E!ects of Speci"c Operational Weighting Factors on Standardized Measurements of Tractor Engine Emissions

Transcription

1 J. agric. Engng Res. (1999) 74, 347}353 Article No. jaer , available online at on E!ects of Speci"c Operational Weighting Factors on Standardized Measurements of Tractor Engine Emissions Per-Anders Hansson ; Olle NoreH n ; Mats Bohm Department of Agricultural Engineering, P.O. Box 7033, Uppsala, Sweden; per-anders.hansson@lt.slu.se Swedish Institute of Agricultural Engineering, P.O. Box 7033, Uppsala, Sweden; mats.bohm@jti.slu.se (Received 9 September 1998; accepted in revised form 17 July 1999) Measurement methods and weighting factors used to calculate typical emissions values for agricultural tractors are primarily developed for other types of vehicles and the results cannot be assumed to be representative of agricultural driving conditions. The purpose of this work was to calculate speci"c emission test weighting factors for a range of typical agricultural tractor work operations. The aim was also to compare emission values calculated using the general standardized weighting factors with values obtained using weighting factors typical for speci"c operations. In order to calculate new weighting factors, time series from sensors mounted on a standard medium-size tractor were recorded during typical agricultural operations. The data recorded were then used to calculate engine speed and loading torque. The recorded points were divided into groups, one for each of the test modes in the standard, and the weighting factors calculated from the relative amount of recorded points in each group. One major conclusion is that the amount of emissions from the tractor engine, measured relative to the output power from the engine, show very large variations depending on the characteristics of the work operation. The use of one single set of weighting factors not adapted to the speci"c operation may therefore give very incorrect results Silsoe Research Institute 1. Introduction There is growing concern about engine emissions from tractors and their impact on the environment. The importance of correct emission values and representative measurement methods has thus become very obvious. Recorded time series of engine speed and loading torque on a tractor engine show that the typical load varies widely depending on the work operation performed (NoreH n and BroweH n, 1988). Available data also show that the amount of emissions from a diesel engine, measured in relation to the energy in the combusted fuel, is highly dependent on engine speed and loading torque (Hedbom, 1994). Standardized measurements of emissions from engines used in agricultural tractors are normally performed according to the ECE R49 or the ISO 8178 standards. Since 1993 ECE R49 has been valid for heavy duty road vehicles in Sweden, while ISO 8178 test cycle C1 (in the following text, cited as ISO 8178) is valid for o!-road vehicles. The emissions are measured at eight (ISO 8178) or 13 (ECE R49) combinations of engine speeds and loading torques and the results are weighted together to form a single value. The weighting factors are general for a wide range of diesel-fuelled vehicles and are not designed for agricultural conditions. The results therefore cannot be assumed to be representative of agricultural tractors. Treiber and Sauerteig (1991) describe the development of ISO They state the need of application-oriented test cycles with corresponding weighting factors for the future optimization of engines used in o!-road vehicles. No such test cycles, however, are reported. Measurements on an engine normally used in wheel loaders (Hedbom, 1994) showed that CO emissions were reduced by 30%, oxides of nitrogen (NO ) emissions were reduced by 10% and hydrocarbon (HC) emissions increased by 13% when measuring according to ISO 8178 instead of ECE R49. The main part of these di!erences was due to the use of other load points and /99/120347#07 $30.00/ Silsoe Research Institute

2 348 P.-A. HANSSON E¹ A. weighting factors. Cornetti et al. (1988) compares ECE R49 with American transient test cycles for engines in diesel trucks. Their results indicate that the ECE R49 weighting factors are representative of driving with heavy duty diesel trucks on European roads. The lack of correct weighting factors for use in the calculation of representative emission values for agricultural operations contributes to substantial uncertainties in important areas; the results from life cycle assessment (LCA) (see for example, Audsley et al., 1997) analyses performed in order to quantify and decrease the environmental load of food production is very sensitive to variations in emission values (Hansson & Mattsson, 1999), and incorrect data may therefore contribute to incorrect advice and measures taken; the optimization of new engines and fuels, for example biofuels, with respect to discharge of emissions will be incorrect if the standardized measurement methods are misleading; national and international emission statistics may be very important in the future, for example in establishing emission rights and taxes; and, since emissions from agricultural and forestry vehicles are major contributors to total emission levels, it is of course most important that the "gures from this sector are correct. The purpose of this work was to calculate speci"c emission test weighting factors for a range of typical work operations. The aim was also to compare emission values calculated with the general weighting factors de- "ned in the two standards with values obtained using the weighting factors typical for speci"c operations. In order to calculate new weighting factors, time series from sensors mounted on a standard medium-size tractor were recorded when performing di!erent typical agricultural operations. The recorded parameters were then used to calculate engine speed and loading torque. The recorded points were divided into groups, one for each of the test modes in the standard, and the weighting factors were calculated from the relative amount of recorded points in each group. The load on the engine of an agricultural tractor was recorded when performing several driving operations. The tractor used was a Valmet 805, a 70 kw standard tractor made in Finland with turbo and four-wheel drive. The tractor was supplied with standard tyres and also in other respects equipped and adjusted for normal Swedish conditions. The data acquisition system used is described by BroweH n (1988). The sample time was 7)5 s and the maximum measurement length was 2)8 h. The recorded signals were used to calculate output parameters such as exhaust and intake air temperature, gear level, engine speed, ground speed, wheel slip and fuel consumption. The calibration of the measurement system is described by NoreH n and BroweH n (1988). The system was calibrated by braking in a test bench and simultaneous determination of fuel consumption, exhaust temperature and engine speed. Also, the ambient temperature was varied in order to "nd correction parameters for this parameter. The calibration results show that it was possible to calculate engine power output with high accuracy (correlation coe$cient r"0)9973) using data for engine speed and fuel consumption. It was also possible, in static conditions, to reach high accuracy using exhaust temperature and engine speed as input parameters. These data showed, however, a slower response to fast variations in the loading torque and the power values used in this study are therefore calculated from fuel consumption. About 40 recordings were made under varying circumstances and when performing di!erent working operations. After a preliminary analysis, the most representative were chosen for a more thorough analysis. For this study, the following operations were considered to be most interesting: (1) harrowing with a 5)7 m wide seedbed harrow adjusted to normal seed depth; (2) sowing with a 3)0 m wide towed seed drill adjusted to normal seed depth; (3) stubble cultivation with a 3)5 m wide disc harrow adjusted to 0)05 m working depth; (4) transport of 8)8 t of sand on the highway and in town tra$c using an 8 t single-axle trailer; (5) transport of grain from the "eld to the farm silo using a 12 t tandem-wheel trailer; and (6) ploughing with a 1)55 m wide mounted reversible plough adjusted to a working depth of 0)23 m. 2. Material and methods 2.1. Measurement of engine load 2.2. Calculation of weighting factors The ECE R49 test cycle states that the emissions are measured for 13 static working modes of operation. Modes 1, 7 and 13 are de"ned at low idle speed without load. Modes 2, 3, 4, 5 and 6 are de"ned at the engine speed with maximum torque (&&intermediate speed'') and with loads of 10, 25, 50, 75 and 100% of the maximum torque, respectively. Modes 8, 9, 10, 11 and 12 are de"ned at the maximum rated engine speed with loads of 100, 75, 50, 25 and 10% of the maximum torque at this speed, respectively. Two measurement modes (7 and 13) are just repetitions of mode 1, which is why the results from the

3 MEASUREMENT METHODS AND WEIGHTING FACTORS OF TRACTOR ENGINE EMISSIONS 349 Table 1 De5nition of break points, p1}p10, on the lines dividing the engine speed}torque diagram into 11 parts for the ECE R49 standard test modes, m1}m12 (see Fig. 1) Point no. Engine speed Torque Fig. 1. Engine speed ~ torque diagram divided into 11 parts (~~, ECE R49, see Table 1) and 8 parts (} }}}, ISO 8178); p1-p10, break points; m1}m12, ECE R49 modes tests describe the emissions at 11 di!erent combinations of engine speed and loading torque. The ISO 8178 test cycle consists of 8 modes, corresponding to modes 1, 4, 5, 6, 8, 9, 10 and 12 in ECE R49. No repetitions of the low idle mode are included and the weighting factors are not the same as in ECE R49. The two test procedures di!er also in some other details, for example regarding the use of the engine cooling fan (Hedbom, 1994). To calculate new weighting factors, the area in the engine speed versus torque diagram is divided into 11 smaller areas, one representative of each combination of engine speed and torque de"ned in the ECE R49 standard. For the analysed time series, the relative frequency of recorded engine speed}torque combinations in each of the 11 areas are then calculated. p1 (R49m1#R49m2)/2 0 p2 R49m2 (R49m2#R49m3)/2 p3 R49m2 (R49m3#R49m4)/2 p4 R49m2 (R49m4#R49m5)/2 p5 R49m2 (R49m5#R49m6)/2 p6 R49m12 (R49m8#R49m9)/2 p7 R49m12 (R49m9#R49m10)/2 p8 R49m12 (R49m10#R49m11)/2 p9 R49m12 (R49m11#R49m12)/2 p10 (R49m2#R49m12)/2 0 In Fig. 1, the positions of the lines dividing the diagram into 11 areas are described. Engine speed and loading torques in the points marked with small circles are de- "ned by the ECE R49 standard and described in the previous part. These points are used to calculate break points at the lines dividing the areas. The break points are de"ned according to Table 1, where R49m1, R49m2, etc., refer to values de"ned for the ECE R49 measurement modes 1, 2, etc. The curves are based on values valid for the Valmet 805 engine, i.e. a maximum torque of 382 N m at 1400 min and a torque of 303 N m at 2100 min (SMP, 1985). The maximum power is 70 kw. The ISO 8178 includes no measurements at intermediate engine speed with 10 and 25% load (modes 2 and 3 in ECE R49) or at maximum speed with 25% load (mode 11 in ECE R49). In the calculations of speci"c work operation weighting factors for ISO 8178, the area for mode no. 3 in Fig. 1 is therefore included in the area Table 2 Emission data following ECE R49 used in the calculations of emission values Diesel Rapeseed oil methyl ester Mode no. CO, g/h NO,g/h HC, g/h Power, kw CO, g/h NO,g/h HC, g/h Power, kw m1 38)2 88)2 6) )3 68)1 3)10 0 m2 52)1 110)5 8)72 5)8 47)0 122)8 2)97 5)8 m3 33)7 197)0 7)64 14)1 28)4 204)9 2)55 14)5 m4 15)3 309)1 8)06 28)1 13)6 321)3 3)04 28)5 m5 12)2 464)8 6)86 42)5 12)6 518)9 3)79 42)6 m6 18)7 583)3 7)38 57)5 16)2 642)2 4)57 55)2 m7 39)2 86)2 5) )4 69)3 3)43 0 m8 29)5 584)2 9)86 70)1 21)0 692)9 6)19 69)3 m9 25)4 409)3 10)90 53)1 18)4 475)1 5)68 52)3 m10 30)3 224)5 11)90 35)5 25)6 258)6 5)16 35)4 m11 35)7 194)0 11)46 19)6 31)6 187)4 4)82 17)6 m12 52)1 128)8 12)30 8)4 40)3 129)7 4)90 6)6 m13 43)1 75)1 5) )0 58)6 3)42 0 NO, oxides of nitrogen; HC, hydrocarbons.

4 350 P.-A. HANSSON E¹ A. Table 3 Description of analysed measurements Operation Time, min Distance, km Average velocity, km/h Average slip, % Average speed, min 1 Average power, kw Harrowing 105)1 10)31 5)88 23) Sowing 168)3 20)55 7)33 6) Stubble cult) 166)0 9)25 6)96 8) Transport 1 151)5 68)3 27)05 3) Transport 2 104)2 21)2 12)20 7) Ploughing 158)1 9)83 3)43 16) for mode no. 4, the area for mode no. 11 in the area for mode no. 10 and the area for mode no. 2 in the area for mode no Calculation of the e+ects on emission values In order to study the e!ects on weighted emission values when weighting factors are varied, available data from emission measurements made according to the ECE R49 standard were used. Thirteen mode data of this kind also include data for the 8 modes de"nediniso8178. Results from each test mode in a ECE R49 test are normally not published and no such data were available for the Valmet 805 engine. However, emission data from a test comparison of using di!erent fuels in a Valmet 420 DS engine were available. This engine is used in tractors similar to the Valmet 805. Both engines are four cylinder turbo-charged engines with a cylinder volume of 4)4 l, and a maximum power of 70 kw. The 420 DS engines are used in Valmet 6400 tractors. The emission measurements were performed at the former Swedish National Machinery Testing Institute (SMP) and summarized in one of their reports (SMP, 1993). The available emission data made it possible to study the sensitivity to variations in weighting factors for both standard diesel and rape seed oil methyl ester (RME) produced by the company Soab AB and based on Swedish rape seed oil. The engine was standard-equipped and factory-adjusted for driving on diesel fuel. The results for each mode in the ECE R49 standard are described in Table 2. Data were available for the emissions CO, NO and HC. Also the discharge of particles was measured in the standard test, but these data are only measured as a cumulative amount for the total test cycle, and it was therefore not possible to include the particles in the calculations. 3. Results 3.1. =eighting factors Some characteristics of the analysed operations are described in Table 3. As an example of the distribution of the operating conditions of the engine, Fig. 2 shows the distribution for ploughing. Each point represents 7)5sof driving and the total number of recorded points represent approximately 1 h of driving. Table 4 shows the distribution of engine speed and calculated loading torque in 11 groups according to the ECE R49 standard. These factors are used as weighting factors when calculating emission values in the next part. Table 5 shows the distribution of recorded loads in the eight modes de"ned by ISO For comparison, the weighting factors de"ned in the standards are also described in the tables. The sum of every line in Tables 4 and 5 is 1) Emission values Fig. 2. Distribution of the load at the engine when ploughing; each data point represents 7)5 s, m1}m12, ECE R49 modes The weighting factors described in Table 4 and the emission test results described in Table 2 are used to

5 MEASUREMENT METHODS AND WEIGHTING FACTORS OF TRACTOR ENGINE EMISSIONS 351 Table 4 Distribution of the engine operating conditions in the groups de5ned from ECE R49 modes Distribution of engine operations by ECE R49 modes Operation m1, m7, m13 m2 m3 m4 m5 m6 m8 m9 m10 m11 m12 Harrowing )02 0)78 0) Sowing 0)17 0)02 0)20 0) )42 0)04 0 Stubble cult) 0) )89 0)07 0)02 Transport 1 0)02 0)01 0)01 0)02 0)01 0 0)01 0)13 0)32 0)34 0)13 Transport 2 0)37 0)06 0)08 0)02 0)02 0 0)04 0)09 0)11 0)16 0)05 Ploughing 0)14 0)01 0)09 0)32 0) )05 0) ECE R49 0)25 0)08 0)08 0)08 0)08 0)25 0)10 0)02 0)02 0)02 0)02 calculate weighted emission values. These values are shown in Table 6. Emissions are calculated both for standard diesel and for RME. Table 7 also shows the results when calculated in relation to the energy produced by the engine. 4. Discussion The standardized measurement methods ECE R49 and ISO 8178 are based on measurement at static load conditions with no respect to the engine's emission characteristics at fast load variations. Other calculations using the same data material as in this study indicate that the rather slow variations in load present in the reported driving operations can be analysed with the semi-static method used (Hansson et al., 1998). This conclusion is con"rmed by Cornetti et al. (1988). However, for operations more frequently including fast variations in load, for example when working with a front loader, the dynamic e!ects have to be included. Dynamic emission test cycles have been developed, for example for passenger cars and buses, but these methods are not applicable to agricultural tractor engines. The emission values published for di!erent diesel engines show rather wide variations. The lack of published data for other tractor engines makes it di$cult to estimate if the engine studied represents the average modern tractor engine. However, unpublished data from measurements made on several Valmet engines indicate that the CO and NO values for the 420 DS engine are close to average, while the tested 420 DS engine gives somewhat lower HC values than average. Most of the working operations analysed include long periods at low loads but high engine speeds. The weighting factors for that type of load are small in both standards. On the other hand, weighting factors for emissions at high loading torque and intermediate engine speed are generally higher than indicated in the recorded measurements. One result of the weighting factors not being adapted to agricultural conditions is that the discharge Table 5 Distribution of the engine operating conditions in the groups de5ned from ISO 8178 modes Distribution of engine operations by ISO 8178 modes (ECE R49 modes in parentheses) Operation m1 (m8) m2 (m9) m3 (m10#m11) m4 (m2#m12) m5 (m6) m6 (m5) m7 (m3#m4) m8 (m1#m7 #m13) Harrowing 0)78 0) ) Sowing 0 0 0)46 0) )35 0)17 Stubble cult )96 0) )02 Transport 1 0)01 0)13 0)66 0)14 0 0)01 0)03 0)02 Transport 2 0)04 0)09 0)27 0)11 0 0)02 0)10 0)37 Ploughing 0 0)05 0)39 0)01 0 0)01 0)41 0)14 ISO )15 0)15 0)15 0)10 0)10 0)10 0)10 0)15

6 352 P.-A. HANSSON E¹ A. Table 6 Calculated emissions Operation Diesel Rapeseed oil methyl ester CO, g/h NO,g/h HC, g/h CO, g/h NO,g/h HC, g/h Harrowing 26) )1 20) )07 Sowing 33) )41 28) )92 Stubble cult. 36) )8 27) )13 Transport 1 38) )4 29) )94 Transport 2 37) )77 35) )99 Ploughing 29) )38 25) )96 NO, oxides of nitrogen; HC, hydrocarbons. of NO emissions is overestimated when using general weighting factors. The recorded harrowing operation shows a high average power output from the engine. The CO and HC emissions (measured in g/kwh) are much lower for this operation than for the others. The results also show that di!erent types of transport work give very varying loads on the engine and big variations in the discharge of emissions, also when related to the transport distance. Hedbom's (1994) conclusion that ISO 8178 tests result in lower CO and NO values but higher HC values than ECE R49 tests is con"rmed in this work. This is very important to take into acount, when engines not tested according to the same standard, are compared. One reason for the di!erences is the higher weighting of maximum speed values in ISO The CO and NO emissions, if measured in g/kwh, are lower at maximum speed than at intermediate speed. 5. Conclusions The driving operation speci"c weighting factors presented are useful when estimating the emissions from a tractor used for a speci"c work operation. The factors may also be useful when optimizing engines for tractors mainly used in a certain type of work, for example heavy soil cultivation or transport work. The amounts of emissions from the tractor engine, measured in relation to output power, show very large variations depending on the characteristics of the work operation. The use of one single set of weighting factors for all engine types therefore results in very large uncertainties. Additionally, the use of one single emission value valid for all operations, as for example in life cycle assessment (LCA) analyses, may give misleading results. One speci"c result of the weighting factors not being adapted to agricultural conditions is that the discharge of NO emissions normally is overestimated when using the standardized weighting factors. References Audsley E (ed.); Alber S; Clift R; Cowell S; Gaillard G; Hausheer J; Jolliett O; Kleijn R; Mortensen B; Pearce D; Roger E; Teulon H; Weidema B; van Zeits H (1997). Harmonisation of environmental life cycle assessment for Table 7 Calculated emissions Operation Diesel Rapeseed oil methyl ester CO, g/kwh NO,g/kWh HC, g/kwh CO, g/kwh NO,g/kWh HC, g/kwh Harrowing 0)39 8)25 0)151 0)31 9)87 0)093 Sowing 1)45 8)89 0)406 1)24 9)50 0)169 Stubble cult. 1)10 6)54 0)353 0)82 7)48 0)155 Transport 1 1)37 8)20 0)409 1)09 9)21 0)183 Transport 2 2)07 10)6 0)483 2)03 11)3 0)226 Ploughing 1)07 8)84 0)346 0)92 9)50 0)146 ECE R49 0)99 10)63 0)249 0)94 11)8 0)129 ISO )83 9)65 0)262 0)76 10)9 0)134 NO, oxides of nitrogen; HC, hydrocarbons.

7 MEASUREMENT METHODS AND WEIGHTING FACTORS OF TRACTOR ENGINE EMISSIONS 353 agriculture. Final Report. Concerted Action AIR3-CT Silsoe Research Institute, Bedford, UK BroweH na(1988). A data acquisition system for "eld applications. Swedish Institute of Agricultural Engineering, Paper presented at AG ENG 88, Paris Cornetti G; Klein K; FraK nkle G J; Stein H J (1988). US transient cycle versus ECE R49 13-mode cycle. Society of Automotive Engineers (SAE), Paper Hansson P-A; BurstroK m A; NoreH n O; Bohm M (1998). BestaK mning av motoremissioner fra n arbetsmaskiner inom jord och skogsbruk [Engine emissions from agricultural tractors and forestry machines]. Department of Agricultural Engineering, Uppsala, Sweden, Report 232 Hansson P-A; Mattsson B (1999). In#uence of derived operation-speci"c tractor emission data on results from an LCI on wheat production. International Journal of LCA, accepted for publication Hedbom A (1994). Emission tests of two Volvo/VME heavy duty o! road engines. AB Svensk Bilprovning, Haninge, Sweden, Report MTC 9307A NoreH n O; BroweH n A(1988). Determination of tractor power based on measurements of fuel consumption and exhaust temperature. Swedish Institute of Agricultural Engineering, Paper presented at AG ENG 88, Paris SMP (1985).Vilken e!ekt har traktorn? [What is the maximum power of the tractor?]. Swedish National Machinery Testing Institute, Uppsala, Sweden, Note 2695 SMP (1993). Sca" 101 i jak mfok relse med andra dieselbrak nslen [Sca" 101 compared with other diesel fuels]. Swedish National Machinery Testing Institute, Uppsala, Sweden, Note 3390 Treiber P; Sauerteig J (1991). Present and future european exhaust emission regulations for o!-road diesel engines. Society of Automotive Engineers (SAE), Paper