Future Perspectives on Diesel Emission Control. Tim Johnson April 6, 2007

Transcription

1 Future Perspectives on Diesel Emission Control Tim Johnson April 6, 2007

2 Summary Regulations European LDD regulations have implications to the US approach Europe is formally proposing Euro VI (HD) this year Wide range of scenarios considered no industry consensus CO2 regulation appears to be the most dominating long term issue Good or bad for LDD? Engine strategies are making impressive progress HD research engines show very low PM levels under steady state full-load conditions Meet US NTE levels without filters Advanced combustion mixed mode engines are emerging for LD and HD NOx solutions are available for ultra-low emissions SCR is addressing cold temperature and secondary emission issues LNTs are performing well today at about 60-70% efficiency; economically attractive for smaller platforms and mixed mode engines DPF systems show continuous improvement Very sophisticated regeneration control strategies. Mixed mode engines could change DPFs quite significantly Some outlying issues yet to be addressed NO 2 and ultrafine PM 2

3 Regulatory Trends

4 To sell European cars into the US market, a minimum of 50-60%% NOx control is needed today to hit a 42-state market (70% US). By 2013, 65% NOx control could get the 50-state market. NOx, mg/km Euro 5, % NOx control 85% NOx control Bin 8 Bin 5 At the Euro 5 limit value of 180 mg/km NOx, 55% NOx control is needed to hit Bin 8 (70% of US), and 85% NOx control is needed to hit Bin 5 (all US). NOx, mg/km Euro 6, Addn l 65% NOx control Bin 8 Bin 5 At the Euro 6 level of 80 mg/km NOx, an additional 65% NOx control is needed to hit Bin 5 (all US). Cutting edge engine-out for 2008/9 is ~110 mg/km, needing 75% NOx control for T2B5. 4

5 Regulatory and engine technology framework 0.14 PM, g/kw-hr; ESC test US2010+ US2010 DPF+NOx Japan 2009 DPF+NOx US2007 DPF US2007 Japan 2005 DPF or SCR US2004 (2002) DOC Euro V 2008 SCR Euro IV 2005 SCR 0 Euro VI 2012? DPF+NOx NOx, g/kw-hr; ESC test Euro III 2000 nothing 5

6 The EU is considering six different Euro VI scenarios. NOx or PM, g/kw-hr NOx PM X 10 Engine-out Threshold Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 US2010 NTE and number-based PM standards also considered. Timing of discussed EC will make a proposal about December 2007 US incremental Research engines 6

7 WHTC has lower temperatures Bosch, Euro V&VI Conf

8 Non-Road has significant challenges US2007-like regs in 2011 Worldwide Cooled EGR limitations DPFs leading concept, but some are looking a at Euro IV/V SCR approach US2010-like regs in 2014 EGR+SCR+DPF seems to be leading Much resistance to urea-scr Best solution seems to be mixed-mode with HC-deNOx (LNT or LNC) and DPF Timing fits Huge watchout NR moves into engineering in Conflicts with US2010, Euro VI, and US LDD Industry resource crunch, especially in supply chain and for dynos Medium sized NR companies appear behind, already 8

9 AB 1493 forces the fleet average 21% below the 2006 EU fleet actual. 37% cut in 7 years required US truck fleet US truck CAFE (unreformed) Slide courtesy of Martec CO 2 grams/km US car CAFE EU fleet -14% in 11 years US car fleet AB 1493 LDT2-7g/km AB 1493 PC/LDT1-37% in 7 years 125 ACEA Voluntary Agreement EU Target US and EU test cycles are not equivalent. ARB standard includes GHG losses from AC system. AB 1493 LDT2 definition includes 8,500-10,000 lb. gvw MDPVs. Fuel economy conversions at lb CO 2 /gallon gasoline and lbs/gallon diesel. US fleet excludes flex fuel credits. 9

10 Engine Developments

11 HDD RESEARCH ENGINES are hitting 0.5/0.015 NOx/PM on the ESC PM, g/kw-hr X X Univ WI virtual engine of 2001 ESC Threshold On average, 60-70% NOx control and 30% PM control are needed to hit steady state requirements Bosch, SAE HDD Symp, 9/ Advanced Engines MAN CTI conf.; 1/ NOx, g/kw-hr

12 HD research engines are showing impressive performance under steady state conditions. 12

13 HD Engine/Regulatory Comments US2010 regulations can be achieved in a number of ways Most likely approach is to use incremental engine technologies to US2007, and add appropriate NOx control Incremental improvements will drop NOx from nominally 1.0 g/bhp-hr NOx to nominally 0.7 g/bhp-hr NOx SCR will be used in fuel-sensitive sectors for 80-85% NOx efficiency; allows g/bhp-hr engine-out 2010 could be fuel-consumption neutral vs. US2007 More advanced engine hardware might drive NOx down to 0.5 g/bhp-hr NOx High EGR, FIE, boost, control Requires 60% NOx control (LNT) SCR will allow g/bhp-hr engine-out for fuel economy gains First cost vs. operating cost and marketing trade-offs L/MDD up to 10,000# will chassis cert. to 0.2 g/mi NOx, and 10-14,000# will chassis cert to 0.4 g/mi Minimal NOx aftertreatment, if any 13

14 Research results show large coverage of US LD test cycle in advanced combustion mode Cummins, Univ WI ERC symposium, 6/05 Clean advance low-nox combustion Traditional diesel combustion Only a small fraction of test cycle load points fall within traditional combustion mode As much of the certification cycle is met using advanced combustion, NOx aftertreatment requirements are reduced or eliminated. However, off-cycle emissions will be high and require denox 14

15 SwRI internal research 11/05 Combined LTC/PCCI/traditional diesel combustion strategy provides low noise, maintained fuel efficiency, and manageable emissions. Running rich LTC with an LNT gives very low fuel penalty. Model-based control of airflow and injection timing/quantity results in precise combustion Additional engine hardware is modest 15

16 Advanced combustion technology will be in both LD and HD platforms in LD Japanese and some US and Euro Autos seem to be developing US-specific engines Saves on total cost and takes out SCR However, might not be best Euro 6 approach Mixed mode for luxury or larger class No de-nox for Euro 6 Others: traditional diesel for Euro 6; add denox for US Bin 5 Diesel engine technical leadership could shift to the Japanese and US. HD Addresses low-load NOx and low load DPF management SCR and LNTs don t work well ($) in urban settings DPF management easier with lower PM load at low load 16

17 SCR

18 US2010 concepts involve SCR, but there are complex tradeoffs with engine design, SCR efficiency and size, and relative urea costs. Leading heavy-hdd US2010 system layout is DPF+SCR, with incremental changes in 2007 engine technology. For any given engine architecture (BSFC vs. NOx), there is a minimum operating cost relationship with SCR efficiency based on urea consumption. Daimler, CTI conference,

19 State-of-the-art urea delivery system described Features: no return line for simplicity and cost savings self-draining lines to prevent freezing and urea deposits mixer to enhance urea decomposition Bosch, MinNox conf. Berlin

20 Urea mixing is especially critical TU Munich, MinNox Conf. 1/07 Poor urea mixing results in injector clogging, catalyst peeling, and poor distribution/hydrolysis of urea. Vortex mixer is a thin shield in front of injector that creates strong localized mixing. Solid urea pellet grinder reduces size to 40 to 150 μm, depending on starting material. Urea is decomposed most effectively with vortex mixer at 250C. 20

21 Zeolite SCR catalysts perform well at LT for short periods of time. T<200C: NH 4 NO 3 formation LT: Rate controlling step is NO NO 2 kinetics PSI, MinNOx conf

22 Fast heat-up strategy is very effective, but costs fuel. 80% cycle denox costs 9% FP. Baseline at 65% efficiency Fast heat-up strategy involves postinjection, only when needed. FEV, MinNOx conf

23 Ammonia sensor in development that will enable closed loop SCR control. NH3 sensor uses lean exhaust reference and a proprietary electrolyte material. 3-5 second response time closed loop NH3 control eliminates slip catalyst, and offers maximum SCR efficiency, as NOx sensors have ammonia sensitivity. ±5-10 ppm accuracy on ESC test after 700 hours. Delphi, CTI conference

24 Lean NOx Traps

25 New LNT+SCR systems are emerging. Longer desulfation intervals. Potentially smaller size Eaton, Euro V&VI Conf Fuel reformer generates hydrogen during the rich spikes to aid in LNT regeneration. NH 3 is formed by rich LNT operation, and this is used in the SCR. Transmission is used to deliver lower speed for richer, hotter gas. 25

26 A new double layer LNT/SCR approach is described. CeO 2 LNT material has good LT activity and low sulfur affinity. The bottom layer acts as an LNT in lean operation, and converts NOx to NH 3 via the water gas shift reaction during rich operation via the water gas shift reaction. The NH 3 is adsorbed by the zeolite top layer for use as an SCR reductant. In model gas testing the CeO 2 LNT material performed well at low temperatures. System is placed in under-body position to reduce temp. Honda, Aachen Colloquium, 10/06 The CeO 2 LNT material has a low sulfur affinity, and thus can be desulfated at 500C. Rich combustion is used to generate CO and H 2 for NH 3 generation

27 Higher PGM loadings on LNTs are not necessarily better. Mixed mode engines have low NOx at T<350C. Synergies with LNTs. After aging at 650C, the high PGM loading helps LT NOx efficiency, low PGM loading are better at higher temperatures. At LT, PGM helps NO2 formation. At HT, 90 gpcf PGM loadings cause nitrate decomposition in lean conditions, dropping capacity. Trend holds to 800C aging. After aging at 650C, the samples with low PGM loading help NOx efficiency at 300C, but advantage diminishes as purge time increases. High PGM loadings have higher oxygen storage capacity, resulting in more purge being needed. Ford, SAE

28 A rapid screen process for LNC compositions is beginning to yield results. Synergies with mixed-mode engines. Possible utopia for diesel: mixed mode engines + low cost LNC. GM, Gordon Conference 8/05 The rapid screening process involves three parties, fast synthesis and testing, with refinement before advanced testing compositions have been evaluated Extrudate results. More than 4500 LNC compositions have been evaluated. Some give superior performance to the standard. Low temperature results need improving, but high temperature results are favorable and significantly better than standard. A leading candidate shows resistance to sulfur, good selectivity, and decent performance at modest C/N ratios. 28

29 First-cost economics favor SCR for engines larger than about 2 liters. Cost Incremental Costs, LNT vs. SCR LNT cheaper ~2 liter SCR cheaper 50-70% PGM savings w/ mixed mode; LNT ~5 liter LNT PGM incremental cost of $80-90 per liter displacement SCR increm. variable cost of $10-15 per liter displ. Fixed urea system cost ~ $180 Displacement Aggressive mixed mode and urea system (potential future systems) have breakeven at 5 liters displacement. Assumptions: Substrate, can, and wash coating costs are the same for LNT and SCR SVR=0.8 assumed for LNT and SCR; NH 3 slip catalyst SVR=0.25, PGM 30 g/cu.ft.; PGM LNT = 80 g/ft 3 ; PGM at $40/g 29

30 PM Control

31 An algorithm is developed to safely control DPF regeneration. Increased fuel injection pressure is the only evaluated parameter that affects soot oxidation temperature. Soot morphology change observed for increased injection pressure. Method used to obtain oxidation rate constant (soot amount/combustion duration). Chart obtained from visual observation. Coefficients for rate constant Hino, SAE Combustion rate constant and inlet temperature are used to detect threshold between normal and abnormal regeneration. 31

32 A lambda sensor after DPF is used to control the regeneration. DPF bed temperature is controlled by oxygen level. Oxygen control is very good in transient conditions. Nissan SAE Adaptive learning tightens A/F control and allows better soot estimation. 32

33 New ZrSi-based DOC washcoats improve light-off and durability properties New ZrSi-based WC drop T50 for propene by 25C vs. alumina After severe aging at 850C, new ZrSi-based WC drop T50 for CO by 30C vs. alumina at 1 wt% Pt, but advantage nearly disappears at 0.5 wt% Pt. Rhodia SAE Properties of Pt washcoats after hydrothermal aging at 750C for 16 hr and sulfation at 300C for 12 hr. 33

34 Passive regeneration might be achievable in LT and low NOx using CSF and HC/CO using advanced combustion EPA DEER 8-04 Low engine-out NOx comes from high EGR, high boost, high injection pressure. However, PM and HCs are rather high and exhaust temperatures can be low. Despite low NOx, LT (<250C) and high PM loading, preliminary reports are that the filter stays clean. Speculation that HCs react on catalyst to burn fresh, active soot. 34

35 Full load HCCI might not be a feasible end point Mixed mode engines hit the sweet spot of low cost DPF and denox technologies Smaller, mature, mostly-passive DPFs Cheap LNCs (HC-SCR) or low-pgm LNTs The incremental engine hardware, fuel infrastructure, and development expense might not justify replacing the DPF+deNOx with a DOC. 35

36 Some secondary issues to resolve

37 NO 2 toxicity is quantified for the first time. -70% 1.2-2X NO 2 increases due to higher LDD penetration with catalysts NO 2 mortality rate is about 0.3% per 10 μg/m3 increase in concentration. TNO, Envir. and Transp. Reims 6/06 APHEA, Euro Respiratory Journal, 27,

38 NO 2 based filter systems are emerging that have very low NO 2 emissions Downstream catalyst selectively converts NO2 to NO using diesel fuel, with little reaction with oxygen. Johnson Matthey, Dresden Conf 5/06 Comment: Move to Pd catalysts, lighter catalyst loadings, FBC, and denox aftertreatment for Euro 6 are auto industry approaches to decrease NO 2 38

39 Ultrafine emissions from filtered truck are described. +100X sulfates 100 km/hr ULSD+ LS oil DPF: -99% soot Filters take out 99+% of soot particles, but can increase sulfuric acid aerosol particles <30 nm, even with LS fuel and lube oil. 100 km/hr cruise for HDD truck. New catalyst component (sulfur trap) added to filter drops sulfuric acid aerosols to below ambient concentrations. Univ. MN SAE

40 Summary Regulations European LDD regulations have implications to the US approach Europe is formally proposing Euro VI (HD) this year Wide range of scenarios considered no industry consensus CO2 regulation appears to be the most dominating long term issue Good or bad for LDD? Engine strategies are making impressive progress HD research engines show very low PM levels under steady state full-load conditions Meet US NTE levels without filters Advanced combustion mixed mode engines are emerging for LD and HD NOx solutions are available for ultra-low emissions SCR is addressing cold temperature and secondary emission issues LNTs are performing well today at about 60-70% efficiency; economically attractive for smaller platforms and mixed mode engines DPF systems show continuous improvement Very sophisticated regeneration control strategies. Mixed mode engines could change DPFs quite significantly Some outlying issues yet to be addressed NO 2 and ultrafine PM 40