New Measurement Techniques & Procedures for Measuring "Real World" Emissions with PEMS and PAMS

Transcription

1 New Measurement Techniques & Procedures for Measuring "Real World" Emissions with PEMS and PAMS Carl Fulper United States Environmental Protection Agency, OTAQ 1 PEMS Conference UC-CERT April 11, 2013

2 Outline 2 New Way to Think About How to Test in the Real World The Modeling Way Using the right tools for the operation to achieve data goals quickly, effectively and accurately Data Reports Data Validation Report QA/QC, time alignment & corrected Emissions Rpt Vehicle Representativeness Improve sample design and recruitment methods PEMS Installation Techniques

3 PEMS Testing in the Real World Measurement is difficult and expensive Vehicles and engines are diverse Vehicles are difficult to acquire or recruit (reduce burden to owne) Techniques are detailed and labor intensive Emissions are highly variable Most test programs install PEMS and owner drives around their normal route for 8-10 hours and test the vehicle over 3-5 days or longer. Represents real driving, but in the modeling world, you could be missing key operating modes or not get enough emission data Is there a better way? 4

4 What kind of operation? put the vehicle thru its paces On choreographed drive route: 5 Start Parking lot Residential street Arterial On-ramp Freeway Off-ramp Idle Grade (hills) or loads Accelerations Decelerations Cold/Hot Start This can be used for all mobile sources (HDVs, MDVs, LDVs and nonroads) in about 2-3 hours. For HDVs: load via trailer & hills For Nonroad: Running set operations

5 NOx (g/hr) Draft Data: Houston Drayage (weighted) vs. MOVES: MY MOVES Drayage mph mph 50+ mph inc. STP inc. STP inc. STP opmodeid

6 VSP Bin Frequency No Grade Shifts to higher or lower VSP bins because of grade (+ or -) With Grade VSP Bin (kw/ton)

7 Benefits: Full profile of a vehicle operations/emissions at moment in time. (no missing data profiles) More accurate emission data for mode of operation Reduces time per choreographed drive route/test Minimize possible failure of PEMS units Allows for scheduled zeros/spans for drift correction Allows for more better designed tests to be conducted. 8

8 PAMS: What parameters are needed? 9 Date/Time Stamp GPS OBD Parameters Normal Engine Speed (RPM) Vehicle Speed Load % Air/Fuel Ratio Catalyst Temperature Short-Term Fuel Trim Long Term Fuel Trim Ambient Temperature MAP or MAF Inlet Temperature OBD Codes Accelerator Position Others OBD Parameters Enhanced - Fuel Economy - Battery State of Charge (hybrid) - Others

9 Instruments and measurements: In-Use Activity Instrument: Logger (by HemData) Keys features: OBD/CAN & GPS (second-by-second) Sleep mode X,Y & Z Accelerometer LDVs/MDVs HDVs/Nonroad (J1939) 10

10 Skid-steer loader: Operation by Day of Week Measurement Period: Monday June 11, 2007 Monday, July 2, 2007 Total operation: 16.6 hours

11 Skid-steer loader: Operation by Hour of Day

12 PEMS vs PAMS When to use one of them or both? PEMS Great for Measure emissions, activity & vehicle parameters Labor intensive to install/uninstall & calibrate High cost PAMS Small unit & easy to install Only records vehicle parameters & GPS Much lower costs Captures real world activity/non-activity for long periods of time 13

13 Use Right Data Gathering Tool: PEMS vs PAMS (Statistical Power) 14 Portable Emission Measurement Systems (PEMS) Full emissions profile of vehicle at that moment of time Represents its aging (odometer, maintenance, environmental factors) Recruit, test, give back vehicle to owner in 2-3 days, allows for more tests per vehicle or more vehicles tested. Include simulated loads emission profiles Repeat recruitment to document vehicle s seasonal or longitudinal deterioration effects Portable Activity Measurement Systems (PAMS) Able to rematch real world activity profiles by vehicle OBD parameters back to emissions. Able to install more PAMS device on vehicles for 2-4 weeks or longer including seasonal effects.

14 How Good is Your Data? Do you know how good your data is from your PEMS unit? How long does it take to know the data gathered from the PEMS unit is good? 15 minutes? 30 minutes?, 1 hour? or hours? Quick Data Stream Validation is the key to your first primary determination on data gathering &what next test your going to conduct next. 15

15 One Solution: NI Diadem Software Generates a graph data report on all data streams to confirm or validate them. Look to confirm minimum/maximum range values Advance Features: Perform QA/QC procedures and provides quick indications that data with normal ranges or within min/max ranges. Allows for creating set reports that calculated emissions for quick comparisons between vehicles Ability to reuse reports for other test programs reducing data validation & QA/QC time and more time on analyzing emission results. 16

16 17 Gives full report on data parameters

17 PARTICULATE MEASUREMENT: CONTINUOUS AND AGGREGA PM Mass (Int. Filter Time) Second by second Particulate! Exhaust Flow Cold start : Peak not shown (2,700 μg/m 3!) Summing the second-by-second gives a back-up to mass on the filters Accumulating mass on three filters 18 DIAdem REPORT Layout

18 19

19 Why Worry about Vehicle Selection & Representativeness? In modeling world.. Is it critical to know the selection criteria for the vehicle(s) tested and how it fits within its like total population Tested LDV #1 Total State LDV population: N = 10,000 Tested LDV #2 20 One Company s LDV Fleet: N=300

20 Vehicle Selection: Is there a better way? Is there a way to screen large numbers of vehicles? Before selecting sample for laboratory measurement? By getting a quick inexpensive measure of emissions? That correlates with lab measurements Best candidate: remote sensing Has been demonstrated for LDV evaporative emissions Has been demonstrated for Heavy-duty vehicles Presently being conducted on Tier 2 RSD/PEMS correlation 21

21 Instruments & measurements: secondary screening NVFEL) Instrument: ESP 4000 HC/CO/NO (remote sensing): Tier 2 LDVs RSD/PEMS corelation 22

22 RSD Equipment 23 Conducted by University of Denver (two weeks in July 2009) Gaseous pollutants (CO2, CO, THC & NOx) Measured at entry gate of Barbour s Cut port Matched license plates to TX DOT database RSD readings: 4,032 Unique vehicles: 1,877

23 Sampling methodology: HDVs Classification Stratified sampling Population divided into separate classes based on RSD NOx emissions and model year groups Each class sampled as an independent sub-population Quota sampling within each class based on Probability that a vehicle is actually in the assigned NOx Bin Frequency that the vehicle will drive in the Port of Houston during PEMS & PAMS instrumentation 24

24 Percent (%) Sampling classes based on RSD scores 1 NOx Bin -2 (2.5%) NOx Bin -1 (20%) NOx Bin 0 (55%) NOx Bin 1 (20%) NOx Bin 2 (2.5%) RSD NOx z-value 1 Data Collection of Drayage Trucks in Houston-Galveston Port Area Draft Report. EP-C May 27, 2011

25 Emissions Analysis Assign second-by-second measurements from PEMS into MOVES operating mode bins Function of vehicle speed and Scaled Tractive Power (STP) Consist of idle, braking, coast, and cruise modes Additional detail provided in the Appendix Calculate average emission rates by model year groups 26

26 Emissions Analysis (cont d) Apply population weights based on RSD sampling methodology Why apply the weights? In order to capture high-emitting trucks in the sample, trucks were recruited with unequal probabilities of selection Thus, it is necessary to adjust the sample distribution to conform to the known distribution in the RSD population Trucks without RSD measurements excluded from the analysis Compare in the wild emissions data to MOVES emission rates for heavy-duty short-haul combination trucks 27

27 PEMS NOx (g/hr) PEMS NOx (g/hr) PEMS NOx (g/hr) PEMS NOx (g/hr) PEMS NOx vs. STP by RSD NOx Bin Cleaner MYG MYG STP MYG STP MYG STP STP Bottom line: correlation between RSD readings and PEMS measurements work but can be improved.

28 NOx (g/hr) Draft Data: Houston Drayage (weighted) vs. MOVES: MY MOVES Drayage mph mph 50+ mph inc. STP inc. STP inc. STP opmodeid

29 Conclusions The Tools (PEMS/PAMS) are there to measure mobile sources By improving are sampling techniques, selection criteria, data gathering techniques, analyses and use of PEMS/PAMS: Good data becomes better data Able to gathering more data (Statistical power) More Cooperative Research/Discussions are needed to support each other in Gathering Data 30

30 Gallery

31 32

32 PEMS Vehicle Example MY 1994 Freightliner Portable Emissions - SEMTECH DS Gaseous pollutants (CO2, CO, THC, NO & NO2), with exhaust flow Teflon membrane filters 47 C Portable Activity Isaac data loggers GPS and RPM (older vehicles) GPS & engine parameters (J1708 & J1939) PM Filters Semtech_DS PM Proportional Sampler System Exhaust System 33

33 PEMS partially inside with Filters/MPS outside cab PEMS filters and DS inside cab with MPS flowmeter outside 34

34 Typical PAMS installation to gather activity: 35 GPS & RPM (older vehicles) GPS & engine parameters (J1708/J1939)

35 36 Questions?

36 37 Appendix

37 Time Alignment? How are you determining time alignment of data from all data streams? Each analyzer has a different, OBD/CAN; flowmeter, GPS, environmental probe, proportional sampling & sampling lines lengths PEMS manufacturers or yourself need to establish procedures to document this properly for each equipment setup 38

38 Time Series: Engine Parameters Exhaust Flow Temperature 600 ( C) Exhaust Mass Flow (SCFM) Engine Speed (RPM ) ,000 2,000 3,000 4,000 5,000 6,000 Elapsed Time (Sec) 7,000 8,000 9,000 10,000