Evaluation methods for RDE Outcome of the lessons learnt and proposal for RDE4 package RDE-LDV Meeting of 7 November 2017 1
Terminology «Trip normalisation» is intended as verification that the trip characteristics fall within the boundary conditions which include several elements: Altitude and cumulative altitude, Ambient temperature, overall and excess of driving dynamics, etc... «Emissions normalisation» is indented as the potential correction of emissions values if the trip is found ex post to be partially outside the boundary conditions and/or far from the WLTP reference Correction factor for extended temperature & altitude ranges Emissions corrections for MAW outside the normal range 2
What is a «normal RDE trip»?? A normal RDE trip shall not be intended as trip representative for the WLTP and it may potentially explore areas of ambient conditions and driving conditions not covered by the WLTP, but still within the agreed boundaries. A single RDE trip is therefore a mix of: Urban, Rural, Motorway driving conditions (ensuring that the behaviour related to the U/R/M dynamics is captured for all). The actual temperature and altitude conditions at the test location A single test can therefore only represent a snapshot (or snapshots) of the possible operating conditions of the vehicle. 3
* Biased (or anbnormal driving) should be detected by the trip normalisation step, prior to any emissions calculation 4
ITEM 1 TRIP NORMALISATION 5
Trip completeness checks Element ICE PHEV U/R/M Distance shares (Annex IIIa) + MAW X X Stop Share (Annex IIIa) X X Urban minimum Distance with ICE (Annex IIIa) X Trip normality checks (Annex IIIa +) Element ICE PHEV Overall driving dynamics (Appendices 5 and 6) X Excess of dynamics (RPA and V*Apos) (Appendix 7a) X X Ambient temperature X X Altitude range X X Cumulative Altitude (Appendix 7b) X X Urban Average Speed X X Motorway Speed X X 6
Lessons learnt Main Trip Normalisation Issues: Data submitted by stakeholders has shown that: Some trips which were normal according to most of the criteria failed completeness checks (quite often for the MAW urban) Some abnormal driving (Emission Provoking Driving = extreme accelerations, wrong gear-shifts,.) was not detected by the present data evaluation (i.e. The combination of all criteria in the second table on the previous slide) and invalid trips result as valid from the Overall driving dynamics check 7
RDE3 Data evaluation Issues - Example 1 (TUG) Normal tests eliminated by EMROAD criterion of normality While not eliminated by PB or V*Apos 8
Does Appendix 7a identify Emission Provoking driving? Based on TUG PEMS testing: same vehicle ND vs EPD driving RPA & V*a pos95 as reported by EMROADv5.96B1 TUG01 ND EPD U RPA 0.23 0.19 0.17 0.15 0.17 R RPA 0.16 0.13 0.11 0.11 0.09 M RPA 0.07 0.06 0.05 0.09 0.03 U v*a 17.08 13.85 10.58 10.82 10.84 R v*a 20.71 17.99 11.61 13.08 10.06 M v*a 19.15 14.38 24.77 19.30 23.19 TUG02 ND EPD 0.17 0.21 0.16 0.19 0.18 0.14 0.16 0.12 0.19 0.15 0.10 0.10 0.05 0.07 0.09 12.93 16.86 12.09 15.05 12.93 19.38 24.12 16.55 24.20 19.92 24.86 26.64 14.60 23.40 16.22 Out of 6 EPD only 1 is identified as dynamic Emission provoking driving can be achieved fulfilling the trip dynamic criteria of Appendix 7a v*a_pos95 is appropriate to identify dynamic driving but not emission provoking driving Therefore EPD should be identified with the evaluation methods
Trip completeness checks Element ICE PHEV U/R/M Distance shares (Annex IIIa) + MAW X X Stop Share (Annex IIIa) X X Urban minimum Distance with ICE (Annex IIIa) X Trip normality checks (Annex IIIa +) Element ICE PHEV Overall driving dynamics (Appendices 5 and 6) X Excess of dynamics (Appendix 7a) X X Ambient temperature X X Altitude range X X Cumulative Altitude (Appendix 7b) X X Urban Average Speed X X Motorway Speed X X 10
Description of revised normality checks considered NB: Options left after the audio-web meeting on October 23 1. Baselines: Appendix 5 2. MAW SF1-1-1 = MAW as described in Appendix 5 setting the scaling factors to 1 Checks for completeness ( 15% MAW per phase) + normality ( 50% normal MAWs per phase) 3. MAW SF1-1-1 = MAW as described in Appendix 5 setting the scaling factors to 1 Checks for completeness ( 15% MAW per phase) + RSI (*) to check the normality (*) RSI definition as proposed by ACEA in September 2017
Further justification for MAW SF1-1-1 - CO2 Reference curve and tol1 The reference CO2 values are from the WLTP. The on-road MAW CO2 variability caused by different factors shall be captured by tol1. Factors causing the differences between the reference CO2 values and the on-road MAW CO2 values (non exhaustive): Vehicle speed-acceleration Vehicle payload Ambient temperature Road grade Vehicle systems consuming and/or recovering electric power Etc... 12
13
Reference data used to Assess the revised normality checks: Source VehicleID Fuel Transmission Rated power [kw] Emission control technologies Designed as RDE-compliant # trips ND: normal driving CS: charge sustaining DD: dynamic driving CD: charge depleting EPD: emission provoking driving BC: battery charge LED: low emission driving * loaded vehicle Other PEMS testing TUG TUG01 Diesel Automatic 140 DPF+SCR 5 1DD +6Alpine: 3 EPD + 1 DD + 2 ND TUG TUG02 Diesel Manual 90 DPF+SCR 5 9Alpine: 5 EPD, 4*, 2 DD + 2 ND JRC PHEV Gasoline/ Electric Automatic 110+75 TWC 1 (CS) 3 (2*CD + 1 BC) JRC Veh01 Gasoline Manual 70 TWC 6 3 (DD) JRC Veh02 Diesel Automatic 193 DOC+EGR+DPF+ SCR 5 2 (DD) JRC Veh03 Diesel Automatic 110 EGR+LNT+DPF 6 2 (DD) JRC Veh04 Gasoline Automatic 140 TWC 6 - JRC Veh05 Gasoline Manual 51 TWC 6 2 (DD) JRC Veh06 Diesel Manual 73-LCV DOC+EGR+DPF+ SCR 4+2* 2 (DD) JRC Veh07 Diesel Automatic 130 EGR+DPF 3 - JRC Veh08 Diesel Automatic 195 DOC+SCR+LNT+DPF 6 2(DD) JRC Veh09 Gasoline Manual 78 TWC 5 2(DD) OEM Veh10 Diesel Automatic 420 EGR+DPF 2 - TWC: Three-way catalytic converter 62 34 DOC: diesel oxidation catalyst DPF: diesel particle filter SCR: selective catalytic reduction EGR: exhaust gas recirculation LNT: lean-nox trap LCV: light commercial vehicle
TUG01 Revised normality checks performance: MAW (EMROAD) MAW SF 1-1-1 MAW SF 1-1-1 + RSI Vehicle Route type Driving style Validity Validity Validity Alpine ND Uc+Rn Uc+U&Rn U&Rn+U&R RSI Alpine ND Uc+Rn Uc+U&Rn U&Rn+U&R RSI Alpine DD Mc+U&Rn Mc+U&Rn U&Rn+U&R RSI Alpine EPD Un U&Rn U&Rn+U&R RSI Alpine EPD Uc+U&Rn Uc+U&Rn U&Rn+U&R RSI Alpine EPD Uc+Un Uc+U&Rn U&Rn+U&R RSI Hilly ND Hilly ND Un Un+U&M RSI Hilly ND Mn 25 +U&M RSI Hilly ND Rn 25 +U&R&M RSI Hilly DD U&R RSI Hilly LED ND: normal driving DD: dynamic driving EPD: emission provoking driving LED: low emission driving c: completeness (>15% MAW) n: normality (>50% normal MAW) RSI: relative severity index (<20%) out: outliers share (<20%) 25 Normality check with tol1+ fixed at 25% RSI set at 20% is very stringent and it invalidates normal driven trips
TUG02 Revised normality checks performance: MAW (EMROAD) MAW SF 1-1-1 MAW SF 1-1-1 + RSI Vehicle Route type Driving style Validity Validity Validity Alpine ND Uc Uc+Un Un+U&R RSI Alpine ND Uc+Mn Uc+U&Mn U&R 25 &Mn+U&R RSI Alpine DD Uc+Rn Uc+U&Rn U&Rn+U&R RSI Alpine DD U&Mc+Rn U&Mc+U&Rn U&Rn+U&R RSI Alpine EPD Uc Uc R RSI Alpine EPD* Uc Uc+Un U&Rn 25 +U&R RSI Alpine EPD* Uc+R&Mn Uc+U&R&Mn U&R&Mn+U&R&M RSI Alpine EPD* U&Mc U&Mc R RSI Alpine EPD* U&Mc U&Mc+Un Un+U&R RSI Hilly ND Hilly ND Hilly ND Hilly ND Hilly ND ND: normal driving DD: dynamic driving EPD: emission provoking driving LED: low emission driving c: completeness (>15% MAW) n: normality (>50% normal MAW) RSI: relative severity index (<20%) out: outliers share (<20%) 25 Normality check with tol1+ fixed at 25% Some ND tests are eliminated because of urban section
Veh06 Veh04 Veh03 Veh02 Veh01 Revised normality checks performance: MAW (EMROAD) MAW SF 1-1-1 MAW SF 1-1-1 + RSI Vehicle Validity Validity Validity R&Mn + R&M RSI JRC Dataset Normal Driving Un Un Un Rn Un Un Un Un Un MAW SF 1-1-1 found efficient on JRC normal trips Does not eliminate any normal tests
Flat driving Relative Severity Index (RSI) Verification On JRC normal trips, driven on relatively flat terrain, the RSI threshold set at 20% rarely invalidates trips. However, on trips driven on hilly routes, complying with cum pos elev gain, RSI 20% is too strict. 30% RSI seems a good compromise to differentiate abnormal from normal driving
Proposals for trip normality Part 1 1. Completeness check for MAW to be eliminated Reason: Redundancy with the design of U/R/M distance shares requirements in Annex IIIa. 2. Verification of trip normality with MAW - Correction of the WLTP scaling factors with 1-1-1 - Higher value for tol1-40% to +45% for urban part - 50% normality criterion or modified RSI (TBD)? Parametric study of tol1 and RSI Reason for the modification: Many normal tests resulted as invalid and vice-versa under RDE3 19
Proposals for trip normality Other issues Relative Positive Acceleration (RPA) requirement could be eliminated Reason: No indication that trips are invalidated by this criteria, i.e. trips are driven with adequate dynamics 20
Proposals for trip normality Other issues Urban and cold start average speed: The lower bound could be lowered to 10 km/h Reason: Real average urban speeds can be lower than 15 km/h and there is no reason to invalidate trip because the average speed is too low, if the stop time requirements are respected. Introduce flexibility in stop requirements (if there is a 300s stop, then eliminate 300s following this) 21
ITEM 2 EMISSIONS CALCULATION 22
Emissions Evaluation Lessons learnt Current calculations in RDE3 correcting emissions for ICE and PHEVs to be potentially revised Reason: Non-systematic relationship for the emissions corrections between the current two tools, EMROAD and CLEAR (Ref: TNO, TUG) Potential problem: Not using the urban MAWs to calculate NOx urban emissions may over-estimate the NOx emissions, due to the merging of very high non-urban accelerations with the urban data. Solution: Modification of the U/R/M speed based definition to a «First acceleration definition», like for heavy-duty 23
Definition of first acceleration in HDV legislation Urban, rural and motorway parts can be determined either on the basis of: geographical coordinates (by means of a map), or first acceleration method. In case the trip composition is determined by means of the first acceleration method, the first acceleration above 55 km/h (70 (or 60?) km/h in the case of vehicles of categories M 1 and N 1 ) shall indicate the beginning of the rural part and the first acceleration above 75 km/h (90 km/h in the case of vehicles of categories M 1 and N 1 ) shall indicate the beginning of the motorway part. Appropriate coverage of the velocities shall be assured as before. 24
Emissions Evaluation For a trip evaluated as normal: Baseline: The MAW emissions calculation Option 1: Raw emissions (Total & Urban) Option 2: Raw emissions (Total & Urban) corrected by the ratio RDE_CO2/WLTP_CO2 (with urban CO2 use CO2 from WLTP phases 1+2) 25
JRC Veh06 Complete trip NO x invalid
TUG02 Complete trip NO x invalid
TUG02 Urban NOx
Option 2 - Emissions link with CO2? Option 2 corrects emissions downwards (in most but not all cases) For tailpipe this is not systematic because state-of-the-art emission controls can deal with high emissions. In the case of NOx, the CO2 vs. NOx relationship is complex and depends on the technology package (Lots of litterature sources to confirm)
Option 2 - Emissions link with CO2? No systematic relationship between CO2 and emissions when proper emission controls are applied Example for a modern Euro 6 (diesel) vehicle (with EGR+SCR+LNT) with tests performed at the JRC
PROPOSED IMPROVEMENTS 31
Proposed Improvements Trip normalisation: MAW completeness checks not needed, provided that the trip completeness is verified by Annex 3A with U/R/M distance shares Correction of MAW CO2 reference curve (1-1-1) and trip normality verification (with 30% RSI or 50% criterion, TBD) Increase of urban tol1 value, in particular to cope with LCVs (CO2 spread expected to be larger due to payload) No need to have lower bounds for trips (RPA lower bound to be deleted)? 32
Proposed Improvements Emissions calculation: Modification of the U/R/M speed based definition: first acceleration method (with velocities tbd, as for HDVs) or MAW average speed? Calculation of emissions: Pollutant/CO2 for intermediate RAW after evaluation of more data, also for PN, with intro at a later date 33