歐7標準下制動排放實驗室測試的統計分析.pdf

1、Statistical Evaluation of Euro 7 Laboratory Tests for Brake Emissions FactorsCarlos Agudelo03.14.2024 CARB Riverside,CAOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-152Variability components ILS2Dyno test capabilitiesWeighing process capabilitiesOSAR 2024|C EOSAR 2024|C E-C E R TC E

2、 R TRiverside,CA.March 14-153Sections Variability components ILS2ISO 5725(26)repeatability,sample effect,lab effect,total Reproducibilitysr,sH,sL,sRCDr,CDROSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-154AcknowledgmentResearch membersTo AUDI AG,BMW AG,Ford-Werke GmbH,Stellantis N.V.

3、,Continental AG,DRiV Inc.,Brembo S.p.A,and Volkswagen AGFor providing brake componentsTo M.Arndt(AVL List GmbH),K.Kolbeck and J.Von-Wild(BMW AG),M.Federici(Brembo S.p.A.),M.Morbach and C.Klsch(DRiVInc.),J.Grochowicz(Ford-Werke GmbH),G.Kanae Filler and D.Lugovyy(HORIBA Europe GmbH),A.Perez and J.Oliv

4、e(Applus+IDIADA),A.Sin and S.Balestra(ITT Inc.),H.Hagino(Japan Automobile Research Institute-JARI),R.Vedula and A.Hortet(LINK),H.Chong(NIER,South Korea),P.Nyhof(TMD Friction Services GmbH),H.Niemann and H.Kaminski(TU DARMSTADT),D.Hesse and C.Hamatschek(TU ILMENAU),and P.Jouy(UTAC France)For their te

5、chnical support during the ILS2TF3 membersCarlos Agudelo,Dr.Barouch Giechaskiel,Dr.Theodoros Grigoratos,Dr.Ing.Jarek Grochowicz,and Dr.Ing.Hartmut NiemannOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-155WHY doing thisbackground and benefitsEuro 7Estimates of relative contributions t

6、o variabilityDetect opportunities to improve UN GTR 24Assess the ability to meet UN GTR 24 requirementsOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-156HOW we calculated the standard deviationsISO 5725-5 for heterogenous materials using non-robust algorithmsT1123456T1123456T1123456T

7、2123456T2123456T2123456m=general averagezijtk=test effectLab 1Lab 8Lab 2Hijt=sample effectBij=lab effectpj=8 labsgj=16 samplesnj=96 resultsnijt=6LjHjrjnijKijKjK”jKjSSHjSSLjSSrjzijtk=test effectHijt=sample effects2Hjs2rjs2Ljs2RjBij=lab effectyijtkmj=general averageHigh-level metrics for entire testEv

8、ent-based for measurand results Using the same brake as ILS1ISO 5725-5 for sr,sH,sL,sRStatistical reportOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-157WHAT we measuredbasis to establish critical differences(CDrfor test-to-test and CDRfor dyno-to-dyno)per ISO 5725-1,-6between tests

9、between labs/dynossrrepeatability(tests)sHSample(friction couple)sLLaboratory/DynamometersRReproducibility2=2+2ISO 5725-5OSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-158STATISTICS for emissions factorslab is main component for PM10variability,while repeatability is main component f

10、or PM2.5Overall mean and Reproducibility(error bar)Percent contribution to total variance by component0246810PM2.5 EF mg/kmPM10 EF mg/kmGenAvg.-mj0%50%100%PM2.5 EF mg/kmPM10 EF mg/kmStdDev repeat.-srStdDev Sample-sHStdDev Lab-sL 35%38%OSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-15

11、9STATISTICS for overall brake and cooling air temperatureslab is main contributing factor to variability050100150200250Avg Cooling Air Temp.CAvg Disc Temp.CMax.Disc Temp.CGen.Avg.mj0%50%100%Avg Cooling Air Temp.CAvg Disc Temp.CMax.Disc Temp.CStdDev repeat.-srStdDev Sample-sHStdDev Lab-sL 8%10%4%Over

12、all mean and Reproducibility(error bar)Percent contribution to total variance by componentOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1510Components of variability for avg-by-time torque by eventlarger scatter for all factors during ILS2ILS1ILS202468100102030405060Trip#Standard de

13、viation NmBrake event/#sLjsHjsrjTrip024681001020304050Trip#Standard deviation Nm Brake event/#sHjsLjsrjTripOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1511Sections Variability componentsDyno test capabilitiesISO 5725(26)ISO 22514(27)GTR 24repeatability,sample effect,lab effect,tot

14、al ReproducibilityBias,repeatability,linearitysr,sH,sL,sRCDr,CDRUMS,QMP,PpkOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1512AcknowledgmentResearch membersBarry Purtymun,from LINK-Arizona,for processing the dyno data using the UN GTR 24 test report outputTyler Odom,from LINK Technic

15、al Center-Dearborn,Michigan,for making the UN GTR 24 test report workThose working behind the scenesCarlos Agudelo,Dr.Ing.Hartmut Niemann,Maximilian Hense,Alejandro Hortet,James Boatwright,Dr.Frederik WeisOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1513HOW to assess the different

16、UN GTR 24 requirementsusing three categories of dataContinuous(ISO 22514-7)Time-based(ISO 22514-2)Attributes(ISO 22514-5)Measurands that vary during the test:Braking/release speedDeceleration/torqueEvent durationConstant measurands during the test:Cooling airflowCooling air temp.and%RHPM/PN sampling

17、 flowMeasurands expressed as Y/N or%:Speed violationsKinetic energy dissipationTest level background complianceOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1514HOW to assess the different UN GTR 24 requirementsusing three categories of dataContinuous(ISO 22514-7)Time-based(ISO 2251

18、4-2)Measurands that vary during the test:Braking/release speedDeceleration/torqueEvent durationConstant measurands during the test:Cooling airflowCooling air temp.and%RHPM/PN sampling flowOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1515HOW we compiled the datause of dyno data from

19、 two standard configurations+brake emissions upgrade(A)M3000 performance(B)M3900 NVHMotor:186 kWFS torque:5600 NmMotor:186 kWFS torque:5600 NmMaximum simulated inertia:260 kgmMaximum simulated inertia:320 kgmMaximum speed:300 km/h rR:0.4 mMaximum speed:375 km/h rR:0.4 mM6330+M4222 per UN GTR 24M6330

20、+M4222 per UN GTR 24Test loads:It=80 kgm;rR=0.340 mTest loads:It=50 kgm;rR=0.290 mBrake size:reff=120 mm;dpiston=60 mm 1Brake size:reff=108 mm;dpiston=60 mm 1OSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1516HOW we selected the levels for the one-way ANOVA(ISO 22514-7)minimum+20thpe

21、rcentiles+maximum level(3-parameter lognormal distribution)Histograms of Set ValuesCumulative Density FunctionAvg-by-time TorqueAvg-by-time Rotational SpeedEvent DurationBraking SpeedOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1517HOW do the statistics work(ISO 22514-7+VDA Guide 5

22、)for brake work using Maximum Permissible Errors to estimate the allowable relative uncertainties combining torque and revolutions to stop(uncorrelated components per GUM)expressed as relative uncertainties(%)3=23+23=2+2MPEKE 5%per UN GTR 24 3=.23+.23.=2+2OSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiv

23、erside,CA.March 14-1518WHAT we calculated for average-by-time torqueMetrics for uncertaintyDyno(A)Dyno(B)BIAS NmuBI0.428 60.382 8LINEARITY NmuLIN0.261 40.751 4REPEATABILITY NmuEVR0.100 20.151 1Combined uncertainty NmuMS0.512 00.856 7Combined exp.uncertainty NmUMS1.024 01.807 6COMBINED EXPANDED UNCER

24、TAINTY0.41%0.96%Performance ratio(1.33Cooling air relative humidity(50%RH)UN GTR 24 bias=0.11%RH Ppk=1.64 1.33Cooling air temperature(20 C)UN GTR 24 bias=1.1 CPpk=4.04 1.3315 C25 C21.1 C0.97C19.90 C22.07 C1.20 C20%80%49.85%18.4%35.7%68.3%14.2%180 m/h 220 m/h199.1 m/h 7.6 m/h194.7 m/h206.7 m/h4.4 m/h

25、OSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1524Sections Variability componentsDyno test capabilitiesWeighing process capabilitiesISO 5725(26)ISO 22514(27)GTR 24ISO 22514(27)GTR 24 DIN EN 12341repeatability,sample effect,lab effect,total ReproducibilityBias,repeatability,linearity

26、Filter(integrity,charge,stability)Room(repeatability,conditions)Sampler(bias,repeatability,linearity,correlation)sr,sH,sL,sRCDr,CDRUMS,QMP,PpkPpkOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1525GTR 24 Weighing room and weighing process requirements(extract)Including microbalance,ro

27、om,filter conditioning,and variabilityRoom Microbalance Process Free of contaminants(22 2)C&(45 8)%RHFilter storage(pre&post)Buoyancy correctionIsolation from vibrationCalibration reportCalibrated weightsResolution 1 g 10 greference filters|m1-m2|10 g or range|m1m4|15 g OSAR 2024|C EOSAR 2024|C E-C

28、E R TC E R TRiverside,CA.March 14-1526DIN EN 12341 C.2.Filter integritycondition&weigh holder wait 1 hour remove weighDIN EN 12341 C.2.Static chargecondition&weigh(without discharge)static discharge reweighOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1527DIN EN 12341 D.1.Accuracy a

29、nd stability of room conditions20 C RT 24 C37%RT 53%|m1 m2|700 tests routine tests OSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1528Experimental setup for level and detection and Two-Way ANOVA Gtr 24-modified to accommodate measurements close to DIN EN 123411443254321Non-rotating b

30、rake20 balanced flow splitters0.950.98 isokinetic ratio 33.3 lpmManual and automatic filter changers5PALAS RBG Basic dust generatorOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1529Experimental setup for level and detection and two-way ANOVA PALAS adjustable generator to inject ISO

31、12103-1 Arizona fine dustAZ fine dustAZ fine dust size distributionAZ fine dust cylindersRBG working principleRBG industrial embodimentInjection at brake enclosure5OSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1530ISO 22514-7 between samplers ANOVATwo sampling systems(manual and aut

32、o)and five filter load levelsBias 10%for the lower limit per EN 12341Linearity within 4%correlation factor 99%OSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1531in closingLaboratory effect dominates total test Reproducibility(from ILS1 and ILS2)With due diligence,a test facility can

33、meet the UN GTR 24 requirementsAutomatic sequential samplers correlate within 5%with manual filter holdersOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1532Carlos A+1 313 319 Director Applications EngineeringLINK GroupPlymouth,Michigan USOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRive

34、rside,CA.March 14-1533STATISTICS for brake torque and coefficient of frictionboth dominated by lab effect050100150200Avg Time Torque NmGen.Avg.mj0%20%40%60%80%100%Avg Time Torque NmStdDev repeat.-srStdDev Sample-sHStdDev Lab-sL00.20.40.6Avg Distance Effectiveness Gen.Avg.mj0%50%100%Avg DistanceEffec

35、tiveness StdDev repeat.-srStdDev Sample-sHStdDev Lab-sL 10%16%Overall mean and Reproducibility(error bar)Percent contribution to total variance by componentOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1534Summary on test variability for brake work and brake temperature,lab variabil

36、ity is key ILS3 needs to(and it is)including test repeats to ensure proper statisticsAverage-by-Time TorquesrjsHjsLjsRjsrjsHjsLjsRj3031.4%0.1%6.1%6.3%5.7%0.0%12.9%14.3%50th1.3%0.0%5.4%5.5%4.7%0.0%12.3%13.5%95th2.0%0.5%11.2%11.3%14.2%0.0%20.6%25.0%EventsILS1ILS2Standard deviations(repeat.,Sample,Lab,

37、Reproducibility)as percent of the general averageOSAR 2024|C EOSAR 2024|C E-C E R TC E R TRiverside,CA.March 14-1535WHAT was the performance for PM10and TPN sampling airflowsusing test-specific values and ISO 22514-2 indexesTPN sampling(0.810.99 lpm)UN GTR 24 bias=0.001%Ppk=7.3 1.33PM10sampling(16.3317.00 lpm)UN GTR 24 bias=0.02%Ppk=1.66 1.3316.03 lpm17.31 lpm16.64 lpm0.40 lpm0.61 lpm16.64 lpm17.04 lpm0.21 lpm0.8 lpm1.0 lpm0.899 lpm0.014 lpm0.0218 lpm0.891 lpm0.913 lpm0.078 lpmSet value16.67 lpmSet value0.90 lpm