Sei sulla pagina 1di 17

Paper Number: 2005-01-2147

Comparison between the exhaust particles mass determined by the European regulatory gravimetric method and the mass estimated by ELPI
Efthimios Zervas, Pascal Dorlhne
Renault

Laurent Forti, Cyriaque Perrin


IFP

Jean-Claude Momique, Richard Monier


PSA Peugeot-Citren

Hok Ing, Batrice Lopez


UTAC
Copyright 2005 SAE International

ABSTRACT
Electrical Low Pressure Impactor (ELPI) is often employed to measure the particle number and size distribution of internal combustion engines exhaust gas. If appropriate values of particle density are available, the particle mass can be estimated by this method. Exhaust particles of three Euro3 passenger cars (one gasoline operating under stoichiometric conditions, one Diesel and one Diesel equipped with Diesel Particulate Filter) are measured using the current European regulations (gravimetric method on the are New European Driving Cycle) and estimated by ELPI particle number and size distribution. Different values for particle density are used to estimate the particle mass using all ELPI stages or only some of them. The results show that the particle mass estimated by ELPI is well correlated with the mass determined by filters for PM emissions higher than 0.025 g/km. This correlation is not very good at lower emissions.

the case of vehicle exhaust gas are the Electrical Low Pressure Impactor (ELPI [2-7]), Scanning Mobility Particle Sizer (SMPS, [8]) and Particle Counters [9, 10], but many others are also presented in literature. A detailed description of several analytical methods can be found in two articles [11, 12]. Currently, SMPS cannot be used for the analysis of particle distribution on the New European Driving Cycle (NEDC), because it has an insufficient resolution time of some minutes. In this work, ELPI was used to estimate the exhaust particles mass using the particle number and distribution of three Euro3 passenger cars (PC): a gasoline vehicle operating under stoichiometric conditions, and two Diesel, with and without Diesel Particulate Filter (DPF). A pragmatic protocol, using the European regulatory conditions (tests on the New European Driving Cycle (NEDC)) is used. The target is to find out if ELPI can be used to estimate exhaust particle mass, using the particle density presented in literature or empirically correlated density values. Four laboratories participated to program: IFP, PSA Peugeot-Citren, Renault and UTAC. The estimated mass is obtained using different assumptions and compared with the particle mass determined by the current European regulatory gravimetric method.

INTRODUCTION
Current European regulations are based on a gravimetric measurement of particles emitted from Diesel passenger cars [1]. But, as the emission levels decrease, other methods are proposed to replace the current gravimetric method, mainly based on exhaust particle number determination. Currently, there are many methods for particle number and/or size determination. The most common used in

TECHNICAL BACKGROUND OF ELPI


The function principle of ELPI is the following: the sample passes through a unipolar positive polarity charger, where particles are charged electrically by 1

small ions produced in a corona discharge. Each particle charge is a function of its size. The charged particles are then size classified in a low-pressure impactor. The stages of the impactor are insulated electrically and each stage is connected individually to an electrometer current amplifier. The charged particles are collected in a specific impactor stage as a function of their size and produce an electrical current, which is recorder by a respective electrometer channel. A larger charge correlates to a higher particle population. The current value of each panel is proportional to the number of particles collected, and thus to the particle concentration in a particular size range. More details about ELPI can be found in literature [2, 13]. Figure 1 presents the principle of this device.

sometimes performed. To achieve it, an effective density of particles must be estimated. Ahlvik [3] used ELPI to measure the particle number distribution of a passenger car (model year 1993) and a heavy-duty engine on two driving cycles (European Driving Cycle (EDC) and Federal Test Procedure (FTP) for the passenger car and US transient cycle and European Steady-state Cycle (ESC), for the heavy-duty engine). A Differential Mobility Analyzer (DMA) was also used for these measurements. The effective particle density was estimated by DMA and was used to estimate the particle mass using ELPI distribution, but the authors did not deeply examine this method. The effect of dilution ratio and driving cycle on particle distribution was also presented. Pattas [4] examined the effect of DPF on particle size distribution using ELPI. The influence of sulphur and Ce content in the fuel on particle distribution was presented on NEDC, where it was shown that DPF decreases the particle number. Shi [14] used ELPI to study the number and distribution of particles emitted from a Diesel engine tested on a 11 mode steady state cycle. A comparison of particle emissions at different engine speeds and load was performed using ELPI and SMPS, where both techniques give similar results. Maricq [6] presented a comparison of particle distribution of three Diesel and three gasoline vehicles (model years 1995-1997), operating at steady speeds and measured by SMPS and ELPI. The effective density of particles was also estimated and it was used to calculate the particle mass. However, no comparison with mass collected on filters is given. ELPI was also used on FTP.

Figure 1. The operating principle of ELPI. Keskinen [2] presented the operation principles of ELPI (charger, impactor cut diameters and electrometer) and compared it with SMPS. An atomizer was used to generate an aerosol, which was measured by SMPS and ELPI. This work concluded that both instruments measure the same particle distribution. Marjamki [13] evaluated the performances of ELPI at a nominal flow rate of 10 l/min. The impactor, charger and their calibration, using two different aerosol generation techniques, is presented. This author presented the collection efficiency of ELPI impactor and the performance of ELPI charger (which were estimated satisfactory) and the comparison of the size distribution obtained by ELPI and SMPS using a di-octyl-sebacate (DOS) aerosol. He concluded that both techniques give similar particle distribution. This instrument is used for the measurement of particle number and distribution in vehicles exhaust gas, but it also can be used to estimate the emitted particles mass. A comparison with the mass collected on filters is

Tsukamoto [15] used ELPI to measure the particle number distribution of a heavy-duty engine exhaust gas. Particle mass was estimated by ELPI measurements and compared with the mass collected on filters. Generally, the estimated mass was 1.5-2 times higher than the mass collected on filters. Khalek [5] analyzed the particle number distribution of a heavy-duty engine exhaust gas on FTP. As the previous author, he concluded that ELPI overestimates the total mass emissions comparing to filter mass measurements. Andrews [16] estimated the effective particle density to 1.5-0.2 g/cm3 as a function of size and concluded that the estimated mass can be very different from the mass collected on filters. This difference can be very important for particles bigger than 1 m. One reason is that ELPI overestimates particle numbers by up to 2 orders of magnitude at this particle size. Virtanen [17] estimated the effective density of particles to 1.1-1.2 g/cm3, as a function of their size and presented that dilution had a strong effect on this density. Ristimki [18] used ELPI and SMPS to estimate 2

the effective density of different aerosols, but not of exhaust gas particles. Choi [19] used ELPI to study the emissions of nanoparticles of a heavy-duty Diesel engine and the effect of the oxidation catalyst on mass and distribution of the emitted particles. Maricq and Xu [20] used DMA and ELPI to determine the effective density and fractal dimension of particles emitted from flames and motor vehicle exhaust gas. This work concluded that the effective densities of particles emitted from of two Diesel engines and a direct-injection SI engine were identical. ELPI has some limitations for particle number measurements. Van Gulijk et al [21, 22] presented a list of no-ideal behaviour of ELPI, as particle bounce, wall or inter-stage loss, overloading or surface built up, losses due to electrostatic effects and to charger non-ideal efficiency. Using a steady-state speed, a continually decrease of small particles and a continually increase of the bigger ones was observed. The authors explained that this was due to impactor overloading. In our point of view, even if a part of these changes is due to impactor overloading, an extremely constant source of particles and different analytical methods must be combined to validate such a conclusion. Van Gulijk et al [23] presented that ELPI underestimated the apparent size of particles and, as a result, their number was overestimated. Three other programmes must be mentioned. ACEA conducted two important research programs studying the emissions of fine particles [24, 25]. In the first study, eleven Diesel and five gasoline vehicles were tested on NEDC and steady speeds by two laboratories. The emissions of regulated pollutants and of the particle number (determined by SMPS) were presented in the final report, but ELPI was not tested in this work. Three Diesel and four gasoline passenger cars were tested in the second program. The stability of ELPI was determined over 18 weeks (one test per week) using a reference vehicle at a steady speed of 100 km/h. With no apparent explanation, the particle number decreases about 35% between the beginning and the end of this programme. The third program is the Swiss contribution to the GRPE particle measurement programme [12]. In this study, 24 particle measurement techniques were tested using a heavy-duty engine and a combustion aerosol generator (CAST system). The limits of detection (LOD) of ELPI was found to be very close (90%) to the measured particle concentration when a particulate filter was used. In these three studies, the ELPI was not used to estimate the particle mass. To estimate the particle mass by ELPI particle number, the particle density must be first measured or estimated. Different particle densities are presented in literature: an average density of 1.0 g/cm3 [14, 15], or 1.7 g/cm3 [26], or 0.5 g/cm3 [7]. However, particle density is a function of size [3, 7, 16, 17, 26]. The smaller particles are spherical

and their effective diameter determines the particle density. The larger the particles the more primary particle they contain, leading to much lower values of effective density [3]. The values of 1 g/cm3 at 50 nm and 0.3 g/cm3 at 300 nm [7], or 1.5-0.2 g/cm3 [16], or 1.2-0.3 g/cm3 [20] or 1.6-0.2 g/cm3 [3], or 1.1-1.2 g/cm3 [17] are suggested. A correlation between the particle mass obtained from calculations and the mass measured on filters is sometimes presented in literature. The conclusions of Tzsukamoto [15] were that an average density cannot be assumed for all particle size and that ELPI, due to the several assumption used (same density of particles of each size, each particle size of each stage, charge efficiency,), generally predicts 1.5-2 times more particle mass. Shi [14] presented that ELPI and SMPS predict 1.3-1.6 times more mass than this collected on filters. Khalek [5] concluded that ELPI overestimates the total mass emissions comparing to filter mass measurements in the case of a heavy-duty engine. Andrews [16] concluded that a comparison of the mass estimated by ELPI can be very different from this collected on filter. Witze [7] used only the first 6 ELPI stages and an empirical method to adjust the particle density so to obtain the same mass as on the filters.

EXPERIMENTAL SECTION
Three passenger cars were used for this study: a gasoline PC operating under stoichiometric conditions (vehicle 1), a Diesel PC (vehicle 2) and a Diesel PC equipped with DPF (vehicle 3). Table 1 provides their main characteristics. Twenty five ppm of commercially used Ce-based additive were added in fuel in the case of the DPF equipped vehicle, in order to decrease the temperature required for the DPF regeneration. Vehicle 1 Renault Mgane Gasoline 1130 1600 4 4 MPI Vehicle 2 Peugeot 307 Diesel 1360 1997 4 2 Common Rail HDI Vehicle 3 Peugeot 307 Diesel 1360 1997 4 4 Common Rail HDI

Type Fuel Inertia class (kg) Displacement (cm3) Number of cylinders Valves per cylinder Injection system Combustion system Max. power (kW) Emission limits EGR type

66 80 Euro3 Euro3 Electric Electric closed closed loop loop After-treatment TWC DOC DOC+DPF Table 1. Main characteristics of the three PC used. 3

66 Euro3

A low sulphur content (less than 10 ppm) is used in this study. The main characteristics of these fuels are presented in the table 2. The same lubricant was used for all three vehicles; it contains less than 0.4% of sulphur. Characteristic Density at 15C (kg/m3) Viscosity at 40C (cSt) Diesel Characteristic Density at 15C 834 (kg/m3) Gasoline 762

air (120C) was used for dilution to minimize the nucleation particles. The total length of the antistatic tube, used to connect the sample probe installed on the dilution tunnel and the ELPI, is less than 3 meters. The four laboratories did not use the same set-up of dilution tunnels: Labs 1 and 2 had the same dilution tunnel for gasoline and Diesel vehicles, Lab 4 used two dilution tunnels (one for gasoline and one for Diesel vehicles), but with a common sampling point, while Lab 3 had two dilution tunnels with two sampling points. These experimental facilities are representative of those used in other European laboratories. Tunnel background measurements (blanks) were performed during 5 minutes before the tests, to check the cleanliness of the tunnel, and during 5 minutes at the end of the test, with engine switched off, to control the initial test conditions. The tunnel background values were not subtracted from the particle number measurements. To correlate the particle mass estimated by ELPI with the mass collected on filters, the following procedure is used: total particle mass is the sum of each stage particle mass over the NEDC. The particle mass of each stage is calculated as the product of particle number, volume of each particle and its effective density. To calculate the volume of each particle, three diameters are used: the lower diameter of each ELPI stage (DL), the intermediate (DI) and the upper one (DU). DI is defined as DI =

1.82 3.00 H/C ratio Octane number 89.1/100. Cetane number 53.8 (MON/RON) 2 Flash Point (C) 82 Distillation (C) Distillation (C) Initial Boiling Point 34.9 62.9 Initial Boiling Point 197 10% 222 20% 77.4 10% 235 50% 107.0 20% 278 90% 144.8 50% 331 95% 154.8 90% 347 Final Boiling Point 184.8 95% 15 Final Boiling Point 359 E 70C (vol %) 30.4 E 100C (vol %) 38.3 E 250C (vol %) 69.6 E 150C (vol %) 92.5 E 300C (vol %) 95.9 Composition E 350C (vol %) Total paraffins (vol 59.4 %) Composition Total olefins (vol 0 8 Sulphur (mg/kg) %) Polycyclic 49.6 Total aromatics Hydrocarbons (% 4.5 (vol %) wt) Ce (ppm, the DPF 1.73 25 Benzene (vol %) vehicle only) Polyaromatics (wt 4.5 %) Total oxygenates <0.2 (vol %) 8 Sulphur (mg/kg) <2 Lead (mg/l) Table 2. Fuel characteristics. Three tests were performed on NEDC (cold start) and the regulated pollutants and CO2 emissions were measured according to current European regulations [1]. The experimental procedure used is the following: a cold NEDC, 24 hours of conditioning at 20C, and three cold NEDC with a conditioning of 24 hours between each cycle. The results of the last three NEDC were taken into consideration. In the case of the DPF equipped Diesel PC, a DPF regeneration is performed at each laboratory before these cycles. Number and size distribution measurements were performed using a DEKATI ELPI, covering particle cut sizes from 8 nm to 10 m. Three laboratories used an ELPI sampling of 10 L/min, while Lab1 used an ELPI sampling of 20 L/min. As the use of a thermodenuder can induce high particle losses [27, 28], a DEKATI ejector type dilutor heated at 130C with hot nitrogen or

DL DU .

The values of effective density tested are: - Four constant with the size values: 1.5 g/cm3, 1.0 g/cm3, 0.8 g/cm3 and, as the initial results were not satisfactory, the best simulated value to adjust the estimated ELPI mass values to the measured filter mass values. - Two values depending linearly with size: one with 1.2 g/cm3 at the first ELPI stage and 0.3 g/cm3 at the last one, and, as the initial results were not satisfactory, the best simulated linear function to adjust the estimated ELPI mass values to the measured filter mass ones. - Two values as previously, but depending linearly on the log of the particle diameter. - The values presented by Ahlvik [3], and as the initial results were not satisfactory, a coefficient (the same for every ELPI stage) is used to adjust the estimated ELPI mass values to the measured filter mass ones. In a typical ELPI particle number distribution, the particle number of the first ELPI stage is usually quite important. However, the total mass corresponding to this first stage is very low because the particle size is very small. From the other hand, even if the particle number of the last ELPI stages is very low, the corresponding mass is important due to big particle size. Witze [7] used only the 4

six first ELPI stages to estimate the particle mass. The impact of ELPI stages number taken into account to estimate the particle mass is also examined in this work: the use of all ELPI stages, of only the six first stages, of all ELPI stages without the 1st one (11 last) and of all ELPI stages from the 2nd to the 9th one are compared. All these combinations give 120 cases. Initially, PM emissions of all three vehicles are used for these calculations. But, as the emissions of the Diesel PC are much higher than the PM emissions of the gasoline and Diesel+DPF PC, the Diesel PC and the low particle emitting PC are also examined separately. The value of (estimated values-measured values)2 is calculated and minimized in the case of parameters optimization. The line Estimated values=a*Measured values+b is calculated for each case. The values of , a, b and r2 of these lines are used to compare the 120 cases: both a and r2 must approach 1.0, while both b and must approach zero.

must be noted that this vehicle is representative of the current European fleet (Euro3).
4E+11

Diesel+DPF PC

150

100 2E+11
11

1.3 10

0 2E+14

Diesel PC

0 60

1E+14
1.5 10
11

30

0 2E+12

0 60

Gasoline PC

RESULTS AND DISCUSSION


1E+12 30
1.1 10
11

EMISSIONS OF REGULATED POLLUTANTS, CO2 AND FUEL CONSUMPTION


0E+0

0 Blank 1 2 3 4 Laboratory MV LOD RPRD RPT /LOQ

The other regulated pollutants, CO2 and fuel consumption are also measured, as indicators of good working conditions. These emissions and their repeatability and reproducibility values are presented in annexe1. A further analysis of these results is presented in another work [29]. The reproducibility values of these emissions are satisfactory, indicating that the particle and number emissions are obtained under good working conditions. The average PM emissions are 0.001 g/km for the gasoline passenger car, 0.0284 g/km for the Diesel PC and 0.0008 g/km for the Diesel PC equipped with DPF. The reproducibility value of these emissions is 67, 29 and 164% respectively for the three vehicles, while the corresponding repeatability value is 60, 12.7 and 191% [29]. The gasoline vehicle and the DPF equipped Diesel one have high repeatability and reproducibility values due to their very low emissions, which are similar to tunnel backgrounds [29]. PARTICLE NUMBER ON THE NEDC Figure 2 presents the particle number emitted from the three vehicles on the NEDC, the particle number of the tunnel background tests and the ELPI limits of detection and quantification (LOD, LOQ). The mean total particle number of the gasoline passenger car is 1.3x10121/km, the same order of magnitude as presented in literature [30]. The reproducibility and repeatability values are quite high, 59% and 45% respectively, due to low particle numbers [28, 29]. The mean total particle number of the Diesel PC is two orders of magnitude higher than the particle number of the gasoline PC: 1.3x10141/km [28-30]. The reproducibility and repeatability values are 47% and 14% respectively. It

Figure 2. Particle number (in 1/km) of the three passenger cars measured at each laboratory (mean of 3 tests), mean value (MV) of all tests, and ELPI limits of detection and of quantification (LOD/LOQ). Blank=mean of all four laboratories blanks for this vehicle. RPRD=reproducibility values, RPT=repeatability values. The mean total particle number of the DPF equipped Diesel vehicle is lower than the particle number of the two previous vehicles: 1.8x10111/km [28-30]. The reproducibility and repeatability values are 131 and 96% respectively. These last values are higher than those of the Diesel vehicle without DPF, due to lower particle numbers [29]. It must be noted that the particulate emissions of this vehicle are representative of the emissions of the future European passenger cars. The particle number of this vehicle is very close or even lower than the ELPI limits of detection or quantification, inducing low reproducibility, as in the case of mass measurements [29]. MEAN PARTICLE SIZE DISTRIBUTION Figure 3 presents the mean particle size distribution of the three vehicles. These distributions are quite similar for the four laboratories. Furthermore, size distribution shows that there is no nucleation during the measurements. The particle numbers emitted by the gasoline vehicle are one order of magnitude higher than particle numbers of tunnel blanks, up to about 300 nm (figure 4). For bigger particles, the ratio between the particles of exhaust gas and the tunnel background 5

Repeatability, Reproducibility (%)

50

Particle number (N, 1/km)

drops to only two.


1E+12 1E+11 1E+10 1E+9 1E+8
Blank Lab 1 Lab 2 Lab 3 Lab 4

the particle number of the tunnel blank measurements. However, this of the DPF equipped Diesel vehicle remains very close to the blank measurements for the entire distribution, especially in the area of the very fine particles. Figure 4 presents that the particle numbers of tunnel blanks, of the DPF equipped Diesel PC and of the upper part of the gasoline PC distribution are very close or even lower to LOD and LOQ. The reproducibility values of each ELPI stage are 63178, 20-161 and 129-208% respectively for the gasoline, Diesel and DPF equipped Diesel vehicles. COMPARISON BETWEEN THE PARTICLE MASS ESTIMATED BY ELSPI AND DETERMINED BY THE GRAVIMETRIC METHOD

Diesel DPF

Particle number (N, 1/km)

1E+7 1E+13 1E+12 1E+11 1E+10 1E+9 1E+8 1E+7 1E+12 1E+11 1E+10 1E+9 1E+8 1E+7 0.00 0.01 0.10 1.00 10.00

Diesel

The 120 cases presented in the experimental section are studied here using the values of all three passenger cars. Annexes 2-5 present the values of , a, b, and r2 of these 120 cases. The first part of this chapter examines the cases with constant, with the particle size, density values. The second one examines the cases where particle density varies with particle size. Constant particle density Figure 5 presents the estimated versus measured values using the particle density of 1.0 g/cm3. This value is initially studied as it is often presented in literature [14, 15]. This figure presents the results when different number of ELPI stages is taken into consideration using the lower diameter of each ELPI stage. The results of all ELPI stages those of the 11 last ELPI stages (when the 1st stage is not taken into account) are identical.
All stages

Gasoline

Dp (m)
Figure 3. Mean particle size distributions measured at each laboratory (mean of 3 tests).

Particle number/LOD

10000.0 1000.0 100.0 10.0 1.0 0.1 1000.0 100.0 10.0 1.0 0.1 0.00
LOQ

Blank Gasoline Diesel Diesel DPF

0.1

6 first stages 2-9 stages

Diesel vehicle

LOD
Gasoline Diesel Diesel DPF

Estimated mass (g/km)

WO 1st stage

1E-2

y=x

Exhaust/blank

1E-3

1E-4

0.01

0.10

1.00

10.00

Particle size (m)


Figure 4. Lower curves: ratio between the mean numbers of the particles emitted from each vehicle and this of the tunnel background tests. Upper curves: ratio between the mean particle numbers and the ELPI LOD. The particle number of the Diesel vehicle is, for the entire distribution, 2-3.5 orders of magnitude higher than

1E-5

1E-5

1E-4

1E-3

1E-2

0.1

Measured mass (g/km)


Figure 5. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different ELPI stages, using the particle density of 1 g/cm3. All 6

three vehicles used, for the lower diameter of each ELPI stage. The estimated mass using all ELPI stages overestimates the particulate mass compared to the measured values by 5.7 times in the case of the Diesel PC and 4.2 in the case of the low emitting PC. When the six first or the 2-9 ELPI stages are used, the Diesel particle mass is overestimated by 1.2 and 1.8 times respectively, but the particle mass of the low emitting vehicles is underestimated by 12.5 and 2.5 times respectively. As the mass estimated by ELPI using the density value of 1.0 g/cm3 do not correlate well with the mass collected on filters, other values are tested. Figure 6 presents the values of , a, b, and r2 for the four cases of constant, with the particle size, particle density. This figure shows that, generally, the particle density of 1.5, 1.0 and 0.8 g/cm3 cannot predict the measured particle mass, as the values of and a are very higher than 1.0 and the values of b and r2 are very higher than 0.0. If the particle density is simulated to obtain the same mass as that collected on filters, the , a, b and r2 values approach to 1.0 and 0.0. The r2 values are the same for the four different values of densities used.
100.000 10.000

density. ALL=all ELPI stages, 6: six first ELPI stages, 29: 2-9 ELPI stages, 11: all ELPI stages without the first one. The negative values of b are presented. Figure 7 presents the values of density obtained from these simulations. It is presented that the use of all ELPI stages or without the first one gives the same results. Fewer stages used more the density is high, due to lower particle numbers taken into consideration.
0.8

Estimated density (g/cm3)

0.6

0.4

0.2

0.0

Sum

1.000 0.100 0.010 0.001

ALL 6 2-9 11 ALL 6 2-9 11 ALL 6 2-9 11 LOWER INTERMEDIATE UPPER DIAMETER DIAMETER DIAMETER

r2

0.90

Figure 7. Simulated values of constant with the size particle density for the three particle diameters used. ALL=all ELPI stages, 6: six first ELPI stages, 2-9: 2-9 ELPI stages, 11: all ELPI stages without the first one. The particle diameter used also influences the obtained particle density, after the order lower>intermediate> upper, because higher particle density is necessary to adjust the smaller particle size when lower diameters are taken into account. It must be noticed that the best simulated particle density is lower than the initial values and, generally, quite low compared to the values presented in literature. Figure 8 presents the estimated versus measured values for the simulated particle densities using the lower diameter of each ELPI stage, when different ELPI stages are taken into account. The results of all ELPI stages and those of the 11 last ELPI stages (when the 1st stage is not taken into account) are identical. The estimated mass using all ELPI stages is closer to the y=x line than the other two cases. The estimated mass values of the Diesel vehicle are quite close to the y=x line for the four configurations, but the points of the low particle emitting vehicles are quite dispersed. Figure 9 presents the estimated versus measured values for the simulated particle densities, using all ELPI 7

d=1.5 g/cm3
0.80

d=1.0 g/cm3 d=0.8 cm3

0.100

d=simulated

-b a

0.010

0.001 100

10

ALL 6 2-9 11 ALL 6 2-9 11 ALL 6 2-9 11 LOWER INTERMEDIATE UPPER DIAMETER DIAMETER DIAMETER

Figure 6. Values of , a, b, and r2 for constant particle

stages, when different particle diameters are taken into account. The results of the lower diameter are closer to the y=x line, while the results of the intermediate diameter comes next, before these of the upper one. Once again, the estimated mass of the Diesel vehicle are quite close to the y=x line for the three diameters, but the points of the low particle emitting vehicles are quite dispersed.
0.1
All stages 6 first stages 2-9 stages Diesel vehicle

particle density and all ELPI stages. All three vehicles used. Particle density as a function of particle diameter Figure 10 presents the estimated versus measured values when the particle density proposed by Ahlvik et al [3] is used. The lower diameter of each ELPI stage is used when different ELPI stages are taken into account. The results of all ELPI stages and the results of the 11 last ELPI stages (when the 1st stage is not taken into account) are identical and not presented here.
All stages

Estimated mass (g/km)

1E-2

WO 1st stage y=x

0.1

Diesel vehicle

6 first stages 2-9 stages

Estimated mass (g/km)

1E-3

y=x

1E-2

1E-4

1E-3

1E-5

1E-4

1E-5

1E-4

1E-3

1E-2

0.1

Measured mass (g/km)


Figure 8. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different ELPI stages, using the constant simulated particle density. All three vehicles used, for the lower diameter of each ELPI stage.
0.1
Lower Intermediate Upper Diesel vehicle

1E-5

1E-5

1E-4

1E-3

1E-2

0.1

Measured mass (g/km)


Figure 10. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different ELPI stages, using the particle density of Ahlvik et al [3]. All three vehicles used, for the lower diameter of each ELPI stage. The estimated mass of the Diesel vehicle are overestimated by 1.8, 1.1 and 1.34 times when all, the 6 first and the 2-9 ELPI stages are used. These values are lower than the results of the constant density of 1.0 g/cm3. However, the estimated particle mass of the low particle emitting vehicles are underestimated by 1.7, 12.5 and 5 times in the case of all, 6 first and 2-9 ELPI stages respectively. As the mass estimated by ELPI using the density values proposed by Ahlvik et al [3] does not give a good correlation with the mass collected on filters, other values are tested. Figure 11 presents the values of , a, b, and r2 for the cases where the particle density is a function of particle size. This figure shows that, generally, the simulated particle densities, used to obtain the same ELPI mass as the mass obtained on filters, give better results than the three non-simulated correlations. 8

Estimated mass (g/km)

y=x

1E-2

1E-3

1E-4 1E-4 1E-3 1E-2 0.1

Measured mass (g/km)


Figure 9. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different diameters of particles, using the constant simulated

Particle density (g/cm3)

The simulated values of density, as a function of particle diameter using the lower diameter of each ELPI stage, are presented in figure 12. As previously, the use of all ELPI stages or only the 11 last ones gives the same results. For this reason, the values of the latter case are not presented in this figure. As in the case of constant, with the size, particle diameter, fewer stages used more the density is high, due to the lower particle numbers taken into consideration. The particle diameter also influences the obtained particle density, according to the order lower>intermediate>upper. As in the case of constant density, the best simulated particle density is lower than the initial values and lower than the density values presented in literature.
10.000 1.000

These three figures present that the estimated mass using all ELPI stages are closer to the y=x line than the other two cases. The estimated mass of the Diesel vehicle are quite close to the y=x line, but the points of the low particle emitting vehicles present quite important dispersions.
2.0
Original All stages 6 first stages

1.0

2-9 stages

Ahlvik 1998
0.0

Linear log
1.0

Sum

0.100 0.010 0.001 0.000

0.5

0.0

Linear
1.0

r2

0.90

0.5
d=linear d=linear sim.

0.0 0.01 0.10 1.00 10.00

0.80 0.100

d=linear log d=linear log sim. d=Ahlvik

Diameter (m)
Figure 12. Variable particle density used. Original and simulated values for all ELPI stages, the 6 first stages and the 2-9 ELPI stages.
0.1
All stages 6 first stages Diesel vehicle

-b

d=Ahlvik sim.
0.010

0.001

10

2-9 stages

Estimated mass (g/km)

1E-2

y=x

1E-3

ALL 6 2-9 11 ALL 6 2-9 11 ALL 6 2-9 11 LOWER INTERMEDIATE UPPER DIAMETER DIAMETER DIAMETER

1E-4

Figure 11. Values of , a, b, and r for the variable particle density. ALL=all ELPI stages, 6: six first ELPI stages, 2-9: 2-9 ELPI stages, 11: all ELPI stages without the first one.
2

1E-5

Figures 13-15 present the estimated versus measured values for the simulated particle densities using the lower diameter of each ELPI stage for the different ELPI stages taken into account, for the linear, linear log and Ahvlik densities. As the results of all ELPI stages and these of the 11 last ones are identical, the points of the latter case are not presented in these figures.

1E-5

1E-4

1E-3

1E-2

0.1

Measured mass (g/km)


Figure 13. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different ELPI stages, using the variable simulated linear particle 9

density. All three vehicles used, for the lower diameter of each ELPI stage. Figures 16-18 present the estimated versus measured values for the simulated particle densities using all ELPI stages for the different diameters taken into account, for the linear, linear log and Ahvlik densities. The results of the lower diameter are closer to the y=x line, followed by the results of the intermediate diameter, while those of the upper one come next. The use of intermediate and upper diameter overestimates several times the particle mass when the linear log and Ahvlik densities are used.
0.1
All stages 6 first stages 2-9 stages Diesel vehicle

Figure 15. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different ELPI stages, using the variable simulated particle density used by Ahlvik et al [3]. All three vehicles used, for the lower diameter of each ELPI stage.
1
Lower Intermediate Upper Diesel vehicle

Estimated mass (g/km)

0.1

y=x

1E-2

Estimated mass (g/km)

1E-2

y=x

1E-3

1E-3

1E-4
1E-4

1E-4

1E-3

1E-2

0.1

Measured mass (g/km)


1E-5

1E-5

1E-4

1E-3

1E-2

0.1

Measured mass (g/km)


Figure 14. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different ELPI stages, using the variable simulated linear log particle density. All three vehicles used, for the lower diameter of each ELPI stage.
0.1
All stages 6 first stages 2-9 stages Diesel vehicle

Figure 16. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different diameters of particles, using the variable linear simulated particle density and all ELPI stages. All three PC used.
1
Lower Intermediate Upper Diesel vehicle

Estimated mass (g/km)

0.1

y=x

1E-2

Estimated mass (g/km)

1E-2

y=x

1E-3

1E-3

1E-4
1E-4

1E-4

1E-3

1E-2

0.1

Measured mass (g/km)


1E-5

1E-5

1E-4

1E-3

1E-2

0.1

Measured mass (g/km)

Figure 17. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for different diameters of particles, using the variable linear log simulated particle density and all ELPI stages, for the three PC. 10

Once again, in figures 16-18, the estimated mass values of the Diesel vehicle are quite close to the y=x line for the three types of density distribution (linear, linear log and Ahlvik), but the points of the low particle emitting vehicles are quite dispersed.
0.1
Lower Intermediate Upper Diesel vehicle

all ELPI stages and the lower diameter in the case of the best simulated, variable with the size, particle density. The two figures show similar results: the mass values estimated by ELPI are generally within the same area than the mass determination using filters in the case of mass values higher than 0.025 g/km, while the lower values are underestimated.
0.06
All stages sim. 6 first stages sim. 2-9 stages sim.

Estimated mass (g/km)

y=x

1E-2

Estimated mass (g/km)

y=x

0.04

1E-3

0.02

1E-4

1E-4

1E-3

1E-2

0.1
0.00 0.00 0.02 0.04 0.06

Measured mass (g/km)


Figure 18. Estimated versus measured values of particle mass (and the best fitted lines y=ax+b), for the different diameters of particles, using the variable simulated particle density used by Ahlvik et al [3], and all ELPI stages. All three vehicles used. COMPARISON BETWEEN THE PARTICLE MASS ESTIMATED BY ELSPI AND DETERMINED BY THE GRAVIMETRIC METHOD OF THE DIESEL VEHICLE WITHOUT DPF In the above figures, the mass estimated by ELPI is close to the line y=x in the case of Diesel passenger without DPF. In this chapter the particle mass estimated by ELPI is focused on this vehicle. Figure 19 presents the estimated versus measured values of particle mass for the different ELPI stages taken into consideration, for the lower diameter and the best simulated constant particle density. The error bars of mass determination using filters and ELPI are added. The 1.96*RSD (relative standard deviation) reproducibility values of these measurements are 29% in the case of mass measurements using filters and 47% in the case of ELPI particle number determination [29]. Two groups of points are formed: one with measured PM mass lower than 0.025 g/km and one with measured PM mass higher than 0.025 g/km. The higher mass values are found quite close to the y=x line; moreover the error bars completely cover it. The lower values are lower than the y=x line and the error bars are just in touch with it. Figure 20 presents the same results using

Measured mass (g/km)


Figure 19. Estimated versus measured values of particle mass, for the different ELPI stages, using the lower diameter of particles and the best constant simulated particle density. Diesel passenger car without DPF.
0.06
Linear sim. Linear Log sim. Ahlvik 1998 sim.

Estimated mass (g/km)

y=x

0.04

0.02

0.00 0.00 0.02 0.04 0.06

Measured mass (g/km)


Figure 20. Estimated versus measured values of particle mass, using all ELPI stages and the lower diameter of particles, for the best variable simulated particle density. Diesel passenger car without PDF. 11

COMPARISON BETWEEN THE PARTICLE MASS ESTIMATED BY ELSPI AND DETERMINED BY THE GRAVIMETRIC METHOD OF THE LOW PARTICLE EMITTING VEHICLES Figure 21 presents the estimated versus measured values of particle mass for the different ELPI stages taken into consideration, for the lower diameter and the best simulated constant particle density, in the case of the low particle emitting passenger cars (gasoline and Diesel+DPF PC). The error bars of mass determination using filters and ELPI are added. The 1.96*RSD (relative standard deviation) reproducibility values of mass measurements are 67% and 164% respectively for the gasoline and Diesel+DPF passenger cars, while the corresponding values for the ELPI number determination are 59 and 131% respectively [29]. Figure 22 presents the same results using all ELPI stages and the lower diameter, in the case of the best simulated variable particle density.
0.002
All stages sim. 6 first stages sim. 2-9 stages sim.

These results suggest that the mass estimated by ELPI is not correlated with the mass collected on filters at these low levels of PM emissions, when particle density is optimized for all three vehicles (PM emissions from 0.0001 to 0.034 g/km). A better optimization in searched in the case of these vehicles using density values specifically adapted for these PM emission levels, without taking into account the emissions of the Diesel PC. Figure 22 presents that even in this case, the estimated ELPI mass values are not correlated with the filter determined mass values.
0.002
Linear sim. Linear log sim. Ahlvik 1998 sim.

Estimated mass (g/km)

0.002

y=x

0.001

0.001

Estimated mass (g/km)

0.002

y=x

0.000

0.001

0.000 0.000
0.001

0.002

0.004

0.006

0.008

Measured mass (g/km)


Figure 22. Estimated versus measured values of particle mass, using all ELPI stages and the lower diameter of particles, for the variable best simulated particle density. Results of the gasoline and Diesel passenger car equipped with DPF.
0.000 0.002 0.004 0.006 0.008

0.000

0.000

0.003

Measured mass (g/km)

Constant sim. Linear sim. Linear log sim.

Estimated mass (g/km)

Figure 21. Estimated versus measured values of particle mass, for the different ELPI stages, using the lower diameter of particles and the constant best simulated particle density. Results of the gasoline and Diesel passenger car equipped with DPF. Figures 21 and 22 present the same cases as the Diesel passenger car without DPF. However, they present that the estimated values of particle mass are very lower than the y=x line, indicating that, for mass measurements bellow 0.0025 g/km, ELPI underestimates the particle mass determined by filters (using the above values of particle diameter and density). Even if the ELPI error bars are very large due to the low repeatability of this method at these particle number levels [29], the line y=x is not always covered by the error bars. It must be noticed that the repeatability values of mass measurements at these levels are also very high.

Ahlvik sim. y=x

0.002

0.001

0.000 0.000 0.001 0.002 0.003

Measured mass (g/km)


Figure 23. Estimated versus measured values of particle 12

mass. All ELPI stages, lower diameter, for different best simulated particle densities specifically adapted for the low particle emitting vehicles (gasoline and Diesel passenger car equipped with DPF).

particle distribution estimated by ELPI. For this reason, a program correlating ELPI with several other devices measuring particle number and size distribution will be undertaken.

CONCLUSIONS
ELPI is used to estimate the exhaust particle mass of three Euro3 passenger cars. One gasoline PC operating under stoichiometric conditions, one Diesel PC and one DPF equipped Diesel PC are tested on the NEW European Driving Cycle. Four laboratories participated ton this work: IFP, PSA Peugeot-Citron, Renault and UTAC. The results of this study show that: - The average PM emissions are 0.001 g/km for the gasoline passenger car, 0.0284 g/km for the Diesel PC and 0.0008 g/km for the Diesel PC equipped with DPF. The reproducibility value of these emissions is 67%, 29% and 164% respectively for the three vehicles, while the corresponding repeatability value is 60%, 12.7% and 191%. - The average particle number is 1.3x10121/km for the gasoline passenger car, 1.3x10141/km for the Diesel PC and 1.8x10111/km for the Diesel PC equipped with DPF. The reproducibility value of these emissions is 59%, 47% and 131% respectively for the three vehicles, while the corresponding repeatability value is 45%, 14% and 96%. - The use of the lower diameter of each ELPI stage gives better results than the use of intermediate or upper particle diameter. - The use of particle number of all ELPI stages gives the same results as the 11 last stages and better than the particle number of only the six first stages, or the 2-9 stages. - A particle density of 1.0 g/cm3 cannot be used to estimate the particle mass. A variable with the size density also gives moderate results. A simulated value gives better results, but with particle densities very lower than those presented in literature. - The estimated by ELPI mass values of the Diesel vehicle are quite close to measured ones, for particle mass higher than 0.025 g/km. Below this value, the estimated values are lower than the measured ones. - The estimated mass values of the low particle emitting passenger cars (gasoline and Diesel+DPF) are very scatted and not correlated with the particle mass collected on filters even if the particle density is simulated excluding the values of the Diesel PC. - The four laboratories continue to work in this field. According to the results presented here, two main work fields are identified; they concern particle number and

REFERENCES
1. Directive 70/220, www.europa.eu.int 2. Keskinen J., Pietarinen K., Lehtimaki M. (1992), Electrical Low Pressure Impactor, J. Aerosol Sci., 23, 353-360. 3. Ahlvik P., Ntziachristos L., Keskinen J., Virtanen A. (1998), Real time measurements of Diesel particle size distribution with an electrical low-pressure impactor, SAE Technical Paper Series 980410. 4. Pattas K., Kyriakis N., Samaras Z., Pistikopoulos P., Ntziachristos L. (1998), Effect of DPF on particulate size distribution using an electrical low pressure impactor, SAE Technical Paper Series 980544. 5. Khalek I.A. (2000), Characterization of particle size distribution of a heavy-duty Diesel engine during FTP transient cycle using ELPI, SAE Technical Paper Series 2000-01-2001. 6. Maricq M.M., Podsiadlik D.H., Chase R.E., (2000), Size distributions of motor vehicle exhaust PM: a comparison between ELPI and SMPS measurements, Aerosol Sci. Technol., 33, 239-260. 7. Witze P.O., Chase R.E., Maricq M.M., Podsiadlik D.H., Xu N. (2004), Time-resolved merasurement of exhaust PM for FTP-75: comparison of LII, ELPI and TEOM techniques, SAE Technical Paper Series 2004-01-0964. 8. Wang S.C., Flagan R.C. (1990), Scanning Electrical Mobility Spectrometer, Aerosol Sci. Technol. 13, 230-240. 9. Willeke K., Baron P.A. (1993), Aerosol Measurement, Principles, Techniques and Applications, Van Nostrand Reinhold, New York. 10. Hinds W.C. (1999), Aerosol technology, J. Willey and Sons, New York. 11. Burtscher H. (1991), Literature study on tailpipe particulate emission measurement for Diesel engines, report for the PMP, www.akpf.org/pub/burtscher_bericht.pdf. 12. Mohr M., Lehmann U. (2003), Comparison study of particle measurement systems for future type approval application, RESEARCH REPPORT NO 202779, Eidgenssiche Materialprfungs und Forschungsanstalt (EMPA), Dbendorf Switzerland, Mai 2003, http://www.empa.ch/plugin/ template/empa/*/20987/---/l=1. 13. Marjmaki M., Keskinen J., Chen D.R., Pui D.Y.H. (2000), Performance evaluation of the electrical low pressure impactor (ELPI), J. Aerosol Sci., 31, 249261. 14. Shi J.P., Harrison R.M., Brear F. (1999), Particle size distribution from a modern heavy duty Diesel engine, Sci. Total Environ., 235, 305-307. 15. Tsukamoto Y., Goto Y., Odaka M. (1990), 13

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

Continuous measurement of Diesel particulate emissions by an electrical low-pressure impactor, SAE Technical Paper Series 2000-01-1138. Andrews G.E., Clarke A.G., Rojas N.Y., Sale T., Gregory D. (2001), Diesel particle size distribution: the conversion of particle number size distribution to mass distribution. Virtanen A., Ristimki J., Marjamaki M., Vaaraslahti K., Keskinen J. (2002), Effective density of Diesel exhaust particles as a function of size, SAE Technical paper series 2002-01-0056. Ristimki J., Virtanen A., Marjamki M., Rostedt A., Keskinen J. (2002), On-line measurement of size distribution and effective density of submicron aerosol particles, J. Aerosol Sci., 33, 1541-1557. Choi B.C., Jang S.H., Juhng W.N. (2003), ELPI measurement of particulate matter from HSDI Diesel engine with Diesel oxidation catalyst, SAE Technical Paper Series 2003-01-3159. Maricq M.M., Xu N. (2004), The effective density and fractal dimension of soot particles from premixed flames and motor vehicle exhaust, Aerosol Sci., 35, 1251-1274. Van Gulijk C., Marijnissen J.C.M., Makkee M., Moulijn J.A. (2001), Restriction for the ELPI in Diesel particulate measurements, J. Aerosol Sci. 32, 11171130. Van Gulijk C., Marijnissen J.C.M., Makkee M., Moulijn J.A. (2003), The choice of instrument (ELPI and/or SMPS) for Diesel soot particulate measurements, SAE Technical Paper Series 200301-0784. Van Gulijk C., Marijnissen J.C.M., Makkee M., Moulijn J.A. Schmidt-Ott A. (2004), Measuring diesel soot with a scanning mobility particle sizer and an electrical low-pressure impactor: performance assessment with a model for fractal like agglomerates, J. Aerosol Sci., 35, 633-655. ACEA programme on emissions of fine particles from passenger cars, Report, December 1999, www.acea.be. ACEA programme on emissions of fine particles from passenger cars 2, Report, July 2002, www.acea.be. Ulfvarson U., Figler B., Krantz S. (1997), Diesel engien development is guided by inadequate particle sampling, SAE Technical Paper Series 970759. OICA (2003), Industry comments on the proposed particulate measurement techniques. OICA contribution to PMP, http://www.unece.org/trans/main/wp29/wp29wgs/wp 29grpe/grpeinf45.html. Zervas E., Dorlhne P., Daviau R., Dionnet B. (2004), Repeatability of fine particle measurement of Diesel and gasoline vehicles exhaust gas, SAE Technical Paper Series 2004-01-1983. Zervas E., Dorlhne P., Forti L., Perrin C., Momique J.C., Monier R., Ing H., Lopez B. (2005), - Interlaboratory Test of Exhaust PM Using ELPI Aerosol

Sci. Technol., in Press. 30. Mohr M., Lehmann U., Margaria G. (2003a), ACEA programme on the emissions of fine particulates from passenger cars (2). Part 2. Effect of Sampling conditions and fuel sulphur content on the particle emission, SAE Technical Paper Series 2003-011890.

CONTACT
Dr. Efthimios Zervas Renault 1, Alle Cornuel, Fr - 91510 Lardy Email : efthimios.zervas@renault.com Tel : +331-69 27 84 77

14

ANNEXE 1: Average, for each laboratory, emission of regulated pollutants, CO2 (in g/km), repeatability (RPT) and reproducibility (RPD) 1.96*RSD values (%) for the three vehicles used. G: gasoline vehicle, D: diesel vehicle, DPF: DPF equipped diesel vehicle. CO (g/km) Lab1 Lab2 Lab3 Lab4 MEAN RPT (%) RPD (%) G 0.67 0.73 0.67 0.73 0.7 11.2 13.1 D 0.2 0.24 0.23 0.22 0.22 20.4 21.3 DPF 0.12 0.28 0.12 0.1 0.16 37.6 113.1 G 0.116 0.147 0.122 0.112 0.124 14.2 27.2 HC D 0.034 0.032 0.035 0.029 0.032 18.3 22.6 DPF 0.033 0.054 0.031 0.028 0.036 27.5 67.2 G 0.045 0.037 0.04 0.051 0.043 45.9 47.2 NOx D 0.333 0.344 0.319 0.36 0.339 4 10.9 DPF 0.254 0.295 0.279 0.307 0.284 6.4 16.7 G 0.000 8 0.001 1 0.001 60.4 67.8 PM D 0.023 3 0.027 5 0.031 3 0.031 3 0.028 4 12.9 29.3 DPF 0.000 6 0.000 7 0.001 0.001 0.000 8 191 163.9 G 156.4 155.8 159.6 161.7 158.4 1.2 3.6 CO2 D 137.1 137.9 136.1 137.6 137.2 1.6 1.7 D+DP F 136.5 135.9 135.9 139.8 137 1.2 2.9 G 6.64 6.59 6.7 6.81 6.68 1.2 2.9 FC D 5.2 5.23 5.23 5.22 5.22 1.6 1.4 DP F 5.14 5.09 5.18 5.29 5.17 1.2 3.3

ANNEX 2 The values of (estimated values-measured values)2 for all cases studied, using all three passenger cars tested. constant simulated 0.0005 0.0007 0.0004 0.0005 0.0007 0.0006 0.0004 0.0007 0.001 0.0005 0.0004 0.001 linear, simulated 0.0004 0.0007 0.0004 0.0004 0.0318 0.0006 0.0004 0.0318 0.5392 0.0005 0.0006 0.4886 linear log 0.028 0.001 0.002 0.028 0.331 0.01 0.026 0.331 3.303 0.084 0.174 3.295 linear log, simulated 0.0004 0.0006 0.0004 0.0004 0.0275 0.0005 0.0004 0.0275 0.1918 0.0005 0.0006 0.1993 Ahlvik simulated 0.0003 0.0004 0.0003 0.0003 0.0004 0.0003 0.0003 0.0004 0.0037 0.0003 0.0002 0.0037

LOWER DIAMETER

INTERMEDIATE DIAMETER

UPPER DIAMETER

All stages 6 first stages 2-9 stages Without 1st stage All stages 6 first stages 2-9 stages Without 1st stage All stages 6 first stages 2-9 stages Without 1st stage

d=1.5 0.62 0.009 0.031 0.619 5.81 0.075 0.197 5.808 60.73 0.403 0.967 60.68

d=1.0 0.244 0.002 0.008 0.244 2.486 0.023 0.07 2.485 26.68 0.154 0.39 26.66

d=0.8 0.142 0.001 0.003 0.142 1.545 0.011 0.038 1.545 16.93 0.088 0.232 16.91

linear 0.082 0.003 0.012 0.082 0.657 0.037 0.096 0.657 5.203 0.224 0.508 5.192

Ahlvik 0.006 0.001 0.002 0.006 0.206 0.017 0.031 0.205 0.379 0.134 0.209 0.375

15

ANNEXE 3. The values of a of the line estimated values=a*measured values+b, for all cases studied, using all three passenger cars tested. constant, simulation. 0.98 0.97 0.99 0.98 0.95 0.97 0.99 0.95 0.92 0.98 0.99 0.92 linear, simulation 1.01 0.97 0.99 1.01 2.81 0.97 0.99 2.81 8.50 0.98 1.14 8.10 linear log 2.77 1.01 1.35 2.77 7.05 2.02 2.70 7.05 19.91 4.08 5.43 19.89 linear log, simulation 1.01 0.98 0.99 1.01 2.66 0.98 0.99 2.66 5.41 0.99 1.14 5.51 Ahlvik simulation 1.00 0.99 1.00 1.00 0.99 1.00 1.00 0.99 1.56 1.01 1.01 1.56

LOWER DIAMETER

INTERMEDIATE DIAMETER

UPPER DIAMETER

All stages 6 first stages 2-9 stages Without 1st stage All stages 6 first stages 2-9 stages Without 1st stage All stages 6 first stages 2-9 stages Without 1st stage

1.5 9.29 1.94 2.86 9.29 26.10 3.85 5.70 26.09 81.06 7.71 11.45 81.02

1.0 6.19 1.29 1.90 6.19 17.40 2.57 3.80 17.40 54.04 5.14 7.63 54.01

0.8 4.95 1.04 1.52 4.95 13.92 2.05 3.04 13.92 43.23 4.11 6.10 43.21

linear 4.05 1.51 2.14 4.05 9.60 3.00 4.27 9.59 25.00 6.01 8.57 24.97

Ahlvik 1.84 1.18 1.43 1.84 5.81 2.39 2.89 5.81 7.59 4.92 5.90 7.55

ANNEXE 4. The values of b of the line estimated values=a*measured values+b, for all cases studied, using all three passenger cars tested. constant , simulati on. -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 linear log, simulati on -0.001 -0.001 -0.001 -0.001 -0.003 -0.001 -0.001 -0.003 -0.006 -0.001 -0.001

LOWER DIAMETER

INTERMEDIATE DIAMETER

UPPER DIAMETER

All stages 6 first stages 2-9 stages Without 1st stage All stages 6 first stages 2-9 stages Without 1st stage All stages 6 first stages 2-9 stages

1.5 -0.008 -0.002 -0.003 -0.008 -0.026 -0.005 -0.005 -0.026 -0.087 -0.009 -0.011

1.0 -0.006 -0.002 -0.002 -0.006 -0.017 -0.003 -0.004 -0.017 -0.058 -0.006 -0.007

0.8 -0.004 -0.001 -0.001 -0.004 -0.014 -0.002 -0.003 -0.014 -0.047 -0.005 -0.006

linear -0.003 -0.002 -0.002 -0.003 -0.008 -0.004 -0.004 -0.008 -0.024 -0.007 -0.008

linear, simulati on -0.001 -0.001 -0.001 -0.001 -0.003 -0.001 -0.001 -0.003 -0.009 -0.001 -0.001

linear log -0.002 -0.001 -0.001 -0.002 -0.007 -0.002 -0.003 -0.007 -0.021 -0.004 -0.005

Ahlvik -0.001 -0.001 -0.001 -0.001 -0.006 -0.002 -0.003 -0.006 -0.006 -0.005 -0.005

Ahlvik simulati on -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.002 -0.002 -0.001 -0.001 16

Without 1st stage

-0.087

-0.058

-0.047

-0.001

-0.024

-0.009

-0.021

-0.006

-0.006

-0.002

ANNEXE 5. The values of r2 of the line estimated values=a*measured values+b, for all cases studied, using all three passenger cars tested. constant, simulation. 0.92 0.89 0.93 0.92 0.88 0.9 0.93 0.88 0.83 0.91 0.94 0.83 linear, simulation 0.93 0.89 0.93 0.93 0.89 0.9 0.93 0.89 0.84 0.91 0.94 0.83 linear log 0.94 0.9 0.93 0.94 0.91 0.91 0.94 0.91 0.87 0.93 0.94 0.87 linear log, simulation 0.93 0.9 0.93 0.93 0.88 0.91 0.94 0.88 0.83 0.92 0.94 0.84 Ahlvik simulation 0.95 0.93 0.94 0.95 0.94 0.94 0.95 0.93 0.9 0.96 0.96 0.9

LOWER DIAMETER

INTERMEDIATE DIAMETER

UPPER DIAMETER

All stages 6 first stages 2-9 stages Without 1st stage All stages 6 first stages 2-9 stages Without 1st stage All stages 6 first stages 2-9 stages Without 1st stage

1.5 0.92 0.89 0.93 0.92 0.88 0.9 0.93 0.88 0.83 0.91 0.94 0.83

1.0 0.92 0.89 0.93 0.92 0.88 0.9 0.93 0.88 0.83 0.91 0.94 0.83

0.8 0.92 0.89 0.93 0.92 0.88 0.9 0.93 0.88 0.83 0.91 0.94 0.83

linear 0.95 0.89 0.93 0.95 0.94 0.9 0.93 0.94 0.91 0.91 0.94 0.91

Ahlvik 0.96 0.93 0.94 0.96 0.94 0.94 0.95 0.93 0.97 0.96 0.96 0.97

17

Potrebbero piacerti anche