An Experimental Methodology for Measuring of

Downloaded from SAE International by Georgios Fontaras, Monday, March 17, 2014 10:00:48 AM 2014-01-0595 Published 04/01/2014 Copyright © 2014 SAE International doi:10.4271/2014-01-0595 saecomveh.saejournals.org

An Experimental Methodology for Measuring of Aerodynamic Resistances of Heavy Duty Vehicles in the Framework of European CO2 Emissions Monitoring Scheme Georgios Fontaras and Panagiota Dilara European Commission

Michael Berner Daimler AG

Theo Volkers Daf Trucks N.V.

Antonius Kies, Martin Rexeis, and Stefan Hausberger University of Technology Graz ABSTRACT Due to the diversity of Heavy Duty Vehicles (HDV), the European CO2 and fuel consumption monitoring methodology for HDVs will be based on a combination of component testing and vehicle simulation. In this context, one of the key input SDUDPHWHUVWKDWQHHGWREHDFFXUDWHO\GH¿QHGIRUDFKLHYLQJDUHSUHVHQWDWLYHDQGDFFXUDWHIXHOFRQVXPSWLRQVLPXODWLRQLV the vehicle's aerodynamic drag. A highly repeatable, accurate and sensitive measurement methodology was needed, in order to capture small differences in the aerodynamic characteristics of different vehicle bodies. A measurement methodology is proposed which is based on constant speed measurements on a test track, the use of torque measurement systems and wind speed measurement. In order to support the development and evaluation of the proposed approach, a series of experiments were conducted on 2 different trucks, a Daimler 40 ton truck with a semi-trailer and a DAF 18 ton rigid truck. Two different torque measurement systems (wheel rim torque sensors and half shaft torque sensors) were used for the measurements and two different vehicle tracking approaches were investigated (high precision GPS and opto-electronic barriers). Results were pooled and compared against results from similar measurements performed by the OEMs at their own proving grounds. The method was proven to be accurate. The analysis showed good UHSHDWDELOLW\DQGUHSURGXFLELOLW\FKDUDFWHULVWLFVDQGDJRRGVHQVLWLYLW\RIWKHPHWKRG%DVHGRQWKH¿QGLQJVLWZDVGHFLGHG that this measurement methodology is suitable and can be included in the European legislation.

CITATION: Fontaras, G., Dilara, P., Berner, M., Volkers, T. et al., "An Experimental Methodology for Measuring of Aerodynamic Resistances of Heavy Duty Vehicles in the Framework of European CO2 Emissions Monitoring Scheme," SAE Int. J. Commer. Veh. 7(1):2014, doi:10.4271/2014-01-0595.

INTRODUCTION The European Commission (EC) in joint collaboration with Heavy Duty Vehicle manufactures (ACEA) and other research bodies has been preparing a new legislative framework for monitoring CO2 emissions from Heavy Duty Vehicles (HDV) in Europe. The core of the upcoming CO2 monitoring methodology is based on vehicle simulation. This approach offers the possibility to accurately capture the highly diverse FKDUDFWHULVWLFVRI+'9VDQGWKHLULQÀXHQFHRQIXHO consumption and CO2 emissions, without heavily increasing WKHFRPSOH[LW\DQGWKHFRVWVIRUYHKLFOHFHUWL¿FDWLRQ$VHULHVRI studies have been performed to demonstrate the advantages

of a simulation-based approach and lay down the lines for a IXWXUHFHUWL¿FDWLRQVFKHPH>1]. Similar simulation-based approaches have been so far adopted by other major HDV markets, e.g. GEM in the USA or the Fuel Economy Standard 2015 of Japan. One of the key parameters affecting fuel consumption and CO2 emissions of HDVs, particularly under highway and extra urban driving conditions, is their aerodynamic performance. $HURG\QDPLFGUDJLVLGHQWL¿HGDVWKHPRVWLQÀXHQWLDO parameter when it comes to vehicle energy consumption for long haul delivery missions and other relevant high speed HDV

Downloaded from SAE International by Georgios Fontaras, Monday, March 17, 2014 10:00:48 AM

Fontaras et al / SAE Int. J. Commer. Veh. / Volume 7, Issue 1 (May 2014) operations (e.g. intercity coaches). Long haul delivery and intercity passenger transport account together for ca. 60 % of CO2HPLVVLRQVJHQHUDWHGE\+'9VLQ(XURSH>2]. Thus it is of key importance for a CO2 monitoring methodology to be able to accurately capture the aerodynamic performance of HDVs, both on an comparative level, e.g. the difference in aerodynamic performance between 2 vehicles, and on an absolute level, e.g. the exact resistance introduced by the vehicle's aerodynamics. A new methodology was necessary for measuring the vehicle aerodynamic drag and also the necessary input variables to be used in a vehicle simulation model. $VWXG\>3] funded by the EC investigated possible options for measuring the aerodynamic resistances of an HDV, mainly focusing on two types of tests: coast-down and constant speed PHDVXUHPHQWV7KHVWXG\FRPSDUHGWKHEHQH¿WVRIDFRDVW down based approach with respect to a steady state speed test one. The analysis considered elements of existing standards for vehicle road load calculation such the SAE standards J2881 >4@->5], the EPA standard EPA-HQ-OAR-2010-0162, the only public standard at the time for HDV road load calculation as well as the provisions of the ISO standard 10521-1, (chapter 5.5), which fully includes the procedures of the previously mentioned SAE standards. Result revealed that although these tests might be suitable for the road load determination of light duty vehicles, a series of issues would not allow for the adoption of one of these standards. Some of which were: • 6LPSOL¿FDWLRQVLQWKHWHVWSURFHGXUHVDQGDVVXPSWLRQV which may result in inaccuracies • The coast-down drivetrain losses (e. g. idle differential friction), different to the normal powered case, which are not taken into account • The wind velocity is only limited and in certain cases not considered for the data analysis in any way. • Only constant road gradients are considered, but even on most appropriate test tracks slopes exist • $FRUUHFWLRQRIWLUHLQÀDWLRQSUHVVXUHWRWHPSHUDWXUH differences between garage and test track environment, suitable only for LDV tires at 3 bar. The conclusions of this analysis resulted in a proposal to adopt a constant speed based approach for the experimental GHWHUPLQDWLRQRI+'9DHURG\QDPLFUHVLVWDQFHV$¿UVWGUDIW proposal on a possible constant speed based test was provided by ACEA, based on internal testing at various OEM sites. This proposed methodology was then further VXSSOHPHQWHGEDVHGRQWKH¿QGLQJVRIWKH/27 FDPSDLJQ and evaluated both by OEMs and independently by the Joint Research Centre (JRC) of the European Commission. This paper presents the results of the evaluation campaign together with a brief overview of the testing methodology.

PROPOSED METHODOLOGY General The general model (eq 1) that describes the aerodynamic resistances applied on a vehicle is considered by the vehicle simulator to be used for CO2 monitoring.

(1)

where: Fair DLUGUDJ>1@ Cd DLUGUDJFRHI¿FLHQW>@ ȕ DYHUDJHDLUÀRZDQJOH\DZDQJOH >@ Acr FURVVVHFWLRQDODUHDRIWKHYHKLFOH>P2] ȡair,ref DLUGHQVLW\DWUHIHUHQFHFRQGLWLRQV>NJP3] vvair ZLQGYHORFLW\>PV@ Scope of the constant speed testing methodology is to GHWHUPLQHWKHDHURG\QDPLFGUDJFRHI¿FLHQW&d) as a function RIWKH\DZDQJOHȕ ZLWKGLUHFWWRUTXHPHDVXUHPHQW7R achieve this, the wheel torque of the driven wheels, the vehicle YHORFLW\WKHDFWXDODLUÀRZYHORFLW\YHKLFOHYHORFLW\SOXVZLQG DQGWKHDLUÀRZGLUHFWLRQDUHPHDVXUHGV\QFKURQRXVO\RYHU straight motion on a test track. Measurements are performed at two different constant vehicle speeds (Vlow and Vhigh) under GH¿QHGFRQGLWLRQV7KH9low of the testing is a constant velocity between 10 - 15 km/h while the target Vhigh should be between of 85 - 90 km/h. In case a vehicle cannot achieve the foreseen high speed, the maximum achievable vehicle speed is applied. Given the abovementioned measured data and information UHJDUGLQJWKHVORSHSUR¿OHRIWKHWHVWVHFWLRQVLWLVSRVVLEOHWR calculate the road load of the vehicle (see Figure 1) based on WKHIROORZLQJTXDOL¿HGDVVXPSWLRQV • rolling resistance force (Frol) independent of vehicle speed1 • air drag force being quadratic to the velocity 7KHDHURG\QDPLFGUDJLVFDOFXODWHGIRU\DZDQJOHVȕ EHWZHHQ DQGEDVHGRQWKHDFWXDOPHDVXUHGDLUÀRZYHORFLW\ In addition to the key elements of the test demonstrated in Figure 1, important parameters of the procedure include the use of high precision positioning instrumentation for accurate recording of vehicle position and ground speed (opto-electronic barriers or very high precision GPS system), weather information and data retrieved from vehicle sensors. Detailed VSHFL¿FDWLRQVDUHIRUHVHHQIRUHDFKLQVWUXPHQWDQGIRUWKH sampled signals.

1. Rolling resistance is influenced by vehicle speed. However the analysis presented in the paper shows that this influence is limited under the defined test conditions and does not affect significantly the calculated air drag value. Studies are ongoing in order to find a way to include the effect of speed on rolling resistance value into the AD calculation method.


Fontaras et al / SAE Int. J. Commer. Veh. / Volume 7, Issue 1 (May 2014)

Preparation Phase During this phase, the various measurement instruments are mounted on the vehicle and their good operation is checked (e.g. the drive-wheels should be checked for proper rotation with installed torque meters or half shafts). The vehicle's mass is measured or calculated and the tyres are checked for the PD[LPXPDOORZDEOHLQÀDWLRQSUHVVXUH7KHSURSHUYHKLFOHDQG WUDLOHULISUHVHQW KHLJKWVDUHYHUL¿HG The positioning equipment is also set up during this phase, HLWKHURQO\RQWKHWHVWWUDFN¿QDOGH¿QLWLRQRIWHVWVHFWLRQYLD installation of the opto-electronic barriers) or on the vehicle and on the test track via high precision differential GPS. 'DWDUHJLVWUDWLRQRIDOOUHOHYDQWPHDVXUHPHQWVLJQDOVLVYHUL¿HG and the engine is left at idling for preconditioning without parking brake.

Warm-Up Phase

Figure 1. Key points of the proposed testing methodology

Test Track In terms of the testing ground different types of test track geometries are foreseen (Figure 2). The important factor in this case is the execution of measurements in both directions in order to cancel out to the best possible extent the effects of ambient wind. The test track must have straight section(s) were the measurements are performed. An extra straight length before each measurement section is foreseen in order to allow IRUWKHVWDELOL]DWLRQRIZLQGÀRZDURXQGWKHYHKLFOHRUWKH drivetrain torque after cornering.

During this phase the vehicle is driven for 90 minutes at Vhigh to assure that the tyres reach a constant pressure and temperature level, and that the powertrain and drivetrain reach a constant coolant and lubricant temperature level. During this phase, the vehicle is driven in both track directions in order to achieve a balanced warm up of the tires and collect data for subsequent control checks, and for determining the misalignment and position error of the mobile anemometer. At the end of the warm up phase, the vehicle is brought to a standstill on a selected area of the test track. The vehicle is slowed down carefully without braking and rolled out for the last meters, with free clutch / neutral gear and engine switched off. Once still and in zero torque conditions, the torque sensors mounted on the vehicle are checked for drift and are subsequently zeroed.

High Speed Test Phase After the zeroing of the torque sensors the vehicle is driven for a minimum of 2 km at Vhigh in order to reach stabilization again. Subsequently the Vhigh test is performed. During testing it must be ensured that: • WKHGULYLQJVSHHGLVFRQVWDQWDWOHDVWIRUWKHGH¿QHG measurement sections and the preceding stabilization sections • the vehicle is driven through the measurement section along a straight line without steering • the amount of recorded measurement sections leads to enough valid “evaluation sections” in the data processing Figure 2. Testing grounds and segments

Test Sequence The test is comprised of 5 phases, the preparation phase, the warm up phase, the high speed test phase, the low speed test phase and control phase as described below.

• a minimum of 20 measurement sections for each driving direction are performed with an equal amount of sections driven in both directions; e.g.: either enough rounds on an one-way circuit track with two measurement sections or an equal amount of measurements driven in each direction on a circuit or straight line track with one measurement section;



Low Speed Test

Table 1. Main vehicle characteristics and main input data origin

The test at Vlow is performed directly after the high speed test. As in the case of the Vhigh test it must be ensured that the driving speed is constant at least for the measurement sections and the preceding stabilization sections and that the vehicle is driven through the measurement section along a straight line without steering. In both Vhigh and Vlow tests the beginning and end of the measurement sections should be clearly recognizable in the measurement data, either via a recorded trigger signal (opto-electronic barriers) or via recorded GPS data.

Control Phase During this phase a series of quality control checks are foreseen, mainly to establish that the drift of the torque sensors remained within acceptable values. The drift check of torque meters, performed in this phase, depends on the type of the torque measurement instrument and it may involve a free roll out of the vehicle or lifting of the driven axle off the ground. In addition to drift, some general checks such as controlling axle/wheel bearings for overheating and a general re-check of WKHYHKLFOHFRQ¿JXUDWLRQDUHDOVRIRUHVHHQ

Figure 3. Vehicles used in the study a: Actros tractor-trailer, b: CF75 rigid truck

Test Field Post Processing A series of post processing corrections on measured data are foreseen. The recorded average vehicle speed is corrected based on the information retrieved from the optical barriers or WKHKLJKSUHFLVLRQ*367KHDLUÀRZYHORFLW\VLJQDORIWKH mobile anemometer is corrected in three steps for the LQVWUXPHQW VHUURUDVGH¿QHGE\WKHFDOLEUDWLRQUHSRUWWKHHUURU generated by the positioning of the instrument on the vehicle PHDVXULQJSRVLWLRQLQVLGHDQDFFHOHUDWHGÀRZGXHWRWKH VKDSHRIWKHYHKLFOH DQGWKHDLUÀRZERXQGDU\OD\HUHIIHFWDQG the yaw angle misalignment (see appendix). Additional corrections are foreseen for compensating the torque sensor drift during the test. In the latter case a linear distribution of drift over time is performed.

The constant speed tests were performed on the Iveco proving ground at Balocco, Italy. Both trucks were driven on the peripheral (outer) track with two different constant speeds, at 15 km/h and 89 km/h, according to the measurement protocol. Measurements were performed on the effective segments (long straights) as indicated in Figure 4. The measurement is performed while the vehicle is cruising at constant speed (2 speeds were recorded, 15 km/h and 89 km/h).

EXPERIMENTAL This paragraph presents the details of the experimental setup, used during the validation campaign of the proposed methodology.

Test Vehicles Two vehicles were used in the study, a 40 ton Euro VI tractortrailer Daimler Actros and a 18 ton DAF CF75 Euro V (see Table 1). Both vehicles were provided by the OEMs in standard RSHUDWLQJFRQ¿JXUDWLRQ

Figure 4. Balocco proving ground overview.

Test Protocol and Equipment The measurements with both vehicles followed the general guidelines and procedural restrictions of the proposed methodology, described previously in the paper. Both vehicles were tested at Vhigh of 89 km/h and Vlow of 15 km/h on the same measurement sections (see Figure 4). The same type of


Fontaras et al / SAE Int. J. Commer. Veh. / Volume 7, Issue 1 (May 2014) mobile anemometer (Table 2) was used for measuring the air ÀRZYHORFLW\DQGDQJOHDQGWKHVDPHJHQHUDOWHVWSURWRFRO regarding the various phases was applied. The methodology permits multiple measurement instruments, for CF75 and Actros were used:

the exception of the boundary layer correction (step 3), which was not foreseen at the moment. A linear torque drift correction was applied in the case of the Actros. For the CF75 no drift correction was performed, but rather a pass fail limit was applied when the torque drift over one test exceeded 10 Nm.

• Application of different torque measurement systems (see Table 3)

7KHFDOFXODWLRQVRIWKHDLUGUDJFRHI¿FLHQWVZHUHSHUIRUPHG ZLWKWKH9(&72&6(WRROZKLFKLVVSHFL¿FDOO\GHYHORSHGE\ the Graz University of Technology (TUG) on behalf of the JRC WRDGGUHVVWKHQHHGVRIDIXWXUHFHUWL¿FDWLRQWHVW7KHWRRO performs all necessary data treatment and quality control FKHFNV>6].

• Vehicle positioning measurements where conducted with high precision differential GPS in the case of Actros and opto-electronic barriers in the case of CF75 • ,QWKHFDVHRI$FWURVGLIIHUHQWYHKLFOHFRQ¿JXUDWLRQV affecting air drag were tested. Table 2. Anemometer characteristic

Two sets of results were calculated and analysed, with and without yaw angle correction. For the results corrected for yaw angle, the correction was based on an average correction curve that 6 OEM's (Daimler, DAF, IVECO, MAN, Scania and VOLVO) have prepared in the framework of this activity. The development of more detailed curves for all types of truck bodies and trailers is still on going.

RESULTS AND DISCUSSION CF75

Table 3. Torque sensor characteristic

Figure 5a presents the results of the constant speed tests performed with the CF75 vehicle (the results are normalized by the average value measured by the OEM). Two sets of results were calculated from the measurements performed with the CF75, without corrections applied for yaw angle (vair,mob) and with corrections applied (vair,mob+yaw). As presented the vair,mob+yaw approach gave the best results in both terms of accuracy and variability (see also Table 4). Only a marginal 1.3 % difference from the value reported by the OEM was calculated fact, which indicates good reproducibility characteristics for the method. The variability of the results was also limited, presenting a standard deviation of 1.6 % fact, which suggests good precision of the method in addition to good accuracy characteristics. A key assumption of the method is that the rolling resistance retains a constant value between the tests. b summarizes the UHVXOWVRIWKHFDOFXODWHGWRWDOUROOLQJUHVLVWDQFHFRHI¿FLHQW7KH calculation of the RRC showed that indeed the calculated value presents limited variations and reproducible results, compared to the OEM measurements conducted on a different test track. The type approval RRC value for the particular set of tyres is also provided for comparison. It should be noted though, that the latter is measured on a drum rig under very strict operating conditions, so the comparison is rather indicatory. The RRC was found only 2.2 % different compared to previous OEM and results with a standard deviation of 10 %. The results are in OLQHZLWKVLPLODUUHVXOWVUHSRUWHGE\RWKHUVWXGLHV>7].

Data Post Processing and Calculations Regarding the post processing of measured data, corrections were applied on vehicle speed (CF75) and wind speed as foreseen by the proposed methodology (see appendix), with


Fontaras et al / SAE Int. J. Commer. Veh. / Volume 7, Issue 1 (May 2014) Figure 6 compares the results of the measurements performed with those previously performed by the OEM. The normalized WRDYHUDJH2(0PHDVXUHPHQWDW\DZ DLUGUDJFRHI¿FLHQW is plotted as a function of the average wind yaw angle. It is observed that the JRC results are totally in line with the values recorded by the OEM during previous measurements with the same vehicle but on a different test site and tyres. This is also an important indication suggesting a good reproducibility of the air resistance measurement method.

Figure 5. Aerodynamic resistance (a) and tyre rolling resistance (b) measurement results for CF75, normalized by the average value measured by the OEM

Differences in RRC were expected, since the rolling resistance performance of tyres is highly affected by many factors, such as weather conditions, tarmac, ground temperature and PLOHDJH$IWHUDQDO\VLQJWKHUHVXOWVQRVLJQL¿FDQW dependencies between RRC and Cd · Acr were found (e.g. Cd · Acr being consistently lower when RRC was higher, see Table 5). Therefore, the results appear to verify the initial assumption of the method of limited RRC differentiations between tests. Table 4. Summary of the differences of the measured values with respect to OEM or TA ones for aerodynamic resistance and rolling resistance

Figure 6. Normalized air drag vs yaw angle results for the CF75. Dashed line corresponds to the curve used for correcting results based RQDYHUDJH\DZDQJOHȕ LQWKH9DLUPRE\DZFDVH

In Figure 6 the generic curve used for correction of the average yaw value in the Vaimob+yaw case is also presented. With the exception of the measurements recorded by the OEM at vey low yaw, the curve appears to accurately capture the impact of ZLQGDQJOHȕRQYHKLFOH VDHURG\QDPLFUHVLVWDQFHV7KLVLVD ¿UVWLQGLFDWLRQWKDWWKHSURSRVHGFRUUHFWLRQIDFWRUDSSURDFK foreseen by the proposed methodology is in the right direction. 7KH¿QDOFXUYHVWREHLQWURGXFHGLQWKHSURSRVHGPHWKRGRORJ\ for the various types of truck bodies and trailer combinations are still being elaborated.

Actros

Table 5. Difference of individual test results from reference value for Cd · Acr and RRC

Figure 7 and Table 6 summarize the results of the tests performed with the Actros vehicle based on the mobile anemometer value without yaw angle compensation (vair,mob). As in the case of CF75 vehicle, the results measured with the EDVHOLQHFRQ¿JXUDWLRQZHUHIRXQGFORVHWRWKHH[SHFWHGYDOXHV (OEM measured average) for both air drag and calculated rolling resistance, with differences of - 0.3 % and 2 % respectively. The variability of the calculated air drag also reached similar values as for the previous vehicle (1.1 % compared to 1.6 %), while the variability of the calculated UROOLQJUHVLVWDQFHFRHI¿FLHQWZDVPXFKORZHUUDQJLQJDW compared to 10.9 %. ,WLVUHPLQGHGWKDWLQWKLVFDVHWKUHHGLIIHUHQWFRQ¿JXUDWLRQVRI the vehicle were tested, baseline, one with slightly improved aerodynamic characteristics (ADlow) one with slightly deteriorated aerodynamic characteristics (ADhigh) in order to


Fontaras et al / SAE Int. J. Commer. Veh. / Volume 7, Issue 1 (May 2014) WHVWWKHVHQVLWLYLW\RIWKHPHWKRG%RWKPRGL¿FDWLRQVZHUH estimated according to OEM in the order of 2 - 4 %, with the exact amount being under investigation via CFD.

The RRC values calculated during the measurements lay very FORVHWRWKHYDOXHVRI¿FLDOO\UHSRUWHGIRUWKHSDUWLFXODUW\UHVLQ most cases and also very close to the results of the tests performed by the OEM. The overall variability also recorded remained at lower levels compared to the tests performed with the rigid truck. Only exception regarding the RRC value measured is that of the ADlow test set. The latter may be attributed to a slightly different stabilization temperature of the tyres during this set of measurements. Still this increase of 7% SUREDEO\KDGQRLQÀXHQFHRQWKH¿QDODLUGUDJFDOFXODWHGDV the CD×A values remain in the expected range. Experience from the present and previous test campaigns suggests that the air drag calculation is rather robust with respect to ÀXFWXDWLRQVLQWKH55&YDOXH1RQHWKHOHVVDQLQYHVWLJDWLRQRQ the topic is still on going.

Repeatability, Reproducibility and Robustness of the Method

Figure 7. Aerodynamic resistance (a) and tyre rolling resistance (b) measurement results for Actros (values normalized by average OEM measured values)

7KHUHVXOWVVKRZWKDWWKHWHVWFRQ¿JXUDWLRQV$'low and ADhigh indeed present the lower and higher air drag, compared to the EDVHOLQHFRQ¿JXUDWLRQ7KHPHDVXUHGGLIIHUHQFHVIDOODOVR within the expected ranges. This indicates that the air drag test method is sensitive enough to capture small changes in aerodynamic characteristics. Quantifying the exact limits of the method's sensitivity will require additional testing. In parallel supplementary work needs to be done with CFD or wind tunnel testing in order to accurately identify the effect of various additions on air drag before attempting to measure them on the test track.

$¿UVWHYDOXDWLRQRIWKHUHSURGXFLELOLW\DQGUHSHDWDELOLW\PHWULFV of the proposed method were calculated based on the results of this study and those reported from previous measurements by the OEMs. Table 7 summarizes the calculated repeatability and reproducibility standard deviations (normalized by the average value recorded). The analysis is based on the results retrieved for corrected yaw wind angle. It was assumed that distribution of the test results is approximately normal Table 7. Repeatability and reproducibility standard deviation of the method

Table 6. Summary of the differences of the measured values with respect to OEM for aerodynamic resistance and rolling resistance (results for baseline configuration)

Given the fact that the method is still developing the achieved ¿JXUHVIRUWKHUHSHDWDELOLW\VWDQGDUGGHYLDWLRQ DQG the reproducibility standard deviation (2.2%-2.9%) are considered satisfactory. Based on these results the repeatability limit (r), which is the value less than or equal to the absolute difference between two results, obtained under repeatability conditions, may be expected to be with a probability of 95%, is in the order of 4.9%-6.7% of the actual &'[$YDOXHPHDVXUHG>8]. :LWKUHJDUGVWRWKHUREXVWQHVVRIWKHPHWKRG¿UVWLQGLFDWLRQV VXJJHVWJRRGFKDUDFWHULVWLFV:KHQFRPSDULQJWKH¿JXUHVRI Table 4 for the C75F vehicle with those of Table 6 for Actros,


Fontaras et al / SAE Int. J. Commer. Veh. / Volume 7, Issue 1 (May 2014) one can observe an improved accuracy (0.3 % compared to 1.3 %) but also lower standard deviations (1.6 % vs 1.1 %) for Cd · Acr. The small differences in the two cases indicate a good robustness of the methodology in the sense, that differences in the instrumentation used during the measurements, ambient conditions and other details had a limited impact on the test results.

SUMMARY/CONCLUSIONS The upcoming European regulation for monitoring CO2 emissions from HDVs is based on vehicle simulation. The regulation will introduce a new methodology for accurately quantifying air drag of HDVs, an important input parameter for the model, based on constant speed tests in a test track. A summary of the methodology and some preliminary tests which aimed to identify its potential were demonstrated in the paper. The results of measurements performed on two HDVs for assessing the proposed methodology suggest good characteristics in terms of measurement sensitivity, precision, reproducibility and robustness. The repeatability standard deviation was calculated in the order of 2% of the air drag value measured while the reproducibility standard deviation in WKHRUGHURI7KHUHVXOWVREWDLQHGVRIDUIXO¿OOWKHQHHGV of the CO2 monitoring procedure for an air drag calculation method that will be in the position to accurately capture the characteristics and advantages of each vehicle body. Additional details and corrections remain to be studied and introduced in order to further improve the repeatability of the method. As presented a correction for the yaw angle of the PHDVXUHGDLUÀRZDURXQGWKHYHKLFOHDOUHDG\LPSURYHVWKH TXDOLW\RIWKHUHVXOWVVLJQL¿FDQWO\$GGLWLRQDOJHQHULFFRUUHFWLRQ IXQFWLRQVZLOOEHHODERUDWHGDQGLQWURGXFHGLQWKH¿QDOYHUVLRQ of the test method. Finally additional approaches are being elaborated on how different types of trailers and body types could be considered by the methodology without the need to perform large series of experimental testing for vehicle air drag calculation. If proven reliable and accurate those will further improve the capability of the simulation based CO2FHUWL¿FDWLRQWRFDSWXUHGLIIHUHQFHVLQ energy performance of different HDVs.

REFERENCES 1. Fontaras, G., Rexeis, M., Dilara, P., Hausberger, S. et al., “The Development of a Simulation Tool for Monitoring Heavy-Duty Vehicle CO2 Emissions and Fuel Consumption in Europe,” SAE Technical Paper 2013-24-0150, 2013, doi:10.4271/2013-24-0150. 2. Hill Nikolas, Finnegan Stephen, Norris John, Brannigan Charlotte, Wynn David, Baker Hannah, Skinner Ian. Reduction and Testing of Greenhouse Gas Emissions (GhG) from Heavy Duty Vehicles -- Lot 1: Strategy, Final Report to the European Commission - DG Climate Action Ref: DG ENV. 070307/2009/548572/SER/C3. Brussels : AEA, 2011.

3. Kies A., Rexeis M., Hausberger S., Schulte L. E.. Reduction and Testing of Greenhouse Gas Emissions from Heavy Duty Vehicles - LOT 2, Development and testing of a certification procedure for CO2 emissions and fuel consumption of HDV. Graz : TU-Graz, 2012. Final report. 4. SAE International Surface Vehicle Recommended Practice, “Measurement of Aerodynamic Performance for Mass-Produced Cars and Light-Duty Trucks,” SAE Standard J2881, Issued June 2010. 5. SAE International Surface Vehicle Standard, “Road Load Measurement Using Coastdown Techniques,” Work in Progress, Jan. 2014. 6. Kies A., Rexeis M, Furian N., Dippold M. Constant Speed Evaluation Tool V1.0, Technical documentation. 2012. Software developed by order of European Commission DG JRC. Report No. I 22/12/Rex EM I 10/12/679. 7. Bode Matthias, Bode Otto, Glaeser Klaus-Peter, Neubauer Joachim, Pflug Hans-Christian. Der Reifenrollwiderstand von Nutzfahrzeugen-Wie korrelieren die Werte bei unterschiedlichen Messverfahren?. Available at: http://www.ipw-automotive.de/vdivortrag.html 8. ISO. 5725-6:1994. Accuracy (trueness and precision) of measurement methods and results-Part 6: Use in practice of accuracy values. Geneva: International Organisation for Standardisation, 1994.

CONTACT INFORMATION For further information please contact Mrs Panagiota Dilara [email protected]

ACKNOWLEDGMENTS The authors would like to acknowledge the support of IVECO in making available the Balocco test track for the needs of this test campaign. The authors would like to thank Mr Konstantinos Anagnostopoulos and Marco Flammini (JRC) for their support in the analysis of the results and Mr Jan Hammer and Leif Erik Schulte (TUV-Nord) for their valuable contribution in the drafting of the test methodology.

DEFINITIONS/ABBREVIATIONS ACEA - Association of European Vehicle Manufacturers EC - European Commission HDV - Heavy Duty Vehicles AD - Air Drag Cd$HURG\QDPLF'UDJ&RHI¿FLHQW RRC5ROOLQJ5HVLVWDQFH&RHI¿FLHQW Ȉ - Standard deviation



APPENDIX APPENDIX CORRECTIONS APPLIED TO THE TEST DATA The following corrections are foreseen for the recorded test data:

VEHICLE SPEED 7KHYHKLFOHVSHHGVLJQDOLVFRQYHUWHGLQWR>NPK@LIEHLQJUHFRUGHGLQDGLIIHUHQWXQLW,QFDVHRIWKHXVHRIWKHYHKLFOH&$1EXVVSHHG signal, the measurement data are corrected by the determination of an averaged calibration factor. The speed signals of every single high speed measurement section are compared with the speed value calculated with the known length of the measurement section and the trigger signals of the opto-electronic barriers or GPS system.

LOCAL AIR FLOW VELOCITY CORRECTION 7KHDLUÀRZYHORFLW\VLJQDORIWKHPRELOHDQHPRPHWHULVFRUUHFWHGLQWKUHHVWHSVIRUWKHLQVWUXPHQWHUURUWKHSRVLWLRQHUURUDQGWKH boundary layer as follows: Step 1. Instrument error correction: The anemometer speed reading is corrected with the calibration factor fvie determined for the individual instrument. Step 2. Position error correction: A correction factor fvpe is obtained with the calibration test data during the warm-up phase. The data must contain at least 5 measurements per driving direction and per measurement section, at low wind conditions (NPK@ va2 DYHUDJHDLUÀRZVSHHGGXULQJFDOLEUDWLRQWHVWPHDVXUHGLQGULYLQJGLUHFWLRQ>NPK@ fvpe VSHHGSRVLWLRQHUURUFRUUHFWLRQIDFWRU>@ vvac DYHUDJHYHKLFOHVSHHGGXULQJFDOLEUDWLRQWHVW>NPK@ 7KHSRVLWLRQHUURUFRUUHFWLRQIDFWRULVDSSOLHGWRWKHGDWDWRUHFHLYHWKHXQGLVWXUEHGDLUÀRZVSHHG8)

where: fvpe VSHHGSRVLWLRQHUURUFRUUHFWLRQIDFWRU>@ vUF 8QGLVWXUEHG)ORZZLQGVSHHG>NPK@

Step 3. Boundary layer correction &RUUHFWLRQIRUWKHZLQGVSHHGDQGZLQGDQJOHYDULDWLRQZLWKKHLJKWEDVHGRQWKHDVVXPSWLRQIRUZLQGSUR¿OHXVLQJWKHSRZHUODZ ERXQGDU\OD\HUĺ:RUNLQSURJUHVV

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE International. Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the paper.