Euro6d RDE EVALUATION METHODS / HYBRIDS > Package ...

78
Euro6d RDE EVALUATION METHODS / HYBRIDS > Package#4 Proposals > 12Sep’17 ACEA TF-RDE SUB-Group EVALUATION METHODS / HYBRIDS

Transcript of Euro6d RDE EVALUATION METHODS / HYBRIDS > Package ...

Euro6d RDE EVALUATION METHODS / HYBRIDS > Package#4 Proposals > 12Sep’17

ACEA TF-RDE SUB-Group EVALUATION METHODS / HYBRIDS

2

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP INTRODUCTION

o Further to our discussion on 03Aug17 the ACEA RDE Evaluation Methods / Hybrids team has been progressing with the studies as mentioned, investigating the latest data that we have available and focussing on 2 key topics:

o Trip Validity – enabling avoidance of intentionally biased driving.

o Trip Normalisation – provision of a single results calculation method.

o With the aim of developing proposals to address the most significant challenges identified in the Evaluation Methods considered so far.

o We would be very grateful for the opportunity to review and discuss these proposals, with the aim of identifying a pragmatic route forwards for Euro6d RDE Package#4.

3

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP

Euro6d RDE – Package#4

Trip Validity

4

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP VALIDITY – FUNDAMENTAL REQUIREMENT

o Following from previous discussions, ACEA’s position continues to be that Euro6d RDE for LDVs incorporates Trip Validity (Completeness and Normality) confirmation as a fundamental principle – with the aim of avoiding “... that the tested vehicles are driven in a biased manner, i.e. with the intention to generate a passed or failed test not by virtue of the technical performance of the vehicle but due to extreme driving patterns.”

o Further, as presented at the 07Apr RDE-LDV meeting and discussed on 30May17, our view is that the vehicle kinematic boundary conditions alone are not sufficient to achieve this aim. And it can be demonstrated that it is possible to drive an intentionally biased RDE trip while still complying with the basic trip and kinematic boundary conditions.

o Therefore it is necessary to retain a further trip validity check, based on analysis of CO2, to assess the effect of intentional driving style on the engine – as per AnnexIIIa 5.4.2

5

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP VALIDITY – SUCCESSFUL ASSESSMENT

o Assessment or Trip Validity criteria options should consider how well they achieve the goal of identifying genuinely biased trips, while not excluding acceptable trips. Reviewing how many trips are excluded from any given set is of no value if the study does not go on to consider whether they should have been identified as invalid.

o However we do also recognise that the current (MAW CO2 window) / previous (PB ‘ranges around reference’) assessments would appear to be excluding too many trips – and probably some of the wrong trips, e.g...

o Therefore further consideration is required...

Valid ~90% Invalid ~10%

Excluded ~20%

6

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Moving Averaging Windowing – Wj value for each window determined by CO2 vs. Characteristic Curve...

Power Binning – results “evaluation by normalisation to a standardised power frequency distribution”... (from CO2, unless wheel hub torque measured)

Plus CO2 ratio introduced for PHEVs in CS mode.

Appendix 6 Verification of trip dynamic conditions with method 2 (Power Binning)

Appendix 5 Verification of trip dynamic conditions with method 1 (Moving Averaging Window)

Driver-Vehicle

biased driving Trip Requirements Route, Ambient …

PEMS: CO2, NOx, PN …

Urban roads (low speed) Rural roads (medium speed) Motorways (high speed) Low and high altitude Up- and down-hill driving Additional vehicle payload Ambient conditions …

Annex IIIA, Appendix 7a & 7b

7

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Annex IIIA, Appendix 7a & 7b Moving Averaging Windowing – Wj value for each window determined by CO2 vs. Characteristic Curve...

Power Binning – results “evaluation by normalisation to a standardised power frequency distribution”... (from CO2, unless wheel hub torque measured)

Plus CO2 ratio introduced for PHEVs in CS mode.

Appendix 6 Verification of trip dynamic conditions with method 2 (Power Binning)

Appendix 5 Verification of trip dynamic conditions with method 1 (Moving Averaging Window)

Urban roads (low speed) Rural roads (medium speed) Motorways (high speed) Low and high altitude Up- and down-hill driving Additional vehicle payload Ambient conditions …

One Method for verifying the

normality

One Calculation for the final emissions

Driver-Vehicle

biased driving Trip Requirements Route, Ambient …

PEMS: CO2, NOx, PN …

8

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Modified MAW OR

Comb. MAW&PB

CO2 Weighting for Conven. & Hybrids

Annex IIIA Appendix 7a Appendix 7b

Focus of today's presentation

Driver-Vehicle

biased driving Trip Requirements Route, Ambient …

PEMS: CO2, NOx, PN …

9

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Modified MAW CO2 Weighting for Conven. & Hybrids

Annex IIIA Appendix 7a Appendix 7b

Driver-Vehicle

biased driving Trip Requirements Route, Ambient …

PEMS: CO2, NOx, PN …

10

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving Definition of CO2 Characteristic Curve as the reference for average driving pattern

50

100

150

200

250

300

350

0 20 40 60 80 100 120 140 160

CO

2 in

g/k

m

Geschwindigkeit in km/h

CO2+TOL2

CO2+TOL1

CO2 Ref

WLTP CO2

CO2-TOL1

CO2-TOL2

PEMS CO2

WLTC-reference RDE #2 • NEDC road load coefficients • + scaling factors

Correct definition of CO2 Characteristic Curve is the essential base

50

100

150

200

250

300

350

0 20 40 60 80 100 120 140 160

CO

2 in

g/k

m

Geschwindigkeit in km/h

CO2+TOL2

CO2+TOL1

CO2 Ref

WLTP CO2

CO2-TOL1

CO2-TOL2

PEMS CO2

WLTC-reference RDE#3 • WLTP road load coefficients • + scaling factors

50

100

150

200

250

300

350

0 20 40 60 80 100 120 140 160

CO

2 in

g/k

m

Geschwindigkeit in km/h

CO2+TOL2

CO2+TOL1

CO2 Ref

WLTP CO2

CO2-TOL1

CO2-TOL2

PEMS CO2

WLTC-reference RDE-LDV Nov’16 • WLTP road load coefficients • no scaling factors

Suggested at

Therefore scaling factors should not be applied to the WLTP CO2 values used in Appendix 5

The validity check is already part of the current legislation: ⇒ Appendix 5: Verification of trip dynamic conditions with method 1 (Moving Averaging Window)

11

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP

1.) Completeness ≥ 15% The test shall be complete when it comprises at least 15 % of urban, rural and motorway windows, out of the total number of windows

2.) Normality ≥ 50% The test shall be normal when at least 50 % of the urban, rural and motorway windows are within the primary tolerance (TOL1) defined for the characteristic curve.

Validity check – how to exclude abnormal driving

Do these 2 criteria exclude abnormal driving?

12

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Completeness –check is not capable to exclude abnormal driving

1.) Completeness ≥ 15%

Validation of horizontal distribution (check of speed classification)

TRIP REQUIREMENTS 6.6 urban 29-44 % rural 23-43 % motorway 23-43 % [classified by speed]

Generally covered in ANNEX IIIA :

TRIP REQUIREMENTS 6.12 minimum in urban, rural and motorway, shall be 16 km

Required only in MAW due to MAW-specific window speed classification correct calculation of emissions for the total trip requires a representative share Mtotal = 0.34*Murban + 0.33*Mrural + 0.33*MMotorway

13

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Normality–check excludes a large amount of abnormal driving! but excludes not all cases of abnormal driving therefor an additional criteria is required

2.) Normality ≥ 50% Validation of vertical distribution (check of extreme window amount)

But missing information about position of minority (a little bit outside or far off?)

The majority of windows shall be within Tol1-range (WLTP describes high frequency distribution of driving)

14

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Normality & RSI excludes most of abnormal driving!

3.) Relative-Severity -Index (RSI) ≤ 20% Validation of average of all window-severities

∅ Relative CO2

window position in %

Window-severity in %

100%= reference

level

25%

0%

50%

75%

50% 150%125%= Upper

Tol1

75%= Lower

Tol1

Window severity S in %

Relative CO2 window position

h in %

conventional PHEV & HEV No severity within Tol1 This is normal driving!

Extreme due to low CO2 [g/km] (CO2 < Tol1 lower)

different for conventional & PHEV/HEV

Extreme due to high CO2 [g/km] (CO2 > Tol1 upper)

③ ①

conventional vehicle only

The test shall be declared as invalid when Relative-Severity-Index of urban, rural or motorway windows is above 20%.

RSI is the average of window severity

RSIk = 1Nk

ΣSj,k k=u,r,m Sj = 𝑓𝑓 hj Window severity is a function

of window position

Tol1upper=125% Tol1lower= 75%

Tol1upper=150% Tol1lower=N/A%

15

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Applying Normality & RSI increase the amount of valid tests!

Analysing the influences ⇒ How much of tests becomes invalid by validity check only?*)

Valid tests in %

70%

60%

80%

100%

Basis

90%

Current

75%

86%

Normality & Normality-index

231 tests 202 tests

269 tests

25% 14%

+11%

*) numbers based on OEM internal databases

Normality & Completeness exclude too many and unessential tests

Increase of valid tests

Normality & Completeness

16

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #04: currently unnecessarily excluded passed by Normality & RSI

Assessing trip#04 from ACEA calculation round robin

Current situation:

Completeness ≥ 15% Normality ≥ 50%

failed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

17

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #08: currently passed failed by Normality & RSI

Assessing trip#08 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

failed

18

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Assessing trip#02 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

failed

Trip #02: currently passed (due to applied “NEDC” scaling factors [1.2/1.1/1.05] for WLTP)

failed by Normality , ok by RSI

19

EURO6D-RDE EVALUATION METHODS EVALUATION METHODS ASSESSMENT

1) The window size (wltp_c02/2) is too large and gives an opportunity for misuse driving strategy (i.e. misuse not only for short events). 2) Works for Hybrids only without lower tolerance (mixing many electrical and ice driven km into the same window (short misuse events remain undetected).

Criteria Moving Ave. Win. (MAW)

Power Binning (PB)

Modified MAW

Combined MAW & PB

Efficiency and accuracy of the method + + + Normalisation against CO2 + o + Robustness against Defeat Device practices + + + Robustness against intentional misuse 0 − 0 1)

Applicability for all technologies − − 0 2)

No random effects − − + Practicality, ease of use, ease of correcting it + + + No windows exclusion − − +

20

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Combine MAW&PB CO2 Weighting for Conven. & Hybrids

Annex IIIA Appendix 7a Appendix 7b

Driver-Vehicle

biased driving Trip Requirements Route, Ambient …

PEMS: CO2, NOx, PN …

21

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Which WLTP experiences could help towards an unique concept? One Tool

22

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Meta Model

Veline Approach in RDE Appendix 6 (Power Binning)

23

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Veline Approach in WLTP

24

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Which RDE experiences could help towards an unique concept? One Tool

25

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

P1

P2

P4

P3

Vehicle CO2 Line (Veline)

Vehicle CO2 characteristic curve

P1 – P4 from WLTP

P1, P3 & P4 from WLTP Which features could be joined

together?

Moving Averaging Window

Power Binning

One Tool

26

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Main Features joined together

Moving Averaging Window

Power Binning

One Tool

Veline Approach for the Characteristic Curve.

Tolerances (Tol.1, Tol.2) around the Characteristic Curve.

CO2 test results in the four phases of the WLTP.

Calculation of the moving averages of the instantaneous test data.

Trip Severity Index calculation.

27

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Pre Check: Does the new tool represent WLTP good enough? New Tool

28

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

P1

P2

P4

P3

Vehicle CO2 Line (Veline)

Vehicle CO2 characteristic curve

P1 – P4 from WLTP

P1, P3 & P4 from WLTP

WLTP

WLTP

Correlation quality

assessment!

P1

P4 P3

P1

P2

P4

P3

Moving Averaging Window

Power Binning

29

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

0 1000 2000 3000 4000 5000 6000 7000 8000 9000

time [sec]

0

50

100

150

spee

d [k

m/h

]

low

medium

high

extra high

RDE Trip = 5 x WLTC

TOTAL URBAN

WLTP by the new method.

-10 0 10 20 30 40 50

P wheels [kW]

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

CO

2 [g

/h]

10 4

RDE MAWs

WLTP p1-p4

Veline w/o Tol

Veline tol 1

Veline tol 2

-1

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

CO

2 [g

/h]

30

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

𝑃𝑃𝑤𝑤,𝑖𝑖 =𝑣𝑣𝑖𝑖

3.6∙ 𝑓𝑓0 + 𝑓𝑓1 ∙ 𝑣𝑣𝑖𝑖 + 𝑓𝑓2 ∙ 𝑣𝑣𝑖𝑖2 + 𝑇𝑇𝑇𝑇 ∙ 𝑎𝑎𝑖𝑖 + 𝑻𝑻𝑻𝑻 ∙ 𝒈𝒈 ∙ 𝒑𝒑

P1

P2

P4

P3

𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾 𝑷𝑷𝑾𝑾,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾; 𝑻𝑻𝑪𝑪𝑪𝑪𝑪𝑪,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

𝑷𝑷𝑾𝑾,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

𝑻𝑻𝑪𝑪𝑪𝑪𝑪𝑪,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

normal

severe

soft

-10 0 10 20 30 40 50

P wheels [kW]

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

CO

2 [g

/h]

10 4

RDE MAWs

WLTP p1-p4

Veline w/o Tol

Veline tol 1

Veline tol 2

-1

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

CO

2 [g

/h]

31

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

𝑃𝑃𝑤𝑤,𝑖𝑖 =𝑣𝑣𝑖𝑖

3.6∙ 𝑓𝑓0 + 𝑓𝑓1 ∙ 𝑣𝑣𝑖𝑖 + 𝑓𝑓2 ∙ 𝑣𝑣𝑖𝑖2 + 𝑇𝑇𝑇𝑇 ∙ 𝑎𝑎𝑖𝑖 + 𝑻𝑻𝑻𝑻 ∙ 𝒈𝒈 ∙ 𝒑𝒑

𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾 𝑷𝑷𝑾𝑾,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾; 𝑻𝑻𝑪𝑪𝑪𝑪𝑪𝑪,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

𝑷𝑷𝑾𝑾,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

𝑻𝑻𝑪𝑪𝑪𝑪𝑪𝑪,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

normal

severe

soft

Which window definition should be used?

-10 0 10 20 30 40 50

P wheels [kW]

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

CO

2 [g

/h]

10 4

RDE MAWs

WLTP p1-p4

Veline w/o Tol

Veline tol 1

Veline tol 2

-1

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

CO

2 [g

/h]

32

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

𝑃𝑃𝑤𝑤,𝑖𝑖 =𝑣𝑣𝑖𝑖

3.6∙ 𝑓𝑓0 + 𝑓𝑓1 ∙ 𝑣𝑣𝑖𝑖 + 𝑓𝑓2 ∙ 𝑣𝑣𝑖𝑖2 + 𝑇𝑇𝑇𝑇 ∙ 𝑎𝑎𝑖𝑖 + 𝑻𝑻𝑻𝑻 ∙ 𝒈𝒈 ∙ 𝒑𝒑

𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾 𝑷𝑷𝑾𝑾,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾; 𝑻𝑻𝑪𝑪𝑪𝑪𝑪𝑪,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

𝑷𝑷𝑾𝑾,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

𝑻𝑻𝑪𝑪𝑪𝑪𝑪𝑪,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

normal

severe

soft Win. Def. Size Conventional Hybrids

1) CO2 CO2 WLTP / 2 + − 2) Time e.g. = 3s o o 3) Distance e.g. = 500m + + Other

1) For Hybrids the window size is too large -> mixing of many electrical and ice driven km. Leads to invalid trips classified as too “SOFT”.

2) The time share of the urban part is predominant. (s. 6. TRIP REQUIREMENTS / 6.12. The minimum distance of each operation: urban, rural and motorway, shall be 16 km).

3) This definition is oriented to the main purpose of a vehicle (drive from A to B).

-10 0 10 20 30 40 50

P wheels [kW]

-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

CO

2 [g

/h]

10 4

RDE MAWs

WLTP p1-p4

Veline w/o Tol

Veline tol 1

Veline tol 2

33

EURO6D-RDE EVALUATION METHODS VERIFICATION OF TRIP DYNAMIC CONDITIONS - ICEV, MHEV

P0 ref,Veline

normal

severe

soft

tolerances vehicle type

tolerance 1 lower / upper

tolerance 2 lower / upper

ICE / MHEV –25 / +25 –50 / +50

P3 > tol1 & < tol2

P4 > tol2

P2 > tol1 & < tol2

P1 > tol2

Severity Index

P(1-4) = abs(( Px - P0 ref ) / P0 ref)

km driven with engine off [overrun fuel cutoff, sailing, electrical drive (Hybrids)]

𝑃𝑃𝑤𝑤,𝑖𝑖 =𝑣𝑣𝑖𝑖

3.6∙ 𝑓𝑓0 + 𝑓𝑓1 ∙ 𝑣𝑣𝑖𝑖 + 𝑓𝑓2 ∙ 𝑣𝑣𝑖𝑖2 + 𝑇𝑇𝑇𝑇 ∙ 𝑎𝑎𝑖𝑖 + 𝑻𝑻𝑻𝑻 ∙ 𝒈𝒈 ∙ 𝒑𝒑

𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾𝑾 𝑷𝑷𝑾𝑾,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾; 𝑻𝑻𝑪𝑪𝑪𝑪𝑪𝑪,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

𝑷𝑷𝑾𝑾,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

𝑻𝑻𝑪𝑪𝑪𝑪𝑪𝑪,𝑾𝑾,𝒎𝒎𝒎𝒎𝒎𝒎𝑾𝑾

The combustion engine shall be considered as deactivated if: engine speed <50 rpm; exhaust mass flow rate <3 kg/h; exhaust mass flow rate drops to <15% of the steady-state exhaust mass flow rate at idling.

Severity Index

for all Px within Tolerance_1 = 0%

34

EURO6D-RDE EVALUATION METHODS VERIFICATION OF TRIP DYNAMIC CONDITIONS - ICEV, MHEV

Tolerances around the Vehicle CO2 Line (Veline) The primary tolerance and the secondary tolerance around the vehicle CO2 line are respectively tol1= 25% and tol2 = 50%.

1. Verification of test completeness The test shall be complete when it comprises at least 15% of urban, rural and motorway windows, out of the total number of windows.

2. Verification of test normality The test shall be normal when at least 50% of the urban, rural and motorway windows are within the primary tolerance defined for the Veline. If the specified minimum requirement of 50% is not met, the upper positive tolerance tol1 may be increased by steps of 1% until the 50% of normal windows target is reached. When using this mechanism, tol1 shall never exceed 30%.

3. Verification of max. amount of extreme windows The trip is valid since max. 20% of the total, urban, rural and motorway windows are outside the secondary tolerance.

4. Verification of test severity The trip is valid since the Severity Indices for the total, urban, rural and motorway are ≤ 20%.

-10 0 10 20 30 40 50

P wheels [kW]

-1

0

1

2

3

4

5

CO

2 [g

/h]

10 4

RDE MAWs

WLTP p1-p4

Veline w/o Tol

Veline tol 1

Veline tol 2

35

EURO6D-RDE EVALUATION METHODS VERIFICATION OF TRIP DYNAMIC CONDITIONS - NOVC, OVC HEV

km driven with engine off [overrun fuel cutoff, sailing, electrical drive (Hybrids)]

normal

severe

soft

tolerances vehicle type

tolerance 1 lower / upper

tolerance 2 lower / upper

ICE / MHEV –50 / +50 –75 / +75

wider range to cover load-point shifting

P_drag = -0.01*P_ice -0.005*P_ed

WLTP CS CO2 combined Two additional points for the “veline” P_add_1: WLTP CS CO2 combined P_add_2: P_drag / CO2-Value = 0g/km

P0 ref,Veline

P3 > tol1 & <= tol2

P4 > tol2

P2 > tol1 & <= tol2

P1 > tol2

The combustion engine shall be considered as deactivated if: engine speed <50 rpm; exhaust mass flow rate <3 kg/h; exhaust mass flow rate drops to <15% of the steady-state exhaust mass flow rate at idling.

Severity Index

P(1-4) = abs(( Px - P0 ref ) / P0 ref)

Severity Index

for all Px within Tolerance_1 = 0%

36

EURO6D-RDE EVALUATION METHODS VERIFICATION OF TRIP DYNAMIC CONDITIONS - FHEV, PHEV (CD & CS MODE)

Tolerances around the Vehicle CO2 Line (Veline) The primary tolerance and the secondary tolerance around the vehicle CO2 line are respectively tol1= 50% and tol2 = 75%.

1. Verification of test completeness The test shall be complete when it comprises at least 15% of urban, rural and motorway windows, out of the total number of windows.

2. Verification of test normality The test shall be normal when at least 50% of the urban, rural and motorway windows are within the primary tolerance defined for the Veline. If the specified minimum requirement of 50% is not met, the upper positive tolerance tol1 may be increased by steps of 1% until the 50% of normal windows target is reached. When using this mechanism, tol1 shall never exceed 60%.

3. Verification of max. amount of extreme windows The trip is valid since max. 20% of the total, urban, rural and motorway windows are outside the secondary tolerance.

4. Verification of test severity The trip is valid since the Severity Indices for the total, urban, rural and motorway are ≤ 20%.

37

Cons • This method is more complex

than mod. MAW Method. • Short time period for testing due

to all involved parties.

Pros • Efficiency and accuracy of the

method • Normalisation against CO2 • Robustness against intentional

misuse • Applicability for all technologies • Matlab App (free) available

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

38

EURO6D-RDE EVALUATION METHODS EVALUATION METHODS ASSESSMENT

Criteria Moving Ave. Win. (MAW)

Power Binning (PB)

Modified MAW

Combined MAW & PB

Efficiency and accuracy of the method + + + + Normalisation against CO2 + o + + Robustness against Defeat Device practices + + + + Robustness against intentional misuse 0 − 0 1) + Applicability for all technologies − − 0 2) + No random effects − − + + Practicality, ease of use, ease of correcting it + + + + No windows exclusion − − + + 1) The window size (wltp_c02/2) is too large and gives an opportunity for misuse driving strategy (i.e. misuse not only for short events). 2) Works only without lower tolerance, mixing many electrical and ice driven km into the same window (short misuse events remain undetected).

39

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP

Euro6d RDE – Package#4

Trip Normalisation

40

EURO6D-RDE EVALUATION METHODS PROPOSAL FOR NEW EVALUATION METHOD

Modified MAW OR

Comb. MAW&PB

CO2 Weighting for Conven. & Hybrids

Annex IIIA Appendix 7a Appendix 7b

Driver-Vehicle

biased driving Trip Requirements Route, Ambient …

PEMS: CO2, NOx, PN …

41

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – FUNDAMENTAL REQUIREMENT

o Following from previous discussions, ACEA’s position continues to be that Euro6d RDE for LDVs incorporates Trip Normalisation as a fundamental principle – with the aim of ensuring that real world emissions correspond to those measured at Type Approval.

o This principle has been incorporated into Euro6 RDE from the outset, and removal at this late stage would present an insurmountable challenge to the engineering of vehicle programmes currently in development for Euro6d-Temp and -Final.

o The investigation of alternative options for Trip Normalisation should therefore focus on...

o How well they perform against the requirements of each trip in question – should it be normalised up or down against the reference.

o And not just how the results vary against other Trip Normalisation options &/or the raw results.

42

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – OPTIONS UNDER CONSIDERATION

o The lack of Trip Normalisation therefore excludes Evaluation Method options 1, 2 & 7.

o The key requirement for a single Evaluation Method, that correctly normalises each RDE trip, then excludes the options which have been demonstrated to introduce random variations / not work for all vehicle technologies:

o MAW mathematical problems / not PHEV = 4 & 6

o PB emissions characterisation problems / not PHEV = 5

o Leaving the CO2 ratio based Evaluation Method options #3 & #8...

43

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – CO2 RATIO

o Option#3 NL CO2 Ratio...

o CO2 ratio has been demonstrated to provide a fair indication of cycle energy and is therefore a good candidate for RDE Trip Normalisation against the WLTP reference:

RDE Final Result/km = RDE Trip Result/km x WLTP CO2 g/km

RDE CO2 g/km

o Plus, for valid trips, this approach includes all emissions data in the analysis – rather than applying a zero weighting to any points outside of +/- Tol2 as per the current Moving Averaging Windowing approach:

Easier to understand if not mathematically simplified in the way that App7c has been.

44

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – CO2 RATIO FOR PHEV

o However – for a PHEV this normalises the trip based on how hard the engine was working, when it was running, but takes no account of the distance accumulated during EV operation:

o But if one were to decide, that because the CO2 Ratio Trip Normalisation does not work for PHEVs, just to use straight RDE Emissions / km:

o Then the situation becomes biased against PHEV technology – as it receives no RDE Trip Normalisation for the level of cycle energy expended, in either EV or ICE mode.

o With bias in the emissions assessment increasing as move more towards Charge Sustaining mode, and ICE operation becomes dominant.

Resulting in a bias against PHEV technology, which increases with EV distance in the RDE test...

45

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – CO2 RATIO x DISTANCE RATIO FOR PHEV

o Therefore ACEA 07Apr17 suggestion = Option #8...

o Utilising CO2 Ratio x Distance Ratio, in an attempt to combine both aspects into a valid method for PHEV RDE Trip Normalisation in any mode:

o Which we recognise, as presented to the RDE-LDV meeting, will not provide a perfectly accurate analysis – there will be variation in the results according to different PHEV operation strategies, all of which must be considered during RDE testing, vs. the Charge Sustaining WLTP test.

o But is a pragmatic option for Euro6d RDE Package#4, utilising the data which will be available within the regulations in this timeframe.

46

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – BIAS IN CO2 x RDE DISTANCE RATIO

o However, it is also recognised that this evaluation has the potential for an overall +ve bias in favour of PHEVs, as it does not take the EV ratio of the Charge Sustaining WLTP reference test into account, as detailed in ACEA’s 18May17 presentation:

And discussed during our 30May17 meeting / after the 01Jun17 RDE-LDV meeting.

o If this is leading PHEV RDE results evaluations to be viewed as too low vs. the raw trip results, then further consideration is required...

47

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – INCORPORATION OF CS WLTP TEST DISTANCE RATIO

o A more rigorous implementation of this approach would be to multiply RDE emissions / km by the ratio of CO2 g/km WLTP to RDE - when the engine is running - for both the RDE and the WLTP test:

RDE Final Result/km = RDE Trip Result/km x CS WLTP CO2 g/kmICE RDE CO2 g/kmICE

o Which can be shown to be equal to:

RDE Final Result/km = RDE Trip Result/km x CS WLTP CO2 g/km x ICE-RatioRDE

RDE CO2 g/km ICE-RatioWLTP

Where: ICE-Ratio = kmICE / kmTOTAL

(Although this would still be subject to variation in results due to differences in PHEV operation modes between RDE & CS WLTP testing.)

48

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – CENTRALISATION OF PHEV +/- ERROR

o However the distance driven by PHEVs in ICE vs. EV during its Charge Sustaining WLTP Type Approval tests will not be available for use in all RDE results evaluations, including 3rd party.

o Therefore, in order to centralise the unavoidable +/- error in PHEV RDE results evaluations, a typical average value for the ICE:EV operation of PHEVs during a CS WLTP test could be incorporated in the calculation.

o A ratio of 85% ICE : 15% EV is suggested as suitable, to be included in the PHEV RDE results evaluation as a 1/0.85 factor:

RDE Final Result/km(t,u) = RDE Trip Result/km(t,u) x CS WLTP CO2 g/km x ICE-RatioRDE(t,u)

RDE CO2 g/km(t,u) 0.85

o Giving a 17.6% increase in all PHEV RDE results.

49

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – 100% SoC PHEV RDE

o Plus the PHEV high EV mode operation clause - as per 07Apr17 proposal:

Where we can now suggest 0.2 as a suitable value for x.

o Which gives the same outcome for a 100% SoC PHEV RDE test as COM’s 07Apr17 proposal...

(Although – as we discussed on 30May17 – our concerns w.r.t. the suitability of this result as an additional value on the CoC for PHEV vehicles remain.)

50

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION – SINGLE FUNDAMENTAL APPROACH

o Thereby providing a single fundamental approach to Euro6d RDE Results Evaluation.

o By utilising the most accurate implementation of Trip Normalisation by CO2, across all vehicle types, that will be possible with the available data.

o For both Total Trip and Urban analysis, with an additional term to be applied for PHEV vehicles :

o And alleviating concerns about excessive reduction in results.

RDE Test Results

Results Evaluation

by All Emissions /

All kms

Y

N

Results Evaluation

by CO2 Ratio

m(t,u),CO2 / mwltc,CO2

< 0.2

Results Evaluation

by CO2 Ratio x

Distance Ratio / 0.85

N

Y

Data ready for results

calculation... OVC-HEV

51

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP TRIP NORMALISATION - FURTHER OPTIONS FOR PHEVS

o All further concepts that have been explored, with the aim of finding a more thorough and accurate RDE Evaluation Method for PHEVs, lead inexorably to the conclusion that more data would be required than is going to available within the Euro6d RDE Package #4 timeframe:

o Direct calculation of cycle energy = fn(speed, accel, slope, vehicle mass)?

o But would require better instrumentation / data on acceleration and instantaneous road gradient data than is available from calculations based on GPS data?

o Or – really – require PowerICE,EV & Battery SoC data, logged from the vehicle (as per HD RDE engine power) to be utilised in a more rigorous evaluation approach?

o Which could be considered for further development of RDE in the future, as per “...keep under review ...the RDE test procedure and adapt ...to accommodate new vehicle and/or measurement technologies and to ensure their effectiveness.”

52

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP

Euro6d RDE – Package#4

Single RDE Evaluation Method

Annex IIIA

Appendix 7a Appendix 7b

One Validity Check One Results Calculation

53

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP SINGLE RDE EVALUATION METHOD – OVERALL PROCESS FLOW

o Euro6d RDE can thus achieve the goal of a fundamentally single overall Results Evaluation Method:

RDE Test Results

Trip Requirements

as per AnnexIIIa

Kinematic Boundary

Conditions as per Appendix

7a & 7b

Trip Validity check...

CO2 based method

Results Evaluation... CO2 Ratio

CO2 x Distance Ratio/0.85 for PHEV (Emissions/km for high SoC)

Test Not Valid

Test Not Valid

Test Not Valid

Y Y Y

N N N

RDE test > PEMS data processing > Exchange

File

54

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP

Thank you for your consideration of these suggestions.

55

EURO6d-RDE EVALUATION METHODS > Modified MAW > MAW+PB

Back-Up ACEA Round Robin

56

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #01: currently passed passed by Normality & RSI

Assessing trip#01 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

57

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Assessing trip#02 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

failed

Trip #02: currently passed (due to applied “NEDC” scaling factors [1.2/1.1/1.05] for WLTP)

failed by Normality , ok by RSI

58

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #03: currently passed passed by Normality & RSI

Assessing trip#03 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

59

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #04: currently unnecessarily excluded passed by Normality & RSI

Assessing trip#04 from ACEA calculation round robin

Current situation:

Completeness ≥ 15% Normality ≥ 50%

failed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

60

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #05: currently passed passed by Normality & RSI

Assessing trip#05 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

61

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #06: currently passed passed by Normality & RSI

Assessing trip#06 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

62

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #07: currently passed passed by Normality & RSI

Assessing trip#07 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

63

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #08: currently passed failed by Normality & RSI

Assessing trip#08 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

failed

64

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #09: currently passed passed by Normality & RSI

Assessing trip#09 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

65

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #10: currently passed passed by Normality & RSI

Assessing trip#10 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

66

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #11: currently passed passed by Normality & RSI

Assessing trip#11 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

67

ACEA TF-RDE - EVALUATION METHODS / HYBRIDS SUB-GROUP Validity check – how to exclude abnormal driving

Trip #12: currently passed passed by Normality & RSI

Assessing Urban-trip#12 from ACEA calculation round robin

Current situation: Completeness ≥ 15% Normality ≥ 50%

passed

ACEA proposal: Normality ≥ 50% RSI ≤ 20%

passed

(Urban assessment only)

68

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#01 from ACEA calculation round robin

total urban

Trip #01: currently passed passed by Combined MAW & PB

passed

69

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#02 from ACEA calculation round robin

total urban

Trip #02: currently passed failed by Combined MAW & PB

failed

70

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#03 from ACEA calculation round robin

total urban

Trip #03: currently passed passed by Combined MAW & PB

passed

71

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#04 from ACEA calculation round robin

total urban

Trip #04: currently unnecessarily excluded passed by Combined MAW & PB

passed

72

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#05 from ACEA calculation round robin

total urban

Trip #05: currently passed passed by Combined MAW & PB

passed

73

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#06 from ACEA calculation round robin

total urban

Trip #06: currently passed passed by Combined MAW & PB

passed

74

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#07 from ACEA calculation round robin

total urban

Trip #07: currently passed passed by Combined MAW & PB

passed

75

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#08 from ACEA calculation round robin

total urban

Trip #08: currently passed failed by Combined MAW & PB

failed

76

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#09 from ACEA calculation round robin

total urban

Trip #09: currently passed passed by Combined MAW & PB

passed

77

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#10 from ACEA calculation round robin

total urban

Trip #10: currently passed passed by Combined MAW & PB

passed

78

EURO6D-RDE EVALUATION METHODS VALIDITY CHECK – HOW TO EXCLUDE ABNORMAL DRIVING

Assessing trip#11 from ACEA calculation round robin

total urban

Trip #11: currently passed passed by Combined MAW & PB

passed