+ All Categories
Home > Documents > Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected...

Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected...

Date post: 25-Dec-2016
Category:
Upload: peipei
View: 215 times
Download: 1 times
Share this document with a friend
10
Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels: Gasoline and M15 methanol gasoline Bin Liang a,b , Yunshan Ge a,* , Jianwei Tan a , Xiukun Han a , Liping Gao a , Lijun Hao a , Wentao Ye a , Peipei Dai a a National Lab of Auto Performance and Emission Test, Beijing Institute of Technology, Beijing 100081, China b Beijing Automotive Research Institute Co., Ltd., Beijing 100079, China article info Article history: Received 23 July 2012 Received in revised form 9 November 2012 Accepted 13 November 2012 Available online 30 November 2012 Keywords: Particulate matter (PM) M15 methanol gasoline Gasoline direct injection (GDI) Mass estimation Number distribution Electrical low pressure impactor (ELPI) abstract Two Euro 4 gasoline passenger vehicles (one gasoline direct injected vehicle and one port fuel injected vehicle) were tested over the cold start New European Driving Cycle (NEDC). Each vehicle was respectively fueled with gasoline and M15 methanol gasoline. Particle number concentrations were measured by the electrical low pressure impactor (ELPI). Particle masses were measured by gravimetric method and estimated from the number distributions using two density distributions (one is constant with the particle size and one is power law related with the size). The first 7 stages of ELPI were used for estimation. The results show that for each vehicle, PM masses measured by gravimetric method, the total PM numbers measured by ELPI and estimated PM masses for M15 are lower than those for gasoline. For each kind of fuel, PM masses by two methods and total PM numbers from the GDI vehicle are higher than those from the PFI one. PM number distribution curves of the four vehicle/fuel combinations are similar. All decline gradually and the maximum number of each curve occurs in the first stage. More than 99.9% numbers locate in the first 8 stages of which diameters are less than 1 mm. PM number emissions correlate well with the acceleration of the two vehicles. The estimated particle masses were much lower than the gravimetric measurements. & 2012 Elsevier Ltd. All rights reserved. 1. Introduction Research on alternative fuels around the world has been widely carried out for reasons like continually rising demands for present energy along with population growing and science and technology development, increasing fuel prices, worries about the exhausting of petroleum and requirements to guarantee energy security (Abu-Zaid et al., 2004; Bilgin & Sezer, 2008). Methanol, also named as methyl alcohol, has received great attention for its lower price and easier production from a lot of resources such as natural gas, coal, combustible trash, city waste and wood (Bilgin & Sezer, 2008; Wei et al., 2008). It is very important to produce methanol from coal which is abundant in the world for a relatively long time, especially for China. In China, taking methanol as an alternative fuel has two realistic reasons. Firstly, the structure of Chinese energy resources is short Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/jaerosci Journal of Aerosol Science 0021-8502/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jaerosci.2012.11.008 * Corresponding author. Tel.: þ86 10 6891 2035; fax: þ86 10 6894 8486. E-mail addresses: [email protected] (B. Liang), [email protected] (Y. Ge). Journal of Aerosol Science 57 (2013) 22–31
Transcript
Page 1: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

Contents lists available at SciVerse ScienceDirect

Journal of Aerosol Science

Journal of Aerosol Science 57 (2013) 22–31

0021-85

http://d

* Corr

E-m

journal homepage: www.elsevier.com/locate/jaerosci

Comparison of PM emissions from a gasoline direct injected(GDI) vehicle and a port fuel injected (PFI) vehicle measuredby electrical low pressure impactor (ELPI) with twofuels: Gasoline and M15 methanol gasoline

Bin Liang a,b, Yunshan Ge a,*, Jianwei Tan a, Xiukun Han a, Liping Gao a,Lijun Hao a, Wentao Ye a, Peipei Dai a

a National Lab of Auto Performance and Emission Test, Beijing Institute of Technology, Beijing 100081, Chinab Beijing Automotive Research Institute Co., Ltd., Beijing 100079, China

a r t i c l e i n f o

Article history:

Received 23 July 2012

Received in revised form

9 November 2012

Accepted 13 November 2012Available online 30 November 2012

Keywords:

Particulate matter (PM)

M15 methanol gasoline

Gasoline direct injection (GDI)

Mass estimation

Number distribution

Electrical low pressure impactor (ELPI)

02/$ - see front matter & 2012 Elsevier Ltd.

x.doi.org/10.1016/j.jaerosci.2012.11.008

esponding author. Tel.: þ86 10 6891 2035;

ail addresses: [email protected] (B. L

a b s t r a c t

Two Euro 4 gasoline passenger vehicles (one gasoline direct injected vehicle and one port

fuel injected vehicle) were tested over the cold start New European Driving Cycle (NEDC).

Each vehicle was respectively fueled with gasoline and M15 methanol gasoline. Particle

number concentrations were measured by the electrical low pressure impactor (ELPI).

Particle masses were measured by gravimetric method and estimated from the number

distributions using two density distributions (one is constant with the particle size and

one is power law related with the size). The first 7 stages of ELPI were used for

estimation. The results show that for each vehicle, PM masses measured by gravimetric

method, the total PM numbers measured by ELPI and estimated PM masses for M15 are

lower than those for gasoline. For each kind of fuel, PM masses by two methods and total

PM numbers from the GDI vehicle are higher than those from the PFI one. PM number

distribution curves of the four vehicle/fuel combinations are similar. All decline gradually

and the maximum number of each curve occurs in the first stage. More than 99.9%

numbers locate in the first 8 stages of which diameters are less than 1 mm. PM number

emissions correlate well with the acceleration of the two vehicles. The estimated particle

masses were much lower than the gravimetric measurements.

& 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Research on alternative fuels around the world has been widely carried out for reasons like continually rising demandsfor present energy along with population growing and science and technology development, increasing fuel prices, worriesabout the exhausting of petroleum and requirements to guarantee energy security (Abu-Zaid et al., 2004; Bilgin & Sezer, 2008).Methanol, also named as methyl alcohol, has received great attention for its lower price and easier production from a lot ofresources such as natural gas, coal, combustible trash, city waste and wood (Bilgin & Sezer, 2008; Wei et al., 2008). It is veryimportant to produce methanol from coal which is abundant in the world for a relatively long time, especially for China. InChina, taking methanol as an alternative fuel has two realistic reasons. Firstly, the structure of Chinese energy resources is short

All rights reserved.

fax: þ86 10 6894 8486.

iang), [email protected] (Y. Ge).

Page 2: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–31 23

of fuel and gas but rich of coal. Secondly, grains are not sufficient to produce ethanol for automobile use (Bilgin & Sezer, 2008;Wei et al., 2008; Zhang et al., 2010).

Methanol also has some technical advantages as an alternative or supplementary fuel since it can be burned purely orin a mixture with gasoline at different ratios (Wei et al., 2008). Due to higher octane number, lower boiling point, lowercarbon/hydrogen ratio and higher oxygen content over gasoline, engines with methanol can work on higher compressionratio and emit less carbon monoxide (CO) and hydrocarbon (HC) (Bilgin & Sezer, 2008; Wei et al., 2008; Zhao et al., 2010).Additionally, methanol has a higher latent heat of evaporation which can cooling the intake air and thus help on increasingthe volumetric efficiency and power output of engines (Bilgin & Sezer, 2008; Zhao et al., 2010). For the above reasons,methanol has aroused great focus as a promising alternative fuel (Bilgin & Sezer, 2008; Liaoet al., 2006; Zhao et al., 2010).In 2009, China issued two standards which are GB/T 23799-2009 methanol gasoline (M85) for motor vehicles and GB/T23510-2009 fuel methanol for motor vehicles. This accelerates the research on methanol gasoline in China.

However, there are two disadvantages need to be mentioned for taking methanol as a vehicle fuel. Firstly, emissions offormaldehyde will increase along with the increase of methanol content in gasoline. Formaldehyde is carcinogenic and hasgreat adverse effects on human health (Zhao et al., 2010). Secondly, burning methanol derived from coal contribute tomore carbon dioxide (CO2) emission than gasoline. The lifecycle CO2 emission for methanol derived from coal is �5.3 t CO2

per ton of methanol burned. For gasoline, the value is �4.03 t (Yang & Jackson, 2012).In order to increase fuel economy, reduce CO2 and meet more stringent emission regulations, automobile manufacturers are

increasingly employing gasoline direct injected (GDI) engine technology (Maricq et al., 2012; Myung et al., 2012). Due to thefuel is directly sprayed into the cylinder with higher injection pressures than port fuel injected (PFI) engine, the atomizationand vaporization rate of the fuel are increased. Also, the volume of fuel and process of combustion can be more accuratelycontrolled at different working conditions. Along with stratified combustion and charge air cooling technology, GDI engine canimprove fuel efficiency, reduce cold start unburned hydrocarbons emissions and reduce CO2 emissions (Zhao et al., 1999).However, since the time for preparing even combustible mixture is short and fuel impingement on surfaces of piston andcylinder happens unexpectedly, GDI engine may emit more particulate matter (PM) (Maricq et al., 2012; Myung et al., 2012;Zhao et al., 1999).

Facing the popularization of GDI technique and increase of concerns on environmental protection, more stringentparticulate matter emission standards applied to GDI engines are introducing around the world. In the European Union(EU), according to commission regulation (EC) no. 692/2008, also called Euro 5 and Euro 6, the particle limit valuesare 4.5 mg/km (http://ec.europa.eu/enterprise/sectors/automotive/documents/directives/motor-vehicles/index_en.htm).In the USA, both California Air Resources Board (ARB) and EPA are considering to tighten next tailpipe PM emissionsstandards. Particle limit values may require decreasing from 10 mg/mi to 6 mg/mi, and then 3 mg/mi (Maricq et al., 2012).In China, Beijing Municipal Environmental Protection Bureau is receiving suggestions from the public on the implementa-tion of the BEIJING regulation, which equals to requirements of Euro 5 on PM emissions (http://www.bjepb.gov.cn/portal0/Table189/info8584.htm).

Studies on regulated and unregulated emissions from engines tested on engine test benches or passenger vehiclestested on chassis dynamometers fueled with methanol/gasoline blends have been reported (Abu-Zaid et al., 2004; Bilgin &Sezer, 2008; Wei et al., 2008; Zhang et al., 2010; Zhao et al., 2010). But few investigated the effects of adding methanol togasoline on PM emissions from GDI engines and GDI vehicles. In this study, two passenger vehicles, respectively equippedwith GDI engine and PFI engine, fueled with gasoline and M15 methanol gasoline were examined. Tests were conducted onthe chassis dynamometer over New European Driving Cycle (NEDC) at cold start mode. PM emissions were analyzed withtwo metrics: numbers measured by ELPI; masses measured by gravimetric method and estimated from measured numberdistributions.

2. Material and methods

2.1. Test vehicles and fuels

Two light-duty passenger vehicles were tested in the comparative experiments. Both were manufactured by ShanghaiVolkswagen Automotive Co., Ltd. and met the Euro 4 emission standard. Table 1 shows detailed technical data of the twovehicles. One equipped with gasoline direct injected engine which is the main stream GDI engine type in China. The otheris port fuel injected. Both are turbocharged and fitted with three-way catalyst. Two types of fuel were used: commercial 93research octane number gasoline and M15 methanol gasoline. The commercial 93 gasoline complies with China fourth-stage fuel standard which equal to Euro 4. M15 methanol gasoline was made by mixing the commercial 93 gasoline withindustrial grade methanol in fraction of 15% by volume. Table 2 shows properties of the two fuels. In the following context,Combo 1, Combo 2, Combo 3 and Combo 4 are used to represent the GDI vehicle fueled by gasoline, the GDI vehicle fueledby M15, the PFI vehicle fueled by gasoline and the PFI vehicle fueled by M15 respectively.

2.2. Test cycle

All tests were carried out on a 40-in. single roll, DC electric chassis dynamometer (PECD 9400, Ono Sokki Ltd., Japan)over cold start NEDC. Before each test, the vehicle must be conditioned at a temperature of (2572) 1C over 16 h. The NEDC

Page 3: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

Table 1Vehicle specifications.

GDI PFI

Type Passat 1.8TSI Passat 1.8T

Displacement (L) 1.798 1.781

Number of cylinders 4 4

Bore� stroke (mm2) 82.5�84.1 81�86.4

Compression ratio 9.6 9.3

Maximum power at engine speed (kW/rpm) 118/4500–6200 110/5700

Maximum torque at engine speed (N m/rpm) 250/1500–4500 210/1750–4600

Odometer (km) 12,076 75,680

Table 2Fuel properties.

Characteristic Gasoline M15

T10 (1C) 56.6 45.1

T50 (1C) 91.3 85.9

T90 (1C) 162.8 158.2

Sulfur (ppm) 30.6 26.0

Aromatics (vol%) 34.1 22.7

Reid vapor pressure (kpa) 53.2 65.4

Fig. 1. PM gravimetric measurement system.

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–3124

includes two parts. The first part is urban cycle which consists of four nonstop repeated ECE-15 cycles. The second part isextra urban driving cycle (EUDC). Urban cycle lasts for 780 s and travels 4.052 km while EUDC, 400 s and 6.955 km.The maximum speed in the urban cycle and EUDC is 50 km/h and 120 km/h respectively.

2.3. Gravimetric measurement of PM

Fig. 1 shows the schematic diagram of the gravimetric measurement system. According to the emission regulations,through the full flow constant volume sampling system (CVS-7400T, Horiba Ltd., Japan), tailpipe gas was diluted by air andPM was collected on the filter. Dilution air was filtered to extract particulate matters by a dilution tunnel (DLT-1230,Horiba Ltd., Japan). 99.97% of particles with size diameter of 0.3 mm were filtered. The temperature of the dilution air wasconditioned to 2572 1C. The relative humidity was conditioned to 4572% relative humidity. The tunnel flow rate andflow rate through the filter were 9 m3/min and 90 L/min respectively. The filters were weighted on microbalance (CPA2P-F,Sartorius AG, Germany) in a Euro 5 standard compliant static controlled weighing chamber. The standard deviation of themicrobalance is less than or equal to 1 mg. In the chamber, temperature and dew point of the environment wereconditioned to 2271 1C and 9.571 1C respectively. Filter weights were corrected for filter buoyancy in air.

2.4. Number measurement of PM

Electrical low pressure impactor (ELPI) (Dekati Ltd., Finland) was used to monitor the size distribution of aerosol particles. ELPIincludes 12 channels and each channel is connected to an electrometer current amplifier. Particles were first charged and thencollected into different stages according to inertia. During each stage, corresponding electrometer current amplifier detects thecurrent (Marjamaki et al., 2000). The current value and the size distribution in each stage are proportional based on particle sizedependent relations between the properties of the charger and the impactor stages (Liu et al., 2011). SEMTECH emission flowmeter (Sensors, Inc., the USA) and SEMTECH-DS analyzer were used to directly measure real-time exhaust flow rates. The flowmeter was integrated with a stainless steel pipe of which outside diameter is four inches.

Page 4: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

Fig. 2. PM number measurement system.

Fig. 3. PM masses of the four combinations tested through gravimetric measurement method.

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–31 25

Fig. 2 shows the schematic diagram of the number measurement system. The stainless steel pipe of the flow meter wasconnected on the tailpipe. Gases from the pipe were partial flow sampled. Sampled gas passed through two dilutors and wasmeasured by ELPI. Both dilutors used compressed air which was dried, high efficiency particulate air filtered to dilute the inlet gas.In order to prevent the high concentrated and high temperature exhaust from condensing and form an even dilution of gases andparticles, the compressed air to the first dilutor and the mixture out of the first dilutor were all preheated to 195 1C.

2.5. Test procedures

Each vehicle was firstly fueled with gasoline. Vehicles with gasoline were measured by gravimetric method on the firstday and were tested by ELPI on the second day. After the tests on gasoline were finished, the fuel tanks were drained. Eachvehicle was then filled with 10 l M15 methanol gasoline and was driven at a constant velocity of 60 km/h on the chassisdynamometer until it had no fuel. The fuel tanks were drained again and each vehicle was filled with 10 L M15 again. Samedriving process was done until the fuel was not enough. Then vehicles were drained for the third time and filled up withM15. After the change of fuel was done, measurements on M15 were conducted accordingly.

3. Results and discussions

3.1. PM mass and number emissions

Fig. 3 presents particle masses measured by gravimetric method for the four vehicle/fuel combinations over cold startNEDC. Fig. 4 exhibits total particle numbers tested by ELPI. Units are milligram per kilometer (mg/km) and number per

Page 5: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

Fig. 4. Total particle numbers of the four tests measured by ELPI.

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–3126

kilometer (#/km) respectively. There are three points need to be mentioned for the two figures. (1) Compared withgasoline, the GDI vehicle fueled with M15 gets 78% and 56% reductions for PM mass and total particle number respectively.For the PFI vehicle, the percentage reductions are 74% and 25%. Two possible reasons may explain these reductions. Firstly,M15 contains more oxygen than gasoline. The ‘‘pre-mixed oxygen effect’’ results in a more complete combustion(Wei et al., 2008; Zhao et al., 2010). Secondly, adding methanol in the gasoline can reduce the soot precursorconcentrations. Compared with gasoline, methanol does not have aromatic components that are prone to soot production(Westbrook et al., 2006). (2) For gasoline, the PFI vehicle’s PM mass and total particle number decline about 76% and 77%from those from the GDI vehicle. For M15, the values are 72% and 60%. (3) The PM mass limit value in Euro 5 for GDIvehicles is same with that for compression ignition (CI) vehicles. But the PM number limit value for GDI vehicles is notregulated. Comparisons between tested numbers and the PM number limit value for CI vehicles were conducted. But themain difference between the particle measurement method of ELPI and that legislated in the Euro 5 need to be mentioned.According to the European method, non-volatile particles are counted by a particle number counter with lower cut point at23 nm. The volatiles and semi-volatiles are removed by using a hot dilution stage at 150–400 1C (the wall temperature setpoint should not exceed the wall temperature of the evaporation tube) and an evaporation tube at 300–400 1C. Since ELPIdoes not have the evaporation tube, total particles were measured. Also the lower cut point of ELPI is about 8 nm. Ratios ofPM masses for Combo 1, Combo 2, Combo 3 and Combo 4 to the Euro 5 limit value are 3.1, 0.69, 0.74 and 0.19 respectively.For the total particle number, ratios are 7.8, 3.4, 1.8 and 1.4 accordingly. Ratios of 3.1 and 7.8 suggest that GDI vehicles facesimilar challenges from Euro 5 standard as diesel passenger vehicles on PM emissions. When GDI vehicles are fueled withM15, PM masses can meet the requirement. Although PFI vehicles have less mass emissions, their particle numbers arealso need attention. Attributed to the development of emission control techniques for diesel vehicles, PM emissions fromEuro 5 or Euro 6 diesel vehicles are decreasing greatly. Contribution of gasoline vehicles to the PM inventory will graduallyincrease and cannot be ignored.

Fig. 6 illustrates transient particle number emission rates from the GDI vehicle fueled with gasoline over cold startNEDC. Particle number emission rates are derived by multiply PM number concentrations measured by ELPI with time-aligned tailpipe emission flow rates measured by SEMTECH-DS. As can be seen in the figure, particle number emissionscorrelate with the vehicle acceleration. Transient particle number emission rates of the other three vehicle/fuelcombinations show similar correlation.

3.2. Particle number distribution

Fig. 5 shows PM number distributions of the four vehicle/fuel combinations. The four curves perform similar trend andall decrease gradually. The maximum values for Combo 1, Combo 2, Combo 3 and Combo 4 are 1.84�1012 km�1,8.46�1011 km�1, 5.39�1011 km�1 and 5.14�1011 km�1 respectively. All these peaks occur in the first stage with anaverage diameter of 21 nm. At each stage, the PM number from each vehicle with M15 is lower than that with gasoline;the PM value from the GDI vehicle is higher than that from the PFI vehicle for the same fuel. Nanometer size particles arefacing more stringent emission standards for its strong relation with mortality (Liu et al., 2011). EPA proposes to revise theannual PM2.5 (particles less than or equal to 2.5 mm in diameter) standard by lowering the level to within a range from12.0 to 13.0 micrograms per cubic meter (mg/m3) (http://www.epa.gov/airquality/particlepollution/actions.html#jun12).China issued GB3095-2012 on Feb. 29, 2012 in which PM2.5 limit values were first introduced. The 24-h average values are35 mg/m3 and 75 mg/m3 for the first and second class districts respectively (http://kjs.mep.gov.cn/hjbhbz/bzwb/dqhjbh/).During these four tests, PM numbers of the stages 1–10 of which diameters are less than 2.5 mm amount for 99.996%,99.993%, 99.985% and 99.989% of the total numbers respectively for Combo 1, Combo 2, Combo 3 and Combo 4.The percentages of stages 1–8 of which diameters are less than 1 mm are 99.982%, 99.970%, 99.949% and 99.957%.

Page 6: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

Fig. 5. Particle number distributions for the four tests measured by ELPI.

Fig. 6. Transient PM number emissions for the GDI vehicle fueled with gasoline over cold start NEDC.

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–31 27

3.3. PM mass estimations

ELPI is mostly used for measurement of particle number distributions. Some researches on conversion of the ELPIparticle number distributions to particle mass were also reported. Liu et al. (2011) used size distributions of the first 7 ELPIstages to estimate particle masses from four diesel buses (two Euro 3 level buses and two Euro 4 level buses). But they didnot compare the estimations with gravimetric measurements. Zervas et al. (2006) used ELPI to estimate emissions fromthree Euro 3 cars (a PFI gasoline car, a diesel car and a diesel car equipped with diesel particulate filter) over NEDC andcompared the estimations with filtered masses. Also they compared the difference of the estimated results between thefirst 7 stages and all stages. They concluded that the use of all stages overestimated the particle masses. When the first 7ELPI stages were used, results are better, but the dispersion is also high. Maricq et al. (2006) used the lowest 7 ELPI stagesto estimate emissions from five cars (a diesel vehicle with an oxidation catalyst, a diesel vehicle equipped with a dieselparticulate filter and an oxidation catalyst, a GDI vehicle and two PFI vehicles) and concluded that particle masses derivedfrom number concentrations agree well with masses measured through gravimetric method on high level PM emissionvehicles. They also stated that on the upper stages, non-idealities like electrical noise, electrometer drift, and small errorsin the diffusion and electrostatic loss will be amplified by mass weighting. Based on these studies, the first 7 ELPI stageswere used in this study for mass calculation.

In order to calculate PM masses from number concentrations measured by ELPI, the effective density of particles is needto be given. Two kinds of density distributions were reported. One is constant with the particle size. The values usedincluded 1.0 g/cm3 (Shi et al., 1999; Zervas et al., 2006), 0.5 g/cm3 (Witze et al., 2004) and 1.7 g/cm3 (Ulfvarson et al., 1997). Theother is a function of the size. Particles emitted from vehicle engines are agglomerates of small carbonaceous primary particles(Maricq et al., 2006). Since particles are highly agglomerated as size increases, the effective density decreases as the particlesize increases (Park et al., 2003). Reported values include values of 1.2–0.1 g/cm3 (Zervas et al., 2006), 1.0 g/cm3 at 50 nm and0.3 g/cm3 at 300 nm (Witze et al., 2004), 1.2 g/cm3 at 30 nm to o0.3 g/cm3 at 300 nm (Maricq & Xu, 2004), 1.5–0.2 g/cm3

Page 7: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–3128

(Andrews et al., 2001) and 1.6–0.2 g/cm3 (Ahlvik et al., 1998). Among the second kind of density distributions, the power lawrelationship shown in Eq. (1) between the effective density and mobility diameter were reported (Maricq & Xu, 2004; Parket al., 2003; Virtanen et al., 2004).

rep dmdf�3

ð1Þ

where re is the particle’s effective density, dm the particle’s mobility diameter and df the fractal dimension which is commonlyused to characterize the structure of agglomerated particles. This relationship is based on the fractal like morphology ofparticles and thus it is a reasonable approximation.

In this study, two density distributions were used. One is constant with particle sizes and the density value is 1.0 g/cm3.The other is a detailed power law density shown in Eq. (2) used by Maricq et al. (2006).

re ¼ r0 dm=d0

� �df�3ð2Þ

where r0 is the primary particle density, d0 the primary particle diameter. r0, d0 and df are given the value of 2.0 g/cm3,20 nm and 2.3 nm respectively. Density distributions used for estimation were showed in Fig. 7.

For the constant effective density distribution, the estimated particle mass is the sum of the particle mass of each stage.The particle mass of each stage is calculated by multiply effective density, volume and particle numbers of each stage.For the power law effective density distribution, the estimated particle mass is calculated via the fitting Eq. (3) used byMaricq et al. (2006).

M¼N0p6r0d0

3�dfð Þmgdf edf

2 lnsgð Þ2=2 ð3Þ

where N0 is the total particle number and mg the geometric mean diameter. Values of r0, df and d0 are same with those inEq. (2).

Fig. 8 shows particle masses estimated from the two density distributions and ratios of the estimated masses togravimetric measurements for the four vehicle/fuel combinations. The upper part of this figure shows the estimatedresults and the lower part gives the ratios. Particle masses for the four vehicle/fuel combinations show a decreasing trendand follow the order Combo 14Combo 24Combo 34Combo 4 when each kind of density distribution is used. For eachcombination, the particle mass derived from constant density distributions is higher than that from power law ones.The exceeding percentage for Combo 1, Combo 2, Combo 3 and Combo 4 is 127%, 181%, 195% and 276% respectively. Foreach combination, estimated masses are much lower than the measured one. Masses calculated from constant densitydistribution are 7%, 17%, 5% and 13% of measured ones for Combo 1, Combo 2, Combo 3 and Combo 4 respectively. Whenpower law density distributions are used, percentages are 3%, 6%, 2% and 4%. Several factors can help to explain why thesebig discrepancies exist.

(1)

For the low PM emitting vehicles, such as PFI gasoline vehicles and diesel particulate filter (DPF) equipped dieselvehicles, ELPI derived PM masses are much lower than gravimetric PM masses. From the data by Zervas et al. (2005,2006), PM masses estimated from the first seven ELPI stages are averagely less than 20% and 10% of measured ones forthe PFI gasoline vehicle and the DPF equipped diesel vehicle respectively over cold start NEDC (a constant densityvalue of 1.0 g/cm3 for each stage and the geometric mean diameter of each stage are used). Data by Maricq et al. (2006)show that estimated PM masses are less than half of filter collected ones for the DPF equipped diesel vehicle over eachphase of the FTP cycle (power law density distribution and fitting equation mentioned above are used). Mostly thepercentages are less than 40%. For PFI gasoline vehicles, except for the cold start phase, ratios of estimated masses to

Fig. 7. Two density distributions used for estimating PM masses from PM number concentrations measured by ELPI.

Page 8: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–31 29

measured ones are less than 20% over the city portion and hot start phase of FTP cycle. Mostly the percentages are lessthan 10%.

(2)

In the previous studies (Maricq et al., 2006; Zervas et al., 2006), particles used for weighing and estimating were allsampled from the full flow CVS dilution tunnel in the same test. The same tunnel flow rate was used in both methodsto calculate the final emission rates. In this paper, particles used for weighing were sampled from the CVS dilutiontunnel in the first test. But particles and the total flow rates used for estimating were directly sampled from thetailpipe in the second test. For the comparison of the two test results, repeatability and measuring differences of flowrate seem to be two important influence factors. Zervas et al. (2006) reported that the repeatability of total particle

Fig. 8. Particle masses estimated and ratios between estimated masses and measured masses.

Fig. 9. Flow rates direct measured from the tailpipe for the two vehicles fueled with gasoline over NEDC.

Page 9: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–3130

number measured by ELPI is 45%, 14%, and 96% for gasoline, diesel, and DPF-equipped diesel vehicles respectively.For filter collected PM, values are 60%, 12.7%, and 191%. These results show that comparative results of PM massesbetween the two methods are much scattering.Fig. 9 gives flow rates direct measured from the tailpipe for the two vehicles fueled with gasoline over NEDC (for eachvehicle, flow rates recorded for M15 are similar with those for gasoline. So the two curves for M15 are not shown).The relationships between flow rates and vehicle speeds are reasonable. But the values are low, especially for the GDIvehicle. Sonntag et al. (2010) also reported that low exhaust flow readings were recorded in direct sampling by aHoriba OBS-1000 which uses the same measurement principle as that used in this paper. Eq. (4) was used by them toapproximately estimate the minimum exhaust flow rate. This equation is based on the assumption that the exhaustgas is at constant pressure and temperature between the cylinder and sampling point. The fuel rate is also ignored forminor contribution to the exhaust flow at low fuel rate (Sonntag et al., 2010).

exhaust flow¼ engine displacement� engine speed� volumetric efficiency� 1=2� �

ð4Þ

Engine speed is given the value of 750 rpm and volumetric efficiency is assumed to be 0.7. The calculated theoreticalminimum exhaust flow rate is 7.8 L/s. This value is close to the flow rates of the first 11 s of idle stage. The averagemeasured flow rate of this stage is 6.4 and 7.4 L/s for the GDI and PFI vehicle respectively. From this comparison, thecalculated minimum flow rate seems to be reasonable. And thus the total measured flow rates are lower. 76% And 66%of the measured flow rates are lower than the calculated minimum one for the GDI and PFI vehicle respectively. Ratiosof the minimum measured flow rate to the calculated one are 21% and 45% for the GDI and PFI vehicle respectively.For the GDI vehicle, 28% of the measured flow rates are less than 40% of the calculated one. These results may partiallyexplain why the ratios of estimated masses to measured ones are low, especially for the GDI vehicle.

(3)

Organic vapor adsorption of filters used for weighing cannot be neglected for low PM emitting vehicles. Chase et al.(2004) reported that the vapor artifact accounts for 10–20% and 30–50% of the 2007 regulatory standard of 10 mg/mifor light duty vehicles for Teflo and TX40 filters respectively. Our data may also reflect this conclusion. Since methanoldoes not have aromatic components, when the vehicles were fueled with M15, filters may collect less organicmaterials than those with gasoline. Thus ratios of estimated/measured masses for vehicles fueled with M15 are higherthan those with gasoline.

(4)

Systematic differences that come from the uncertainties in the soot effective density and the geometric standarddeviation of the soot mobility distribution may result in an estimated 20% uncertainty in ELPI derived PM masses(Maricq et al., 2006).

4. Conclusions

Particle masses and particle number concentrations from a GDI vehicle and a PFI vehicle over NEDC were measured.Each vehicle was respectively fueled with gasoline and M15 methanol gasoline. PM mass estimations from measurednumber concentrations using two different density distributions were conducted.

For each vehicle, PM masses measured by gravimetric method, the total PM numbers measured by ELPI and estimatedPM masses for M15 are lower than those for gasoline. Studies by Maricq et al. (2012) indicated that when the ethanolcontent increases to 430%, reduction in PM mass and number emissions are 30–45%. Storey et al. (2012) reported that PMmasses from the stoichiometric GDI vehicle showed obvious reduction with E10 and E20. The percentage reduction for E10and E20 are about 25% and 60% respectively. Take these three studies together, similar effects were found on PM mass andnumber emissions between the adding of methanol to gasoline and adding of ethanol to gasoline.

For each kind of fuel, PM masses by two methods and the total PM numbers from the GDI vehicle are higher than thosefrom the PFI one. Similar results were reported by Maricq et al. (2006). Their studies indicated that PM masses from a GDIvehicle fueled with gasoline over NEDC are in the range of 20–35 mg/km; for phases 1–3 of the FTP cycle, PM masses fall inthe range of 15–20, 10–15 and 5–10 mg/km respectively. For the PFI vehicle, PM masses of the three stages of the FTP cycle areall lower than 1 mg/km. Zhao et al. (1999) also mentioned that PM masses of a GDI vehicle are on the order of 10 mg/mile andthose from a PFI vehicle are in the range of 1–3 mg/mile.

PM number distribution curves of the four vehicle/fuel combinations are similar. All decline gradually and themaximum number of each curve occurs in the first stage. More than 99.9% numbers locate in the first 8 stages of whichdiameters are less than 1 mm. PM number emissions correlate well with the acceleration of the two vehicles. Two densitydistributions and numbers of the first seven ELPI stages are used for particle mass estimation. But the estimated resultswere much lower than the gravimetric measurements.

Acknowledgments

The authors gratefully acknowledge financial support from Energy Foundation (G-1206-16338 ).

Page 10: Comparison of PM emissions from a gasoline direct injected (GDI) vehicle and a port fuel injected (PFI) vehicle measured by electrical low pressure impactor (ELPI) with two fuels:

B. Liang et al. / Journal of Aerosol Science 57 (2013) 22–31 31

References

Abu-Zaid, M., Badran, O., & Yamin, J. (2004). Effect of methanol addition on the performance of spark ignition engines. Energy Fuels, 18, 312–315.Ahlvik, P., Ntziachristos, L., Keskinen, J., & Virtanen, A. (1998). Real time measurements of diesel particle size distribution with an electrical low pressure

impactor. SAE paper no. 980410.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. SAE paper no. 2001-01-1946.Bilgin, A., & Sezer, I. (2008). Effects of methanol addition to gasoline on the performance and fuel cost of a spark ignition engine. Energy Fuels, 22,

2782–2788.Chase, R.E., Duszkiewicz, G.J., Richert, J.F.O., Lewis, D., Maricq, M.M., & Xu, N. (2004). PM measurement artifact: organic vapor deposition on different filter

media. SAE paper no. 2004-01-0967.GB3095-2012. Ambient air quality standards. Available at: /http://kjs.mep.gov.cn/hjbhbz/bzwb/dqhjbh/S.Liao, S.Y., Jiang, D.M., Cheng, Q., Huang, Z.H., & Zeng, K. (2006). Effect of methanol addition into gasoline on the combustion characteristics at relatively

low temperatures. Energy Fuels, 20, 84–90.Liu, Z.H., Ge, Y.S., Johnson, K.C., Shah, A.N., Tan, J.W., Wang, C., & Yu, L.X. (2011). Real-world operation conditions and on-road emissions of Beijing diesel

buses measured by using portable emission measurement system and electric low-pressure impactor. Science of the Total Environment, 409,1476–1480.

Maricq, M.M., Szente, J.J., & Jahr, K. (2012). The impact of ethanol fuel blends on PM emissions from a light-duty GDI vehicle. Aerosol Science andTechnology, 46, 576–583.

Maricq, M.M., & Xu, N. (2004). The effective density and fractal dimension of soot particles from premixed flames and motor vehicle exhaust. Journalof Aerosol Science, 35, 1251–1274.

Maricq, M.M., Xu, N., & Chase, R.E. (2006). Measuring particulate mass emissions with the electrical low pressure impactor. Aerosol Science and Technology,40, 68–79.

Marjamaki, M., Keskinen, J., Chen, D.R., & Pui, D.Y.H. (2000). Performance evaluation of the electrical low-pressure impactor (ELPI). Journal of AerosolScience, 31, 249–261.

Myung, C.L., Kim, J., Choi, K., Hwang, I.G., & Park, S. (2012). Comparative study of engine control strategies for particulate emissions from direct injectionlight-duty vehicle fueled with gasoline and liquid phase liquefied petroleum gas (LPG). Fuel, 94, 348–355.

Park, K., Cao, F., Kittelson, D.B., & Mcmurry, P.H. (2003). Relationship between particle mass and mobility for diesel exhaust particles. EnvironmentalScience and Technology, 37, 577–583.

Shi, J.P., Harrison, R.M., & Brear, F. (1999). Particle size distribution from a modern heavy duty diesel engine. Science of the Total Environment, 235,305–317.

Sonntag, D.B., Gao, H.O., & Holmen, B.A. (2010). Comparison of particle number and mass emissions from diesel transit buses across temporal and spatialscales. UTRC’s student report. Available at: /http://www.utrc2.org/publications/research-project-reports?tid=332&page=1S.

Storey, J., Barone, T., Thomas, J., & Huff, S. (2012). Exhaust particle characterization for lean and stoichiometric DI vehicles operating on ethanol–gasolineblends. SAE paper no. 2012-01-0437.

Ulfvarson, U., Figler, B., & Krantz, S. (1997). Diesel engine development is guided by inadequate particle sampling. SAE paper no. 970759.Virtanen, A., Ristimaki, J., & Keskinen, J. (2004). Method for measuring effective density and fractal dimension of aerosol agglomerates. Aerosol Science and

Technology, 38, 437–446.Wei, Y.J., Liu, S.H., Li, H.S., Yang, R., Liu, J., & Wang, Y. (2008). Effects of methanol/gasoline blends on a spark ignition engine performance and emissions.

Energy Fuels, 22, 1254–1259.Westbrook, C.K., Pitz, W.J., & Curran, H.J. (2006). Chemical kinetic modeling study of the effects of oxygenated hydrocarbons on soot emissions from diesel

engines. Journal of Physical Chemistry A, 110, 6912–6922.Witze, P.O., Chase, R.E., Maricq, M.M., Podsiadlik, D.H., & Xu, N. (2004). Time-resolved measurements of exhaust PM for FTP-75: comparison of LII, ELPI

and TEOM techniques. SAE paper no. 2004-01-0964.Yang, C.J., & Jackson, R.B. (2012). China’s growing methanol economy and its implications for energy and the environment. Energy Policy, 41, 878–884.Zervas, E., Dorlh �ene, P., Forti, L., Perrin, C., Momique, J.C., Monier, R., Ing, H., & Lopez, B. (2005). Interlaboratory test of exhaust PM using ELPI. Aerosol

Science and Technology, 39, 333–346.Zervas, E., Dorlh�ene, P., Forti, L., Perrin, C., Momique, J.C., Monier, R., Ing, H., & Lopez, B. (2006). Exhaust gas particle mass estimation using an electrical

low pressure impactor. Energy Fuels, 20, 498–503.Zhang, F., Zhang, X., Shuai, S.J., Xiao, J.H., & Wang, J.X. (2010). Unregulated emissions and combustion characteristics of low-content methanol–gasoline

blended fuels. Energy Fuels, 24, 1283–1292.Zhao, F., Lai, M.-C., & Harrington, D.L. (1999). Automotive spark-ignited direct-injection gasoline engines. Progress in Energy and Combustion Science, 25,

437–562.Zhao, H., Ge, Y.S., Hao, C.X., Han, X.K., Fu, M.L., Yu, L.X., & Shah, A.N. (2010). Carbonyl compound emissions from passenger cars fueled with methanol/

gasoline blends. Science of the Total Environment, 408, 3607–3613.


Recommended