test cycle.

By increasing the urea overall injection of urea by 12%the conversion efficiency was increased to 82% still witha low NH3 slip se figure 14.

0,0000,0020,0040,0060,0080,0100,0120,014

1 185 369 553 737 921 1105 1289 1473 1657Time (s)

Mol

ar fl

ows

(mol

/s)

Nox in (mol(s)Nox out (mol/s)NH3 out (mol/s)

Figure 14. Inlet and outlet NOx and NH3 molar flows in atransient test cycle. The total NOx conversion was 82%,NOx on the inlet 9.09 mole, NOx on outlet 1.52 mole andNH3 slip 0.16 mole. In total 751 g urea was injectedduring the test cycle.

Figure 15 shows a case when urea is over-injected toprovoke a larger slip of NH3. This led to a NOxconversion of 87% giving an ammonia slip nowmeasured to about half the outlet NOx emission.

0,0000,0020,0040,0060,0080,0100,0120,014

1

150

299

448

597

746

895

1044

1193

1342

1491

1640

1789

Time (s)

Mol

ar fl

ow (m

ol/s

)NOx in (mol/s)NOx out (mol/s)NH3 out (mol/s)

Figure 15. Inlet and outlet NOx and NH3 molar flows in atransient test cycle. The total NOx conversion was 87%,NOx on the inlet 9.08 mole, NOx on the outlet 1.16 moleand NH3 slip 0.77 mole. In total 886 g urea was injectedover the test cycle.

Figure 16 shows the results of two consecutive ETC-tests with no intermediate time delay. The first part of therun shows 85% conversion with a urea injection of 800 g.The second run shows 89% conversion with a ureainjection of 803 g. It is seen in the figure that in thesecond, hot cycle, there is a substantial reduction of theNOx emissions in the beginning of the cycle. This higheractivity is, not only, due to a higher start temperature ofthe catalyst, when the cycle is continued, but also due toa surface coverage of NH3 on the surface at the start ofthe cycle.

00,0020,0040,0060,0080,01

0,0120,014

1 383 765 1147 1529 1911 2293 2675 3057 3439Time (s)

Mol

ar fl

ow (m

ol/s

)

NOx in (mol/s)NOx out (mol/s)

Figure 16. Two test cycles, each 1800 seconds long inseries with NOx conversion of 85 and 89%.

NOX CONVERSION EFFICIENCY V.S. UREA INJEC-TION - Figure 17 shows the results of several transients.The NOx emissions in g/kWh and the average NH3 slipare shown as a function of the amount of urea injectedduring the test cycle.

Figure 17 also illustrates that it is possible to achieveNOx conversion up to 80% on an Euro II engine with areasonable ammonia slip in the range of 10 to 15 ppm.Injection of more urea lowers the NOx emissions on theexpense of a higher NH3 slip, which should be eliminatedby a subsequent oxidation catalyst. This is achieved byusing a single NOx conversion map, described above.Lower NOx emissions without increase in NH3 slip can

be reached by using a more sophisticated urea injectioncontrol strategy [9].

0%

20%

40%60%

80%

100%

0 500 1000 1500

Total urea injection (g/cycle)

NO

x co

nver

sion

(%)

0

50

100

150

200

"Ave

rage

" NH

3 sl

ip

(ppm

)

NOx conv. cold cycleNOx conv. hot cycleNH3 ppm

Figure 17. NOx conversion (%) and NH3 slip (ppm) as afunction of total urea injection in the ETC cycle.

PRESSURE DROP – The pressure drop (dP) over theOEM Urea-SCR silencer was 20% lower than the dPover the original standard silencer of the engine type, seFigure 18. This low dP over the Urea-SCR-silencer wasachieved by careful elaboration and computerized flowmodeling of all flow conditions in the Urea-SCR silencer.

0

200

400

600

800

1000

0 500 1000 1500 2000 2500Massflow (kg/h)

dP (m

mVs

)Original silencerOEM Urea-SCR Silencer

Figure 18. Pressure drop over Urea SCR silencer andstandard original silencer.

SILENCING PERFORMANCE - The differences insilencing performance between the Urea-SCR silencerand the original standard silencer is seen Figure 19. Thestandard silencer is seen as the zero level. The soundmeasurements were done via a microphone inside anexpanded part of the exhaust system.

The OEM Urea-SCR silencer attenuates 0-4 dB betterthan the standard original silencer up to 125 Hz. Theattenuation is 2 dB lower up to 1100 Hz. Between 1100and 2500 HZ the attenuation is 6 dB lower than theoriginal. This might be due to some flow-inducedregenerated noise. When evaluating the differences insilencing performance one should bear in mind that ingeneral it is more difficult to obtain good attenuation atlower frequencies, and that at high frequencies exhaustnoise is usually outweighed by other noise sources of avehicle.

-10,0

-8,0

-6,0

-4,0

-2,0

0,0

2,0

4,0

6,0

8,0

10,0

12,0

25 32 40 50 63 80 100 125 160 200 250 315 400 500 630 800 100012501600200025003150400050006300800010000

Frequency 1/3-octave [Hz]

Average 1000-2000rpm Average+St.de

Figure 19. Differences in sound attenuation (dB)between the Urea-SCR OEM silencer and engine originalsilencer.

RESULTS; UREA-SCR RETROFIT SYSTEM

DEVELOPMENT OF UREA INJECTION STRATEGIESBASED ON REAL TIME CALCULATIONS - The retrofitinjection control system is as described above based onreal time kinetic calculations using readings fromdifferent sensors as input and not based on engine orurea injection maps. This injection control systems alsoincludes a more sophisticated injection control strategy inorder keep a high degree of NOx conversion and at thesame time minimize slips of ammonia. The moreadvanced injection control strategy uses both calculatedNOx conversions together with event-based calculations.

In Figure 20 below, the results from two sets oftransients are shown where the maximum allowed NOxconversion was set to be 80%. In the first case with asimple urea injection strategy, using NOx conversioncalculations solely (referred to as simple) and in thesecond the more sophisticated strategy as describedabove (referred to as advanced). The conversionefficiency and the amount of urea injected were almostidentical for the two different tests, see Table 4.However, the ammonia slip was only one forth using theadvanced strategy compared to the simple one. At thesame time the NOx conversion efficiency was improvedby 3% using the advanced strategy compared to thesimple one, indicating a better use of the injected urea.

Table 4. Results from ETC test with a simple andadvanced urea injection strategy.

Injectionstrategy

NOxg/kWh

NH3g/kWh

Ureag

ηηηη

Simple 2.28 0.018 325 73%

Advanced 2.13 0.006 323 76%

Figure 20 shows the ammonia molar flow as a functionof time for the two cases above. The test using thesimple strategy showed ammonia slip peaks in the rangeof 1.6 mmole/s, corresponding to about 40 ppm, whilethe test with advanced strategy gave no high peaks ofammonia slip. In the figure ammonia emission is givenas mole/s. As a comparison the engine emits about 4.6moles of NOx during the whole ETC test cycle, whichgives an average NOx molar flow of 118 mmole/s.

00,000020,000040,000060,000080,0001

0,000120,000140,000160,00018

1 222 443 664 885 1106 1327 1548 1769Time (s)

NH

3 sl

ip (m

ole/

s)

NH3 slip w simple strategyNH3 slip w advanced strategy

Figure 20. Ammonia slip during a ETC test cycle withsimple and advanced urea injection strategy.

The inlet and the outlet NOx flow from anotherexperiment with the advanced injection strategy aboveare shown in Figure 21. It is seen that the conversionefficiency is low in the beginning of the test cycle. Theexhaust temperature upstream the catalyst wasconditioned to be 100°C before starting the ETC tests.The urea injection was not started until the exhausttemperature exceeded 200°C, leading the fact that nourea was injected under the first 150 seconds. Theoverall NOx conversion was 84%. The NOx emissiondownstream the Urea-SCR silencer was 1.3 g/kWh andthe NH3 emission 0,03 g/kWh.

0,0000,0010,0020,0030,0040,0050,0060,0070,008

1 601 1201 1801Time (s)

NO

x an

d N

H3

flow

(mol

e/s)

NOx in mole/s

NOx out mole/s

NH3 out mole/s

Figure 21. Inlet, outlet NOx and ammonia slip of a ETCtest cycle with advanced urea injection strategy

NOX CONVERSION EFFICIENCY V.S. UREA INJEC-TION - Several transient experiments with the ETC cyclewere conducted in order to develop the urea injectionstrategies. NOx conversions and ammonia slip as afunction of total mass of urea injected from theseexperiments are presented in Figure 22. The figure

shows that it was possible to obtain NOx conversions ofup to 85% with ammonia slips that can be regarded aszero. It should be pointed out that these experimentswere conducted with a “cold start”, meaning an inletexhaust start temperature of 100°C. It can also be seenthat for some experiments with the same mass ureainjected, different ammonia slips were measured. Thisvariation is related to different injection strategies, givingvarying ammonia slip. The amount of diesel fuelconsumed during a test cycle was 5.7 kg giving a relativeurea consumption of 5 to about 8% of the fuelconsumption in terms of weight.

020406080

100

0 100 200 300 400 500Total urea dosage (g/cycle)

NO

x co

nver

sion

(%)

00,10,20,30,40,5 NH

3 slip (mol)

NOx conversion (%)NH3 slip (mole)

Figure 22. NOx conversion and ammonia slip as afunction of the total amount of urea injected during theexperiments with ETC-transient test cycles.

TESTS AT MTC – A verifying function test of thecomplete Urea-SCR system was conducted at MTC inSweden. Table 5 includes the values of the emissionstandard in force from 1999 and the emission standardwith effect as from 2002 for off road engines. It alsoincludes a measurement of the emissions at Volvo onthe engine used in this work. Finally, it includes ameasurement of the engine base emissions at MTCusing the ISO 8178 C1 cycle. The lower NOx andparticulate emissions in the MTC test can be explainedby differences in the fuel used; low sulfur and lowaromatic give less NOx and particulate.

Table 5. EU emission standard of off road engines 1999(Stage I) and 2002 (Stage II) and measured engine outbase emissions at Volvo and at MTC.

NOxg/kWh

HCg/kWh

COg/kWh

PMg/kWh

1999 9.2 1.3 5.0 0.54

2002 6.0 1.0 3.5 0.20

Volvo1 8.0 0.4 1.3 0.28

MTC2 7.1 0.4 1.9 0.181 Certification fuel standard.2 Swedish low sulfur Environmental Class 1 fuel.

ISO 8187 C1 test. - Three ISO 8178 tests with and threewithout the Urea-SCR system were carried out and theaverage results from these tests are presented in Table6. From these tests it was concluded that the Urea-SCRsystem is able to reduce 77% of the NOx emissions. Thesystem also reduces HC emissions by 87% and PMemissions by 27%, while CO was increased by 30%.

These tests proved that the Urea-SCR system haspotential to reduce regulated emissions and enabling aEU Stage I engine combined with a Urea-SCR system tocomply with the future EU Stage II emission standard foroff road engines, with effect as from 2002.

Table 6. Results from triple ISO 8187- C1 test.

NOxg/kWh

HCg/kWh

COg/kWh

PMg/kWh

Fuelg/kWh

- SCR 7.11 0.38 1.86 0.18 214

+ SCR 1.63 0.05 2.6 0.13 214

ηηηη -77% -87% 32% -27% 0%

ETC-test - The urea injection system was tuned to givevarious levels of NOx reduction in the different ETCtests. The first results given in Table 7 are from atransient test where the injection system was set to givea high degree of NOx reduction with a minimum slip ofNH3. In these tests the Urea-SCR system was active andin use also in a period of preconditioning of the enginebefore the actual test. This start procedure made thetemperature in the SCR reactor more favorable for NOx-reduction and at the same time the catalyst also had apre-adsorbed urea/ammonia layer making the catalystactive from the very beginning of the test cycle. Under

these conditions 88% of NOx and 83% of HC werereduced, while CO was slightly increased by 7%. Therewas no significant increase in fuel consumption over thetest-cycle w. or w/o. the Urea-SCR system.

Table 7. Results from a single ETC test with ECU tunedfor high NOx conversion and low NH3 slip.

NOxg/kWh

HCg/kWh

COg/kWh

PMg/kWh

Fuelg/kWh

- SCR 8.12* 0.42* 1.97* 0.27* 226*

+ SCR 0.92 0.07 2.10 NA 225

ηηηη -88% -83% 6.8% NA 0%

* Triple experiments on without catalyst

The potential NOx reduction efficiency of the Urea-SCRsystem was tested in a ETC cycle by overdosing urearelatively to NOx, se Table 8. This measure led to analmost complete reduction of NOx. The reading in massNOx was 0.02 g/kWh when the Urea-SCR system wasused, and this is far beyond the NOx limit of 0.2 to 0.5g/kWh suggested by EPA for year 2007 in USA. SI-MSmeasurement of NH3 downstream the catalyst gave anammonia slip of about 100 ppm in average throughoutthe test cycle. It will not be possible to control this kind ofammonia slip originating from such high level of NOxconversion by using an injection strategy solely, anammonia oxidation catalyst downstream the Urea-SCRcatalyst will also be required.

The HC conversion was 76%, which is slightly lower thanthat obtained in the test shown in Table 7. Theexplanation is that HC and NH3 compete over the activesites on the catalyst surface. A higher partial pressure ofammonia (urea) gives a higher degree of ammoniasurface coverage leading to fewer sites for hydrocarbonsto adsorb and react.

One remarkable result was the high particulate reductionof 52%. This high PM reduction was controlled by fourtests with the Urea-SCR system and three tests without.The average reduction in these tests turned out to be53%, in good agreement with the results shown in Table6.

Table 8. Results from a single ETC test with ECU tunedfor maximum NOx conversion.

NOxg/kWh

HCg/kWh

COg/kWh

PMg/kWh

Fuelg/kWh

- SCR 8.12* 0.42* 1.97* 0.27* 226*

+ SCR 0.02 0.10 2.05 0.13 228

ηηηη -99.7% -76% 6,8% -52% 1%

* Triple experiments without Urea-SCR system.

CONCLUSION

The OEM Urea-SCR silencer project has achieved 5 outof 6 design targets. The space requirement target wasnot reached, as the silencer is 250 mm longer than theoriginal.

The sound level is reduced at lower frequencies, whereacoustic performance is of particular importance. Thepressure drop is lower than with a standard silencer,which is remarkable in that both urea injection/mixtureand flow across the catalyst are processes, which per secause pressure drop. The lower pressure drop acrossthe unit reduces fuel consumption of the engine. 85%NOx reduction was demonstrated with 32 liter of catalystshowing that the Urea-SCR technology has a NOxreduction potential for meeting the very stringentEuropean emission targets for 2008.

The importance of injection strategy and the relationbetween urea/NOx ratio, ammonia slip and obtainedNOx-removal was elucidated by transient tests with aUrea-SCR retrofit system and a 10 liter 190 kW VolvoHD excavator engine on test rig. NOx-removals of 77 to88% were demonstrated in ETC and the European ISO8178C1 off road test cycles with essentially no ammoniaslip. 99.7% deNOx was obtained with a high NH3-slip ofabout 100 ppm, illustrating the potential of the urea SCRtechnology. The silencer contained 29 liter catalyst.

The described Urea-SCR technology has been proven tobe efficient in reducing not only NOx but also HC and PMemissions under both stationary and transient conditions.

ACKNOWLEDGMENTS

The authors thank the Swedish authority “Miljöteknik-delgationen” for supporting the Urea-SCR Retrofit projectfinancially. The authors also thank Scania for supportingthis project with engines and auxiliary equipment. NorskHydro is aknovleged for their supply of Urea to theprojects.

REFERENCES

1. Havenith et al., “Entveichlung eines Verbrauchs-Optimmierten Euro 5 Nfz-Dieselmotors mit Harnstof-SCR-Abgasnach-behandlung.”, VDA TechnischerKongress, 28. -29. Sept. 2000, Frankfurt.

2. Held et al, “Catalytic NOx reduction in Net oxidizingexhaust gas”, SAE-paper 900496.

3. Walker et al, “Development of an Ammonia/SCRNOx Reduction System for a Heavy Duty NaturalGas Engine”, SAE paper 921673.

4. Andersson et al, “Reducing NOx in Diesel Exhaustsby SCR Technique: Experiments and Simulations”,AIChE Journal, Vol. 40, No. 11, p 1911-1919, 1994.

5. Havenith et al, “Transient Performance of a UreadeNOx Catalyst for Low Emissions Heavy-DutyDiesel engines”, SAE paper 970185.

6. Fritz et al, “On-Road Demonstrations of NOxEmissions Control for Diesel Trucks with SINOXUrea SCR System”, SAE paper 1999-01-0111

7. Nobutaka Kihara et al., “Real Time OnboardMeasurements of Mass Emissions of NOx, FuelConsumption, Road Load and Engine Output forDiesel Vehicles”, SAE paper 2000-01-1141

8. Christoffersenet al, “Progress in SCR DENOX andSilencing Technology for Diesel Engine ExhaustSystems”, Proceedings of the 22nd CIMACInternational Congress on Combustion Engines, vol.5, p. 1131-1146.

9. Gabrielsson et al, “Urea-SCR NOx RemovalEfficiency For a Diesel Engine in the SuggestedEuropean ETC Test Cycle”, 32nd ISATA Conference,14th – 18th June 1999 –Vienna- Austria

CONTACT

HALDOR TOPSØE A/SPär GabrielssonNymollevej 55DK-2800 LyngbyDenmarkPhone +45-45272184Fax +45-45272999e-mail: pg@topsoe.dkHomepage : www.haldortopsoe.com

DEFINITIONS, ACRONYMS, ABBREVIATIONS

ECU - Electronic control unitETC - European transient test cycleMTC - Motor Test CenterSI-MS - Soft Ionization Mass SpectrometerSCR - Selective Catalytic ReductionPM - Particulate matterNOx - Nitrogen oxideHC - HydrocarbonsCO – Carbon monoxideNH3- ammonia