• 沒有找到結果。

Real-world vehicle emissions and VOCs profile in the Taipei tunnel located at Taiwan Taipei area

N/A
N/A
Protected

Academic year: 2021

Share "Real-world vehicle emissions and VOCs profile in the Taipei tunnel located at Taiwan Taipei area"

Copied!
10
0
0

加載中.... (立即查看全文)

全文

(1)

Real-world vehicle emissions and VOCs profile in the Taipei

tunnel located at Taiwan Taipei area

Mei-Yin Hwa

a,c,

*, Chu-Chin Hsieh

b

, Tsung-Cheng Wu

b

, Len-Fu W. Chang

a

a

Graduate Institute of Environmental Engineering, National Taiwan University, Taiwan

b

Department Environment and Safety Engineeing, National Yunlin University of Science and Technology, Taiwan

c

Department of Environmental Engineering, Tung Nan Institute of Technology, Taipei, Taiwan Received 14 September 2001; received in revised form 3 January 2002; accepted 15 January 2002

Abstract

An in situ field experiment was conducted in a highway road tunnel in the Taipei City to determine the motor vehicle emission factors (EF) of different kinds of air pollution species. These are carbon monoxide (CO), oxides of nitrogen (NOx), non-methane hydrocarbons (NMHC) and VOCs species. About 56 species of VOCs were sampled by canister

sampler and followed by the GC-MS analyzing. Furthermore, the tunnel-drafting rate was determined by SF6tracer

method.

The EF for the highway vehicles determined from this experiment are 3.64, 0.90, 0.44 and 0.24 gm km1veh1for CO, NOx, NMHCand the totally measured VOCs, respectively. A comparison of the EFs from the road tunnel

experiment to the estimates by the USEPA MOBILE5b (M5b) and the modified Taiwan EPA MOBILE-TAIWAN2.0 (MT2.0) provides a first-hand evaluation of the model characteristics. M5b and MT2.0 both tend to underpredict CO by 10% and 20%, respectively. While M5b overpredicts NOxand NMHCby 40% and 20%, respectively; MT2.0 has

fairly good predictions for these two species. From the GC-MS analysis of the canister samples, it was found that the most abundant species from the traffic-emitted VOCs in Taipei road tunnel are toluene, ethene and 1,2,4-trimethyl-benzene (1,2,4-TMB) by the weight basis. However, ethene, acetylene and toluene are the most abundant in VOCs based on volume. The VOCs’ weight composition in terms of the carbon bond classification is 28% by the paraffins, 33% by the olefins and 39% by the aromatics, respectively. In order to evaluate the ozone formation potential from the typical road emission in Taipei area, the maximum increment reactivity is calculated. It was found that about 1015 mg of O3is induced by per vehicle per kilometer traveled emission. Among them, ethene, 1,2,4-TMB and propene from the

road vehicle’s emission contribute most to the ozone-formation reactivity. r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Emission factor; Tunnel study; VOCprofile; Taiwan; MIR

1. Introduction

Ozone pollution attains a serious air quality issue in several urban areas over Taiwan. In recent years, the number of days with ozone hourly concentration exceeding the National ambient air quality standard

(NAAQS) (120 ppb) has replaced PM10to be the major

contributor for air quality episodes in Taiwan. Ozone is not emitted directly into air; it is formed in the troposphere as a result of complex photochemical reactions that involve VOCs and NOxas its precursors.

The anthropogenetic VOCs exhibit a wide range of variation in reactivity with respect to ozone formation (NRC, 1991). Several VOC reactivity-scaling methods have been introduced and proposed in the articles. For example, the California Air Resource Board (CARB,

*Corresponding author. Department of Environmental En-gineering, Tung Nan Institute of Technology, Taipei, Taiwan.

E-mail address:hwa@duck.tnit.edu.tw (M.-Y. Hwa).

1352-2310/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 2 - 2 3 1 0 ( 0 2 ) 0 0 1 4 8 - 6

(2)

1990) has adopted the maximum incremental reactivity (MIR) introduced by Carter (1994) to evaluate the VOC reactivity of alternative motor vehicle fuels (CARB, 1990). NRC(1991) defined the propylene-equivalent concentration, Propy-Equiv (j), for each VOCspecies ‘‘j’’ to account for the combined effect of OH-radical in ambient air and the VOCs’ concentration.

Mobile source becomes a significant origin of the VOCs and NOx in the urban area of Taiwan. The

emission inventory was compiled by CTCI (1999), which indicated that mobile sources contribute 92% of COs total emission, 37% of NMHCs total emission, and 69% of NOxs total emission in Taipei metropolitan area. In

order to figure out an adequate mitigation strategy concerning the ozone problem in Taiwan, it is necessary to obtain a reliable mobile source emission profile. Many emission factor (EF) models have been widely used to do the job. Examples are the MOBILE5, the EMission FACtor (EMFAC7) (CARB, 1996) and the COmputer program to calculate emissions from road transport (COPERT II) (Ahlvik et al., 1997; Winther, 1998) and so on. However, the reliability of a mobile source emission model can be affected by many factors that vary from place to place and traffic fleet by fleet. Thus, the design and implementation of an adequate in situ mobile source emission study become necessary in order to serve as a first-hand comparison and evalua-tion. Highway and/or road tunnel studies have been used to measure the real-world emissions of motor vehicles. The method, compared to the method of the individual test of vehicle on a dynamometer, has the advantage on their simplicity, cost-effectiveness and sufficiency in statistical sample size. The important findings from the results of Van Nuys Tunnel study (Ingalls, 1989) revealed that motor vehicle emissions of CO and NMHC were about a factor 2 higher than the predictions by EMFAC7C. Similar road tunnel studies conducted in different places can be found in the articles, such as Gubrist Tunnel experiment of Switzer-land (Staehelin et al., 1995, 1997, 1998; John et al., 1999), Fort McHenry Tunnel experiment of Baltimore Harbor, Maryland (Pierson et al., 1996) and Tuscarora Mountain Tunnel of Pennsylvania (Sagebiel et al., 1996). All these kinds of activities contribute signifi-cantly to the improvement of the development of traffic emission models. In Taiwan, the tunnel study was first conducted in southern Taiwan area by Tsai et al. (1998) for the Chung–Cheng tunnel in the Kao-Hsiung city (Tsai et al., 1998; Hsu et al., 2001). However, the tunnel is relatively short and thus the entrance effect becomes a major concern. Later on, an experiment was carried out in the Lishin and Zefun tunnel of Taipei county for the emissions of volatile aromatic compounds from the combined traffic com-posed by motorcycles and light-duty (LD) vehicles (Hsieh et al., 1999).

In this study, the Taipei tunnel was selected to study the highway traffic emissions. It is on a connecting freeway that provides the channel to access the turnpike No. 3, an artery freeway of Taiwan Island, from southeastern Taipei urban area. Most of the traffic is commuters; they use the turnpike No. 3 to access their work places from towns within 100 km distance. The experiment was voted to determine the EF of CO, NOx,

NMHCand important VOCs species from motor vehicles that traveled through the tunnel with different congestion patterns in a daily cycle. Furthermore, a direct deduction from the dilution factor of a controlled release of tracer gas, SF6, was used to determine the

induced tunnel-drafting rate (the volumetric flow rate). It is believed that this method can significantly reduce the uncertainty in the determination of the traffic-induced drafting inside a highway tunnel.

2. In-situ measurement 2.1. Tunnel description

The Taipei tunnel is located on the connecting highway to the turnpike No. 3 that goes through the southeast corner of the great Taipei Metropolitan area. This area has inhabitants of more than 5 million. The tunnel facilitates a two-bored, no signal light control and divided highways for the eastbound and westbound traffics, which connect the city traffic to the major turnpike. There are two lanes on a one-way bore, and each bore has 60 m2 in cross-section, and 800 m in length, 10.6 m in width, 6.8 m in height. Also, it facilitates a 1.1 m wide sidewalk on the right-hand side of the road and thus provides the mounting space for the instruments in our experiment. The measurement was carried out in the eastbound bore. The tunnel’s orientation is rather straight and the grade is minimal. Since the traffic light control is located beyond 3 km from the entrance of the tunnel, the traffic speed within it is varied mainly by the congestion factor. All vehicles’ operation can be safely assumed to be in the hot-stabilized mode. Four axial-type blowers mounted under the tunnel’s ceiling are equipped to enhance the tunnel drafting. Two of them run automatically at 8:00 a.m. and 18:00 p.m. for each 1-h congestion period. 2.2. Run description

The sampling experiment was carried out from 7:00 a.m. to 17:00 p.m. on 1 July 2000 (Saturday, half-workday). It was a typical sunny and hot summer day in Taipei area. The averaged daytime ambient temperature outside the tunnel during the experiment period was recorded as 32.01C, while the temperature inside the tunnel was about 0.41Chigher. Each sampling that

(3)

sampled the concentrations inside the tunnel lasted for 60 min. Two types of vehicles were classified, namely the LD vehicle and the heavy-duty vehicle (HD). Since, the highway regulation in Taiwan does not allow the two-wheel motorcycle to access the normal highway, the LD contains only passenger cars and small gasoline-fueled or diesel-fueled pick-ups. While, the HD contains those buses and HD trucks used for cargo transport. All of them are powered by the diesel-fueled engine. The traffic volumes and traffic speeds were detected and recorded by the inductance sensors originally mounted by the freeway control center. Serving as a back-up record in our experiment, a video camera (SONY DCR-TRV-310) was used to tape-record the traffic status within the tunnel.

2.3. Instrumentation

Air samplers were located over the sidewalk space in the tunnel. Array of sampling racks was allocated at a distance of 100, 400 and 700 m, respectively, from the entrance point of the tunnel. Each sampling rack has 4 sampling inlet ports, and was located at 1.0, 2.0, 3.0 and 4.5 m heights above the tunnel floor, respectively. Fig. 1 shows the sampling system layout. This arrangement allowed a cross-sectional averaged concentration to be obtained. Air samples contaminated by traffic emissions were collected by either Tedler bags or stainless-steel canisters (6 l in size) located at the aforementioned positions, respectively. Either the Tedler bag or the canister was controlled by floating flow meters at a flow rate of about 100 ml min1level continuous for 60 min. The tracer gas, SF6, was released at the 100 m down the

entrance position from a near floor port, the total mass released in the 1-h period was controlled and recorded. Tedler bag samples were analyzed immediately after each run by automatic monitors for their concentrations of CO, NOx, and SF6. The specifications of analyzers

used are API-300 NDIR CO analyzer, API-200 chemi-luminescence NO–NO2/NOxanalyzer and the HP-5890

GC-ECD SF6 analyzer. The GC-ECD was equipped

with a stainless steel column (Porapak Q 80/100,

2 m  4 mm ID), and the oven temperature was set to 801C, the injector’s and detector’s temperature were set to 2201Cand 2501C, according to the operation manual. All the aforementioned instruments were calibrated at multi-points by standard gas and zero gas before and after the experiment according to the usual routines of quality control.

The canister samples were brought back to the laboratory to analyze its NMHCconcentration by GC-FID method and the speciated composition of VOCs by the GC-MS method. The US EPA Method 15 (TO-15, USEPA, 1997) provides the guidance of GC-MS analysis. The VOCs standard (Enviro-Mat Ozone Precursor Mixture, Matheson Gas Products, Georgia, USA) was prepared and served as the standard. Canister content was injected into a 500 ml pre-conditioning system for thermal conditioning (Entech Instrument, Inc., model 7100, CA). After that, the sample was injected into the gas-chromatograph (GC) (HP 5890 Series-II plus coupled with a HP 5972 MS). The mass-spectrometer (MS) scanned the range of m=z from 40 to 150 using SCAN mode. A non-polar HP-1, 60 m  0.25 mm ID  1.8-mm film thickness capillary column was used to separate individual compound in the sample. NMHCwere analyzed by capillary GC-FID (Shimadzu GC-14B). The column used in this GC-FID is a 15 m  0.53 mm ID silica steel capillary with the 1 m  0.53 mm ID molecular sieve 5A (60–80 mesh). The oven temperature of GCwas set to 1501C.

HP5972 MS and the data processing software (HP G1034CChemstation Software) were used for measur-ing and analyzmeasur-ing species in VOCs. Quality control and assurance requirements include establishing GCreten-tion times, calibraGCreten-tion R2, method detecGCreten-tion limit (MDL) and the study of reproducibility for all analyzing compounds. An internal standard (toluene-d8) was used for each analysis to confirm the stability of the MS. Additionally, cyclohexanal-d12 was analyzed daily as an instrument performance reference standard. For VOCs collected by canisters, method detection limits were between 0.1 and 0.7 ppbv for GC-MS used. The RSD is between 2.2% and 6.5%, which is well within the limit of

2m 4.5m

1.1m 3m 1m

Inlet; 4 inlets of one bag

800m Inlet (West) Outlet (East) 400m 100m 100m 6.8m

Tracer gas injection Sampling site

(4)

725%. The precision in this analyzing routine was between 7% and 8% that is less than the required 730%.

2.4. Determination of the tunnel ventilation flux The tunnel ventilation flux was determined by tracer gas method. A fixed mass emission rate of SF6 was

maintained and recorded at the release point. In the pre-run test, the cross-sectional averaged concentration of SF6 at several downwind sections were sampled and

analyzed to check the mixing status of SF6in the tunnel.

It was found that the assumption of complete mixing over the cross-sectional area of the released SF6can be

adopted at the distance of 75 hydraulic diameters down the releasing point. In the real run period, the averaged concentration of SF6sampled through 5 different ports

located at the last instrument array (100 m before the tunnel exit) was used to infer the drafting rate (ventilation volume flow rate) in the tunnel. Since the vehicle’s body movement induces the draft in the tunnel, the high-intensity turbulence within the drafting flow renders the direct measurement by anemometer method to be extremely difficult and unreliable. With the known SF6concentration at the near exit section and the

pre-known mass emission rate of SF6, the effective drafting

rate can be determined by the mass conservation principle.

2.5. Determination of the vehicle’s EF

The total mass emission rate from the passing vehicles between two cross-sections in the tunnel can be inferred by the mass conservation principle.

M=t ¼ Q  ð %Cout %CinÞ: ð1Þ

Here, M (mg) is the total pollutant’s mass emitted by passing traffic vehicles in-between two cross-sections in the period of time ‘‘t’’. Q is ventilation flux (air flow volume per unit time); %Cinand %Coutrepresent the

sectional averaged concentration at the upwind cross-section and the downwind cross-cross-section, respectively. Giving the traffic volume counting, N (number of passing vehicles) within period ‘‘t’’; and the known distance between two sampling cross-sections, L; the average emission factor, EF in gm veh1km1 for a given run can be obtained

EF ¼ M  ðN  LÞ1: ð2Þ In this study, the difference between LD and HD vehicles are neglected, instead the overall fleet emission was calculated.

The same emission factor can also be obtained from the model’s computation. The applications of MOBI-LE5b (M5b) and Mobile-Taiwan2.0 (MT2.0) are such examples. MT2.0 is a modified version of MOBILE4.1,

with a minor modification with parameters pertinent to Taiwan’s specific vehicle, fuel and driving pattern characteristics. However, these characteristics appear to be reflected in the so-called ‘‘the zero miles emission rate’’ and the ‘‘deterioration rate’’ factors used in the model. Due to the insufficient sample size in performing the test for MT2.0, we regard that the determination of these parameters might be a significant concern. Furthermore, the hot summer climate in Taiwan causing the ‘‘hot soak’’ emission becomes an unknown effect. All these uncertainties need to be solved by the real road test of traffic emission. As for the purpose of model evaluation and comparison, we did not alter the model parameters except those of in situ road fleet vehicle composition and the traveling speed as measured. Both models and tunnel experiments provide the EF of CO, NOxand NMHCin the unit of gm km1veh1to reflect

the emissions on an averaged freeway traffic condition during summer in Taiwan climate.

3. Results and discussion 3.1. Traffic flow

The hourly traffic volume ranged from 1643 (7–8 a.m.) to 2819 (1–2 p.m.) with an average of 2296 veh1h1on the experiment day. Among 94–98% of the traffic volume were LD vehicles, in which 4–10% was diesel-fueled. The experimentally averaged speed of LD in the tunnel was recorded at 67 km h1, the variations were from 64 (1–2 p.m. on Saturday, half-working-day rush hour) to 71 (7–8 a.m.) km h1. The average speed of HD vehicles ranged from 60 (1–2 p.m.) to 69 km h1(11–12 p.m.), the experimentally averaged speed was 65 km h1.

3.2. Tunnel ventilation flux

The ventilation flux in a tunnel is vital in controlling the pollution level inside it. Concerning our experiment in the determination of the EFs of on-road vehicles’ emission from tunnel measurement, accurate measure-ment of the effective ventilation flux is key to determine the emission rate inferred from concentration measure-ment. The drafting effect inside a tunnel is mainly caused by the aerodynamic drag force from the moving vehicles. The shape factor of the car, its moving speed and the sparsity among vehicles determine the effective drafting flow. However, due to the nature of turbulence and intermittency in the induced drafting flow, the point measurement by a traditional anemometer becomes difficult and even unreliable.

The result from the SF6tracer method is presented in

Fig. 2. Interesting features are revealed by the figure. The tunnel cross-section averaged ventilation velocity is

(5)

seen to be inversely proportional to the vehicle’s speed, although with a rather large scatter from the regression line. The drafting speed is seen to increase as the in-tunnel traffic volume (the number of passing vehicles per time) increases. This makes sense that traffic volume may affect the total number of the dragging bodies in the tunnel. The ventilation flux from our experiments ranged from 250 m3s1at 7–8 a.m. to 290 m3s1at 1–3 p.m., with an average of 272 m3s1.

3.3. Emission factors

Due to the insignificant variation of the HD/LD traffic volume ratio in our experimental runs, the

attempt to differentiate the EF of LD vehicle and the HD vehicle by multi-regression method has failed. Thus, only the composite EF for all the on-road vehicles is reported in this paper. Table 1 provides the comparison of our EFs for CO, NOxand NMHCto EFs obtained

from other domestic experiments and experiments in other regions. The EF of CO obtained in this experiment ranges from 3.27 to 4.08 gm km1veh1with an average of 3.64 gm km1veh1. The EF for NOx ranges from

0.67 to 1.21, with an average of 0.90 gm km1veh1. The EF of NMHC(obtained by GC-FID measurement) and the EF of total sum of 56 VOCspecies (obtained by GC-MS measurement) are 0.44 and 0.24 gm km1veh1, respectively. Compared with the EFs from the Chung– Cheng tunnel experiments, their EFs of CO and NMHC are higher by factors of 1.8 and 3.4, respectively, while the EFs for NOxfrom both the experiments are about

the same. Reasons for these discrepancies are discussed as follows. Since the Chung–Cheng road tunnel is located at Kao-Hsiung city of southern Taiwan, it passes under a surface railway with a length of about 100 m. Thus the sloping effect (with road grade of about 76%) becomes significant. Furthermore, the entrance of Chung–Cheng tunnel is only about 30 m from a cross-road signal. We expect that the acceleration–decelera-tion driving modes of the passing vehicles may significantly increase the emissions of CO and NMHC. On the other hand, the inauguration of more stringent new car emission standard effective by 1999 might be another minor effect (Chung–Cheng tunnel experiment was carried out in 1997). All these may contribute to

R2 2 = 0. 53 R = 0. 17 55 60 65 70 75 80 85 90 4.1 4.3 4.5 4.7 4.9 Vehicle speed (km hr -1) 0 500 1000 1500 2000 2500 3000 Vehicle Volume Speed Volume Ventilation velocity (m s-1)

Fig. 2. Ventilation velocity and vehicle volume vs. speed. Table 1

Comparison of emission factors (g km1veh1) with other tunnel experiments

Tunnel CO NOx NMHC

Taipei (All vehicles) 3.6470.26 0.9070.18 0.4470.06

Chung–Cheng (All vehicles) 6.25 1.02 1.51

6.571.5(spao60) 1.9 70.5b(spao60) 6.071.8(spa>60) 1.070.5b(spa>60) Gubrist (LD) 4.1870.38 1.0570.09 0.4670.04c (HD) 1.1872.33 15.5670.79 0.2970.27c (All vehicles) 3.8970.64 2.4571.59 0.4470.06 Fort McHenry (LD) 3.9570.34 0.5070.06 0.3970.06d (HD) 6.1171.75 8.9770.28 0.9670.28d (All vehicles) 4.63 2.060.48 Tuscarora (LD) 3.0470.30 0.2470.16 0.1870.04d (HD) 3.7571.00 12.0970.53 0.4270.12d (All vehicles) 3.63 3.81 0.30 a Speed of vehicle (km h1). b21 Species of NMVOC.

ct-VOC, to be calculated by regression.

(6)

smaller EF for CO and NMHC from our experiment. Compared to EFs by Gubrist tunnel experiment (Switzerland), the values of CO and NMHC from our experiment are quite similar, while their EF of NOxis a

factor of 2.7 higher than the result from the Taipei tunnel experiment. The comparison to the results from the Tuscarora Tunnel experiment on the Pennsylvania Turnpike and the Fort McHenry Tunnel experiment in Baltimore finds a similar conclusion about EFs of CO and NMHC. The difference is concentrated on the values of EF of NOxfrom different regions. To explore

the reason, it is found that the fraction of HD vehicles (diesel-fueled) on the experimental day’s traffic was extremely low. It only occupied 2–6% of the total traffic volume. The fraction of HD vehicles in the total traffic volume, by the Gubrist experiment was averaged as 24% (workday average), the Tuscarora experiment was 6–80% (average 18%), and the Fort McHenry experiment was 4–73% (average 17%), is significantly larger than ours. This fact causes a smaller NOx

emission factor of total on-road traffic volume from Taipei tunnel experiment. It is unclear, but is speculated

that the special date we chose might be of bias concerning the on-road HD vehicle’s population. Further study regarding the on-road fleet’s character-istics is needed.

Table 2 shows the EF of different VOCcompounds. On the mass basis, the top three abundant VOCin Taipei Tunnel are toluene, ethene and 1,2,4-TMB. The averaged EF for them are 29.0274.95, 26.2374.89 and 14.2872.94 mg km1veh1, respectively. With little

speculation, the characteristic VOCs emissions may directly reflect the specific formula of gasoline or diesel fuel provided by local refinery industry.

It can be seen that toluene’s emission factor is significantly higher in Taipei tunnel than in Gubrist tunnel in Switzerland. The aromatic compounds like benzene, xylene and toluene, have been measured in significant concentrations by Hsieh et al. (1999). Other reasons like the hot-soak emission or the climate effect and so on, seem unlikely in explaining the measured VOCs spectrum in Taipei tunnel and other places in Taiwan. Concerning the EF of olefin compounds, they are recognized as the most reactive species in ozone

Table 2

Measurements of EF of Taipei tunnel and Gubrist tunnel study (mg km1veh1)

Code Species Taipei (all) Gubrist (GAV)a Code Species Taipei Gubrist

A Ethane 4.2770.96 4.2970.89 AC3-Hexene(cis) 0.3670.11 —

B Ethene 26.2374.89 24.1476.12 AD Cyclopentane, methyl 2.6470.93 — CPropane 2.4070.77 0.1571.15 AE PENTANE,2,4-dimethyl 0.4470.07 — D Propene 11.6172.04 13.9373.81 AF Benzene 12.2173.26 10.3872.34 E Acetylene 11.5673.02 12.8373.22 AG Cyclohaxne 0.9870.13 — F Isobutane 4.5770.94 1.7170.95 AH Hexane,2-methyl b — G n-Butane 6.5671.96 9.7075.32 AI Pentane,2,3-dimethyl b 8.6474.02 H 1-Butene 8.2771.55 1.9270.61 AJ Hexane,3-methyl 2.9470.43 — I trans-2-Butene 1.6170.38 1.4470.57 AK 2,2,4-Trimethyl-pentane 0.2970.18 — J cis-2-Butene 1.8470.46 1.2670.53 AL Heptane 1.4670.24 0.9370.44

K 3-Methyl-1-Butene 0.6770.13 0.7270.30 AM Cyclohaxne, methyl 0.9470.18 — L Isopentane 12.5074.09 18.2277.28 AN Pentane,2,3,4-trimethyl b — M Pentene 1.6170.38 0.6170.39 AO Toluene 29.0274.95 16.0274.84 N Pentane 9.5273.05 6.1674.45 AP Heptane,2-methyl 1.0570.29 — O Isoprene b AQ Heptane,3-methyl 1.02 70.33 — P trans-2-Pentene 2.7670.77 1.2270.75 AR Octane 1.3170.29 0.2270.23 Q cis-2-Pentene 1.5970.43 0.7870.48 AS Ethylbenzene 5.8871.55 3.670.91 R 2-Butene,2,2-dimethyl 3.0971.05 — AT m; p-Xylene 8.9572.38 10.7873.01 S Butane,2,2-dimethyl 1.3270.25 4.5372.14 AU Styrene 4.8171.33 — T Cyclopentene 0.4770.12 — AV o-Xylene 7.8872.14 4.7770.56 U 1-Pentene,4-methyl b AW Nonane 0.5470.13 0.0770.11 V Cyclopentane 0.8970.21 — AX Benzene,1-methyl-1-ethyl 1.0970.44 — W Butane,2,3-dimethyl 1.3370.69 — AY a-Pinene b

X Pentane,2-methyl 5.2771.72 — AZ Benzene, propyl- 1.6870.60 0.6870.29

Y Pentane,3-methyl 6.3971.53 — BA b-Pinene b —

Z 1-Pentene,2-methyl 0.8370.29 — BB Benzene,1,3,5-trimethyl 2.3170.38 1.5170.38 AA Hexane 4.1871.56 1.7370.56 BCBenzene, 1,2,4-trimethyl 14.2872.94 4.5971.09

AB 2-Hexene 0.9370.10 — BD 1,3-Butadiene 2.5670.38 1.6170.21

aGasoline vehicles, were calculated by multi-regression model. bBelow the detection limit—not reported.

(7)

formation, and are in par with the findings in other tunnel experiments.

3.4. Comparison with EF models

EF as determined from our tunnel experiments at each run are listed in Table 3 in order to compare with the corresponding EFs obtained by M5b and MT2.0. The ratios of measured EF to that by the model are also plotted in Fig. 3. In general trend, the models tend to underestimate the CO emission, while they overestimate both the NOxand NMHCemissions. On average, M5b

underestimates the measured CO emission factor by 10%, while it overestimates the NOx by 40%, and

NMHCby 20%. The MT2.0 also underestimates the measured CO emission factor by 20%, while it fairly estimates the measured emission factor of NOx and

NMHCwith slight run-to-run variations. Furthermore, the experiment run composite ratios of CO/NOx and

NMHC/NOxare also listed in Table 3. The same ratios

calculated by M5b and MT2.0 with parameters perti-nent with each run condition are also listed for comparison.

From the amplitude of deviations from its mean, the ratio provides the information of model sensitivity for their response to real condition. It can be seen that both models are of less sensitivity to respond to the real world change as compared to the measured ratios. We speculate that the changes like the driving pattern change due to traffic congestion and the on-road fleet composition change could alter the real EF to a larger extent. In short, it must be very careful in using the emission model to estimate the real world’s traffic emission, especially when sitewise or diurnalwise traffic emissions are needed. We suggest that it might be an interesting topic when talking about the ozone episode. Whether it is caused by meteorology factors or the city traffic congestion factors needs to be explored.

3.5. VOCs composition and fraction

The speciation of VOCs was obtained from the analysis of canister samples. Profiles of VOCs at the entrance and exit sampling sites are shown in Fig. 4. The most abundant three compounds observed, in both entrance and exit, were ethene, toluene and acetylene.

Comparing the percentage of entrance with exit, ethene and acetylene were increased but toluene was decreased. Observed VOCs emission profiles (as weight percen-tage of total VOCs) are shown in Table 2. The most abundant compounds were toluene, ethene, 1,2,4-TMB, isopentane, benzene, propene and acetylene; each of the former two compounds over 10%, each of the latter four compounds over 5%. The weight composition in terms of the carbon bond classifications is paraffins 28%, olefins 33% and aromatics 39%; with the number of carbon being C2–C4 34%, C5–C6 30% and C7–C9 36%.

The results showed that EF of CO, toluene and 1,2,4-TMB were less relative to vehicle speed, their correlation coefficient (R) was –0.259, 0.233 and –0.220, respec-tively. But that of 1,3-butadiene was 0.647; this means the EF of 1,3-butadiene increased with increasing vehicle speed. The R of NOxand vehicle speed was 0.844, which

means EF of NOx increased with increasing vehicle

speed.

The relations between CO and benzene were less relative (R ¼ 0:424), or were CO and 1,3-butadiene

Table 3

Model-predicted emission rates (g km1veh1) and CO/NOx, NMHC/NOxratios from the Taipei tunnel

CO NMHC NOx CO/NOx NMHC/NOx Observed 3.6470.26 0.4470.06 0.9070.18 4.2370.94 0.5170.13 MOBILE5b 3.2470.14 0.5370.01 1.2870.13 2.5670.34 0.4270.04 Mobile-Taiwan2.0 2.7270.09 0.4470.01 0.8870.10 3.1570.43 0.5170.06 CO 0.6 1.0 1.4 NOx 0.0 1.0 2.0 NMHC 0.4 1.0 1.6 0 1 2 3 4 5 6 7 8 9 10 11 Run No. Ratio Ratio Ratio MB5/obs MT2/obs

Fig. 3. Run ratios of model-predicted and observed emission factors for CO, NOxand NHMC, (MB5: MOBILE5b, MT2:

(8)

(R ¼ 0:177). But for benzene and toluene, they are more relative to each other (R ¼ 0:78). However, the vehicle speed ranged from 63.6 to 70.2 km h1, more data should be collected to make the evidence clear. 3.6. Ozone formation potential

In order to evaluate the relative importance of individual VOCspecies concerning their ozone forma-tion potential, the Maximum Incremental Reactivity (MIR) scale (Carter, 1994) for individual compound measured in Taipei tunnel is used. With the known VOCs profile measured by canister GC-MS method during this study, ozone formation potential is compiled to the form of ozone formed per vehicle per kilometer traveled. Their total is 1.02 gm O3km1veh1.

On the basis of 100 gm VOCs emitted, the ozone formation is 427 gm (from Section 3.3, average emission rate of total VOCs is 0.24 g km1veh1) which is similar to the LD fleet results of Tuscarora (409 gm) and of Ft. McHenry (403 gm). Also, individual species is grouped into paraffin, olefin and aromatic compounds to get the group lump-sum ozone formation potential. The results are shown in Table 4. It can be seen that the olefin

compounds contribute most to the potential ozone formation, i.e. 53.9%. The aromatic compound group contributes in less but significant percentage, 38.7%. Paraffin is least in their contribution in ozone formation, regardless of their obvious amount of emission. The most important compounds and their reactivity to contribute ozone formation are listed in Table 5. Top three towere ethene, 1,2,4-TMB, and propene, each contributing over 100 mg O3km1veh1.

Due to the significant number of two-wheel motor-cycles used by local commuters in almost every Taiwan urban areas, the emission difference needs to be explored. Furthermore, about half of the motorcycles

** 0 5 10 15 20 25 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z AA AB 0 5 10 15 20 25 A C A D A E A F A G A H A I A J A K A L A M A N A O A P A Q A R A S A T A U A V A W A X A Y A Z BA BB BC BD *Compounds list in Table 2 ** The I on each bar means ±

*

% (by vol.)

% (by vol.)

Entrance Exit

Fig. 4. Observed VOCs profiles of entrance- and exit-sampling site. Table 5

Major VOCs compounds of ozone formation potential (per 100 g VOCs emitted)

Species g O3 Species g O3

Taipeia Zefunb Lishinb Taipeia Zefunb Lishinb

Ethene 81.6 — — o-Xylene 21.5 29.3 51.3 1,2,4-TMB 52.8 14.4 34.5 1,3-Butadiene 11.7 — — Propene 45.9 — — trans-2-pentene 10.2 — — Toluene 32.9 126.1 98.7 1,3,5-TMB 9.8 115.5 170.2 1-Butene 30.9 — — Ethylbenzene 6.7 7.5 10.4 m; p-Xylene 27.9 58.8 81.6 Benzene 2.2 3.9 3.9

aCalculated from the EFs of this study.

bCalculated from each VOCs% (wt.) in NMHC (mg m3as C).

Table 4

Reactivity of VOCs (mg O3veh1km1)

Paraffins Olefins Aromatics Overall

Max 91.3 661.5 494.5 1218.4

Min 57.8 425.4 280.4 812.4

Ave 75.2 547.6 392.7 1015.5

Std 12.0 87.9 61.6 143.9

(9)

in Taiwan are powered by the two-stroke engine, which is notorious for its heavy VOCs emission. Table 5 also provides the comparison among the characterizing ozone potential by different aromatic compounds’ emission profiles from highway and usual town-way traffic. Since the traffic fleets of Zefun tunnel and of Lishin tunnel are characterized by combined passenger cars and motorcycles, their emissions must reveal the distinct difference from the highway tunnel’s emission. All the highways or freeway in Taiwan do not allow access by to two-wheeled vehicles. It can be seen that the significant differences in ozone formation are due to toluene, xylene and TMB.

4. Summary and conclusion

Real-world EF have been obtained for CO, NOx,

NMHCand individual VOCspecies from a tunnel experiment characterizing the usual highway emissions in northern Taiwan area. The results are compared with the traffic emission models like USEPA MOBILE5b (M5b), and the modified version MOBILE-TAI-WAN2.0 (MT2.0), in order to evaluate their applic-ability in Taiwan. A comparison with the results from other prominent tunnel experiments carried out in a different region/country was also made and discussed. The presentation of VOCs-profile obtained in our experiment reveals the emission characteristics of on-the-highway traffic fleet of Taiwan. When compared to the town-way tunnel experiments in the same northern Taiwan, the VOCprofile reveals distinct characteristics. The two-stroke engine used on motorcycles, which only allowed accessing the local town-way road, causes significant aromatic compounds’ emission. Specific findings and conclusion are listed below.

* EF of CO and NMHC obtained from our freeway

tunnel experiment are fairly close to the data by Gubrist tunnel Experiment, Fort McHenry tunnel Experiment and the Tuscarora tunnel Experiment. However, the emission factor of NOxis at least of

factor 2 less than other experiments. When the results from Chung–Cheng tunnel Experiment are com-pared, their EF for CO and NMHC are observed to be higher. This reveals that Taiwan town-way’s driving pattern with more acceleration/deceleration mode may increase the CO and NMHC’s emission.

* The emission factor model, MT2.0, predicts fairly

well the real-world emissions in an overall average sense, although it underestimates the CO emission by 20%. Using the experimental run-to-run parameter input, both models become less sensitive in the output of emission amount, as compared to the experimental results. This fact suggests that MT2.0 or M5b might not be able to give a realistic emission

change due to the city congestion effect or the on-road traffic fleet composition variation.

* The aromatic compounds are abundant in the Taipei

Tunnel’s emission in addition to the olefin species, ethene. Among them are toluene and 1,2,4-TMB, and their emissions are significantly higher than the findings from other regions/countries. It suggests that the aromatic compounds are actually the main gasoline additive or high fraction components in the fuel provided by local refinery industry.

* Using MIR scale for the ozone formation potential,

different VOCspecies are rated for their ozone formation per vehicle per kilometer traveled. The average traffic emission of total VOCs from Taipei Tunnel Experiment produces 1.02 gm ozone per vehicle per kilometer traveled. Among this, ethene, 1.2,4-TMB and propene contribute over 100 mg O3

per vehicle per kilometer traveled.

Acknowledgements

We thank the Environmental Protection Administra-tion of Taiwan in sponsoring this study. The authors would like to express their appreciation of Dr. S.C. Wu of NTU, R.C. Chang of CPC Co., Sinotech Engineering Consultants LTD., CTCI Co., TERTEC Co., Freeway Control Center, C.B. Pen and Z.H. Lin for helping this study and experiment successfully.

References

Ahlvik, P., Eggleston, S., Gorissen, N., Hassel, D., Hickman, A.-J., Joumard, R., Ntziachristos, L., Rijdeboer, R., Samaras, Z., Zierock, K.-H., 1997. COPERT II Computer Program to Calculate Emissions from Road Transport— Methodology and Emission Factors, European Environ-ment Agency, Denmark.

CARB (California Air Resources Board), 1990. Proposed regulations for low-emission vehicles and clean fuels, Staff Report and Technical Support Document. State of Cali-fornia Resources board, Sacramento, CA.

CARB, 1996. Methodology for estimating emissions from on-road motor vehicles, Vol. 2 EMFAC7g, California Air Resources Board, Technical Support Division, Sacramento, CA.

Carter, W.P.L., 1994. Development of ozone reactivity Scales for volatile organic compounds. Journal of the Air and Waste Management Association 44 (7), 881–899.

CTCI Co., 1999. Environmental Protection Administration, EPA-88-FA31-03-1059 (in Chinese).

Hsieh, C.C., Chang, K.H., Kao, Y.S., 1999. Estimating the ozone formation potential of volatile aromatic compounds in vehicles tunnels. Chemosphere 39, 1433–1444.

Hsu, Y.C., Tsai, J.H., Chen, H.W., Lin, W.Y., 2001. Tunnel study of on-road vehicle emissions and the photochemical potential in Taiwan. Chemosphere 42 (3), 227–234.

(10)

Ingalls, M.N., 1989. On-road vehicle emission factors from measurements in a Los Angeles area tunnel. The Air and Waste Management Association’s 82nd Annual Meeting, Paper 89–137.3, Anaheim, California.

John, C., Friedrich, R., Staehelin, J., Schlapfer, K., Stahel, W.A., 1999. Comparison of emission factors for road traffic from a tunnel study (Gubrist tunnel, Switzerland) and from emission modeling. Atmospheric Environment 33 (20), 3367–3376.

National Research Council, 1991. Rethinking the ozone problem in urban and regional air pollution, National Academy Press, Washington, DC.

Pierson, W.R., Gertler, A.W., Robinson, N.F., Sagebiel, J.C., Zielinska, B., Bishop, G.A., Stedman, D.H., Zweidinger, R.B., Ray, W.D., 1996. Real-world automotive emissions— summary of studies in the Fort McHenry and Tuscarora mountain tunnels. Atmospheric Environment 30 (12), 2233– 2256.

Sagebiel, J.C., Zielinska, B., Pierson, W.R., Gertler, A.W., 1996. Real-world emissions and calculated reactivities of organic species from motor vehicles. Atmospheric Environ-ment 30 (12), 2287–2296.

Staehelin, J., Schlapfer, K., Burgin, T., Steinemann, U., Schneider, S., Brunner, D., Baumle, M., Meier, M., Zahner, C., Keiser, S., Stahel, W., Keller, C., 1995. Emission factors from road traffic from a tunnel study (Gubrist tunnel,

Switzerland). Part I: concept and first results. Science of the Total Environment 169, 141–147.

Staehelin, J., Keller, C., Staheh, W.A., Schlapfer, K., Steine-mann, U., Burgin, T., Schneider, S., 1997. Modelling emission factors of road traffic from a tunnel study. Environmetrics 8, 219–239.

Staehelin, J., Keller, C., Stahel, W., Schlapfer, K., Wunderli, S., 1998. Emission factors from road traffic from a tunnel study (Gubrist tunnel, Switzerland) Part III: results of organic compounds, SO2 and speciation of organic exhaust

emis-sion. Atmospheric Environment 32 (6), 999–1009. Tsai, J.H., Hsu, Y.C., Chen, H.W., Lin, W.Y., 1998. Motor

vehicle emission factors of NMHC, CO, NOx and

VOCs Pattern in photochemical reactions inside the tunnel, The Air and Waste Management Association’s 91st Annual Meeting & Exhibition, paper 98-MA4A.02, San Diego, CA.

USEPA, 1997. Compendium method TO-15: determination of volatile organic compounds (VOCs) in air collected in specially prepared canister and analyzed by Gas Chromato-graphy/Mass Spectrometry (GC/MS), Center for Environ-mental Research Information, Office of Research and Development, U.S. EPA, Cincinatti, OH, USA.

Winther, M., 1998. Petrol passenger car emissions calculated with different emission models. Science of the Total Environment 224, 149–160.

數據

Fig. 1. Sampling locations and placement of equipments of Taipei Tunnel experiment.
Fig. 2. Ventilation velocity and vehicle volume vs. speed.
Table 2 shows the EF of different VOCcompounds.
Fig. 3. Run ratios of model-predicted and observed emission factors for CO, NO x and NHMC, (MB5: MOBILE5b, MT2:
+2

參考文獻

相關文件

Strategy 3: Offer descriptive feedback during the learning process (enabling strategy). Where the

Now, nearly all of the current flows through wire S since it has a much lower resistance than the light bulb. The light bulb does not glow because the current flowing through it

There are existing learning resources that cater for different learning abilities, styles and interests. Teachers can easily create differentiated learning resources/tasks for CLD and

• Thresholded image gradients are sampled over 16x16 array of locations in scale space. • Create array of

The quantity of landscape planting in campus was evaluated and used as an indicator to divide the campus into different landscape zones where the air anions

Shinar, "Effects of an in-vehicle collision avoidance warning system on short- and long-term driving performance," Human Factors, vol. Abdel-Aty Mohamed, “Investigating

Service quality, customer satisfaction and customer loyalty are also different while people used Taipei Smart Card in different situation.. We suggest that the Taipei Smart

Due to the high technology change and rigorous global competition environment, the dynamic customer preference needs and short product lifecycle intensify the