Real-time, on-road measurements of diesel exhaust aftertreatment device PM removal efficiency

Transcription

1 Acknowledgements All data presented here were collected by the author. Some of the examples shown were part of studies funded by the United States Federal Highway Administration, Pittsburgh Clean Cities, New York City Department of Transportation, Port Authority of New York and New Jersey, IdleAire Corporation, California Air Resources Board, and others. Some work has been done as portion of author s graduate studies. Part of the work, including compilation and preparation of this presentation, has been done at the Josef Božek II. Internal Combustion Engine and Automobile Research Center, funded by the Czech Ministry of Education, project no. M

2 th ETH Conference on Combustion Generated Nanoparticles Zurich, 3th - 5th August 2007 Real-time, on-road measurements of diesel exhaust aftertreatment device PM removal efficiency Michal Vojtíšek-Lom Technical University of Liberec Department of Vehicles and Engines, Faculty of Mechanical Engineering Hálkova 6, 46 7 Liberec, Czech Republic, michal.vojtisek@tul.cz, tel Thomas R. Lanni Bureau of Mobile Sources, Division of Air Quality New York State Department of Environmental Conservation Albany, New York, USA (New York City measurements portion)

3 Background Modern engines increasingly rely on the combination of electronic engine controls and exhaust aftertreatment devices to achieve low emissions. Various deviations from ideal operation typically result in substantially higher emissions. Small number of vehicles and small portion of total operating time have a disproportionately high contribution to the total emissions.

4 Example: On-road tests of CNG buses (Pittsburgh, USA, )

5 Example: On-road tests of CNG buses HC emissions deterioration rates HC HC [g/mile] [g/mile] HC HC [g/mile] [g/mile] test mileage test mileage test mileage test mileage

6 Example: On-road tests of CNG buses Differences among identical vehicles Vehicle Vehicle Vehicle 20 CO (g/km) NOx (g/km)

7 Why on-road measurements? When evaluating benefits of a fleet-wide deployment of a new technology, high emissions vehicles and high emissions episodes cannot be ignored as outliers, but must be characterized and accounted for. On-road measurements using portable on-board systems allow for testing of a large set of vehicles under a wide range of operating conditions. Measurements can be done during normal everyday operation of the tested equipment.

8 Evaluation of aftertreatment devices Use one monitoring system. Run separate identical tests with and without aftertreatment, compare. + Relatively easy test Tests have to be repeatable Real-time or proportional sampling instrumentation Use two monitoring systems, one upstream, one downstream of the aftertreatment + Tests do not have to be repeatable More difficult realization of the test Instrumentation has to measure in real-time ( Hz)

9 Repeatability of on-road measurements Instrumentation: - Five-gas (HC, CO, NO, CO 2, O 2 ) garage analyzer - Light scattering device for PM concentration measurements - Sampling: No dilution, no heating of the line - Exhaust flow: Computation based on engine rpm, intake air temperature and pressure, engine parameters and exhaust composition - Second-by-second data Vehicle: 999 International 24-passenger bus, DT-444-E 45 kw, 7.3-liter V-8 electronically controlled direct-injection turbodiesel

10 On-road tests on a test track can be repeatable Bus operating on different blends of diesel fuel and di-methyl ether (DME) Emissions measured using Manhattan Bus cycle driven on a,6 km test track speed [mph] Manhattan Bus Driving Cycle - target road speed Manhattan_Bus Diesel - no DME 4% DME / diesel blend 25% DME / diesel blend PM emissions [ug/s] Test-to-test variance: < 0% for computed total PM mass

11 Measurements in ordinary traffic are poorly repeatable Example: VW Golf / Jetta cars powered by diesel fuel and vegetable oil. Multiple runs along a 20,9 km suburban-highway-city route. Unlike in laboratory tests, comparison is done on a distance basis (x-axis: distance from start of the test) road speed [km/h] VO- D- D-2 VO-2 D-3 D-4 VO-3 VO-4 VO-5 D-5 D elapsed km

12 Example: VW Golf car powered by diesel fuel and vegetable oil. Multiple runs along a 20,9 km suburban-highway-city route. NOx emissions [mg/s] elapsed km VO- D- D-2 VO-2 D-3 D-4 VO-3 VO-5 D-5 PM emissions [mg/s] elapsed km VO- D- D-2 VO-2 D-3 D-4 VO-3 VO-5 D-5

13 Simultaneous upstream-downstream sampling More weight, higher power consumption, decreased system reliability Impossible to sample from a laminar flow upstream of the aftertreatment due to exhaust system geometry of most engines But: Most diesel particles < um; if sampling PM or PM 0.5, sampling error is relatively small compared to other factors Pilot studies done using a simple light-scattering device

14 Simple on-board systems can be capable of repeatable measurements in the lab even at ~0.0 g/kwh levels (but no claims about accuracy) 2000 International 3400 truck, DT-466-E turbodiesel, CRT trap CBD cycle driven on a chassis dynamometer PM concentrations measured with a light scattering device, exhaust flow computed 0000 PM [micrograms per second] CBD_Cold CBD_Hot_ CBD_Hot_2 CBD_Hot_3

15 On-board PM concentrations measurement using a light scattering instrument Samples raw, undiluted exhaust from the tailpipe using 6 mm sample line Real-time measurements of PM scattering efficiency Robust, easy to use, low power consumption Data requires considerable interpretation 4. Scattering detector SAMPLE FLOW LASER BEAM PATH DETECTOR 3. Sample reheat 6. Pump 5. Flow control. Sample inlet 2. Impaction of condensate and large particles

16 On-line diesel oxidation catalyst evaluation Simultaneous Upstream / Downstream Sampling (2 monitoring Systems) Exhaust sampling port before catalyst, ¼ probe, unheated line, no dilution Exhaust sampling at tailpipe ¼ probe, unheated line, no dilution Construction equipment, World Trade Center, New York

17 On-line diesel oxidation catalyst evaluation Three sampling locations: Identical sample ports upstream and downstream of DOC, ordinary probe at the tailpipe Two monitoring systems (light scattering device for PM concentrations) PM [mg/m3] 0 Sampled PM[mg/m3] from - Unit port AD 9upstream PM[mg/m3] of DOC - Unit F Sampled from port upstream of DOC Sampled from port downstream of DOC Sampled from port upstream of DOC Sampled from Port downstream of DOC Sampled from port downstream of DOC Smpl line swap Sample lines changed Sampled from port Sampled upstream from of tailpipe probe DOC Sampled from tailpipe probe Smpl line swap 5:00 5:5 5:30 5:45 6:00 (MQ 00 kva electric generator, 6.8-liter, 35-hp John Deere engine) Sampled from port upstream of DOC Sampled from tailpipe probe

18 Engine-out vs. tailpipe-out: New York City ferries (Staten Island Ferry, Manhattan <-> Staten Island) True in-use testing 2 x Caterpillar 356 V-6 55 kw drive engines Baseline for biofuels and SCR evaluation Sampling at turbocharger outlet AND at stack end Engine A out Engine B out Stack out (alternating) F2070C 000 F207B V207B V2070D Stack-PM Stack-Opacity Stack opacity Exhaust PM [mg/m3] Stack opacity [%] :00 22:30 23:00 23:30 0:00 0:30 :00 :30 2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00

19 On-line diesel oxidation catalyst evaluation In-use emissions testing on construction equipment during regular operation at the World Trade Center no. 7 site, New York, NY Operation not repeatable and difficult to simulate Simultaneous upstream and downstream sampling PM [g/h] Downstream Upstream Eff[%] 50% 40% 30% 20% 0% DOC efficiency [%] 0 6:58 6:59 7:00 7:0 7:02 7:03 7:04 7:05 7:06 7:07 7:08 7:09 7:0 0%

20 On-road emissions from recycled frying oil study VW Golf / Jetta,9 TDI cars with GreaseCar vegetable oil conversion Fuels: Highway diesel fuel and a mix of recycled frying oil from different sources Vegetable oil heated to ~60 C fuel temperature at injection pump inlet 20,9 km suburban / highway / urban test route Emissions measurements upstream and downstream of DOC Data still being analyzed Suburban arterials Freeway Urban road speed [km/h] CENTER CITY VO- D- D-2 VO-2 D-3 D-4 VO-3 VO-4 VO-5 D-5 D elapsed km

21 Passenger car DOC efficiency diesel fuel VW Jetta,9 TDI, 20,9 km test route in ordinary traffic PM emissions [mg/s] Simultaneous sampling Diesel-Upstream Diesel-Downstream elapsed km PM emissions [mg/s] Sampling on two separate runs Diesel-Upstream Diesel-Downstream elapsed km

22 Passenger car DOC efficiency vegetable oil VW Jetta, 20,9 km test route in ordinary traffic 0 Simultaneous sampling VO-Downstream VO-Upstream PM emissions [mg/s] elapsed km 0 Sampling on two separate runs VO-Upstream VO-Downstream PM emissions [mg/s] elapsed km

23 Passenger car DOC efficiency diesel vs. vegetable oil VW Jetta, 20,9 km test route in ordinary traffic PM emissions [mg/s] 0 0. Diesel fuel Diesel-Upstream Diesel-Downstream elapsed km 0 Vegetable oil VO-Downstream VO-Upstream PM emissions [mg/s] elapsed km

24 Passenger car DOC efficiency difference in catalysts 20,9 km test route in ordinary traffic 0 Poor catalyst performance Upstream- Downstream- PM emissions [mg/s] elapsed km 0 Normal operation Diesel-Upstream Diesel-Downstream PM emissions [mg/s] elapsed km

25 Passenger car DOC efficiency diesel fuel vs. vegetable oil 20,9 km test route in ordinary traffic Idle (also prolonged idling) Low exhaust and catalyst temperatures PM dominated by organics Transient high loads Overfueling conditions Relatively infrequent PM dominated by soot 0 PM emissions vs. engine torque - Diesel fuel 0 PM emissions vs. engine torque - Vegetable oil PM concentration [relative] 0. Diesel.Downstream Mean.Downstream Mean.Upstream PM concentration [relative] 0. Veg.Oil-Downstream Mean.Upstream Mean.Downstream CO2 in raw exhaust [%] (as a surrogate of engine torque) CO2 in raw exhaust [%] (as a surrogate of engine torque) During bulk of the medium load operation, PM concentrations are lower for vegetable oil; but mean PM concentrations are higher, suggesting that a large contribution to the total comes from transients

26 Passenger car DOC efficiency diesel fuel vs. vegetable oil Mean vs. median values median values suppress transients 0 PM emissions vs. engine torque - Diesel fuel 0 PM emissions vs. engine torque - Vegetable oil PM concentration [relative] 0. Diesel.Downstream Mean.Downstream Mean.Upstream PM concentration [relative] 0. Veg.Oil-Downstream Mean.Upstream Mean.Downstream CO2 in raw exhaust [%] (as a surrogate of engine torque) CO2 in raw exhaust [%] (as a surrogate of engine torque) 0 PM emissions vs. engine torque - Diesel fuel 0 PM emissions vs. engine torque - Vegetable oil PM concentration [relative] Diesel.Downstream Median.Upstream Median.Downstream CO2 in raw exhaust [%] (as a surrogate of engine torque) PM concentration [relative] Veg.Oil-Downstream Mean.Upstream Mean.Downstream CO2 in raw exhaust [%] (as a surrogate of engine torque)

27 Passenger car DOC efficiency diesel fuel vs. vegetable oil Diesel oxidation catalyst efficiency - Diesel fuel 20% Diesel oxidation catalyst efficiency - Veg. oil 20% PM conc. drop across DOC 0% 0% -0% -20% -30% -40% -50% PM conc. drop across DOC 0% 0% -0% -20% -30% -40% -50% -60% exhaust CO2 [%] (as a surrogate of engine torque) -60% exhaust CO2 [%] (as a surrogate of engine torque) Diesel oxidation catalyst efficiency - Diesel fuel 20% 20% Diesel oxidation catalyst efficiency - Veg. oil PM conc. drop across DOC 0% 0% -0% -20% -30% -40% -50% PM conc. drop across DOC 0% 0% -0% -20% -30% -40% -50% -60% 0. 0 PM concentration [relative] -60% 0 0 PM concentration [relative]

28 PM dynamics within the exhaust system Secondary growth of PM Deposition, storage and re-entrainment of PM Most pronounced during and following idle and low loads with long exhaust systems with high aerosol fraction of PM Must be differentiated from the aftertreatment device effects

29 Effect of prolonged idling on PM emissions 999 Freightliner truck, CAT 3406 engine, ~50,000 miles 8-hour extended idling test Idle Aire Technologies, Knoxville, TN, December 7, 200

30 Effect of prolonged idling on emissions during subsequent driving Class 8 truck idled for 8 hours at high idle, then driven for ~32 miles on an interstate highway PM emissions were sampled at the turbocharger outlet (engine-out) and at each stack (tailpipe) using three portable, on-board systems 999 Freightliner truck, CAT 3406 engine, ~50,000 miles 8-hour extended idling test Idle Aire Technologies, Knoxville, TN, December 7, 200

31 Field test excavator with diesel particulate filter Emissions measured simultaneously upstream & downstream of DPF during a field test designed to mimic real-world operation Typical test run, construction equipment with diesel particulate trap PM emissions [mg/s] idle loaded idle loaded mode idle Engine-out Tailpipe 0:00 0:02 0:04 0:06 0:08 0:0 0:2 0:4 0:6 PM ~ 99% reduction Test run after prolonged idling: Elevated PM during high load operation PM emissions [mg/s] 0 Engine-out Tailpipe :00 0:02 0:04 0:06 0:08 0:0 0:2 0:4 0:6 PM

32 Example: Effect of driving style on particulate matter emissions 99 International school bus 95-hp DTA-360 turbodiesel Area of bubble proportional to emissions Each point = second of driving Normal driving Aggressive driving Very calm, zen master driving Source: Author data, personal research

33 Crude evaluation of overall diesel oxidation catalyst function by CO measurements Extreme case: Bus catalyst operational only during sustained high load operation Measurement is far from accurate (repair-grade analyzer used), but useful for the purpose. 0,9 Engine rpm 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0, Cold GPS_mph start rpm Heavy urban traffic Freeway 0 0 3:5 3:30 3:45 4:00 4:5 4:30 4: ,08 0,07 Green catalyst efficiency 00% 50% Emissions g/s emissions CO (g/s) 0,06 0% B_CO[g/s] 0,05 Red emissions upstream of DOC -50% P_CO[g/s] Blue d_co[%] emissions downstream of DOC 0,04-00% 0,03-50% 0,02-200% 0,0-250% 0-300% 3:5 3:30 3:45 4:00 4:5 4:30 4:45 Catalyst efficiency PTI vs. UB unit difference

34 Discussion PM sampling, PM measurement using light scattering Dilution at the sampling port upstream of aftertreatment adds to the complexity of the system No dilution, no heating of the sampling system secondary growth of PM in the sampling system, plus condensation of water vapor contained in the exhaust on particles Low concentration counting of individual particles possible but generally not possible without dilution High concentration aggregate number but some size information can be obtained with very fast sampling Response ~ d 6 for particles << wavelength of light Response ~ d 2 for particles >> wavelength of light Response ~ d 0 if particles grown to uniform size For soot, mass ~ d fd, where fd = fractal dimension of particles Complexity of issues requires careful use and thoughtful, applicationspecific calibration with other methods; still, results might not be accurate

35 Discussion accuracy of on-road measurements On-road testing Wide intermediate area Development and pilot deployment of new technologies HC HC [g/mile] [g/mile] Characterization of the effects Inspection programs High-volume, fast, lowcost field testing Goal is to identify high emitters (order-ofmagnitude deviations) Basic + applied research Low-volume laboratory testing Can be costly and complex Goal: Laboratory quality data test mileage test mileage Accuracy and other parameters of onboard test equipment and methods need to be carefully matched to the needs of the application.

36 Crude evaluation of overall particulate trap function by visual inspection Visible soot: Signs of mechanical problem No visible traces of soot = Trap function probably OK

37 Conclusions Aftertreatment device evaluation requires, after initial development but before mass deployment, testing of a large set of vehicles under a variety of conditions; such testing can be done on the road Studies done using a simple on-board system show that repeatable measurements can be obtained by replicating dynamometer driving cycles on a closed road; tests can be then done alternately with and without the aftertreatment device Ordinary operation (i.e., of a vehicle on the road) is generally poorly repeatable; in this case, measurements can be done by simultaneous sampling upstream and downstream of the aftertreatment device Studies done using a simple on-board system show that errors due to non-isokinetic sampling upstream of the device appear to be relatively small (compared to other sources of error) Simultaneous emissions measurements upstream and downstream of an aftertreatment device appear to be a feasible way of evaluating its efficiency in real-world operation.