汽機車污染排放關鍵影響因素之分析

Abstract

隨著臺灣經濟之成長、道路基礎設施的擴建，同時也使得機動車輛之使用愈來愈頻繁，且數量不斷增加。民國79年，臺灣地區車輛之登記數，汽車約為230萬輛，機車約為710萬輛;但至2007年，分別各增加為670萬與1400萬輛，其成長速度不容忽視。顯然地，這不僅導致都市與城際間之交通擁擠問題，甚至造成更嚴重之空氣污染。在許多都會地區，機動車輛所排放之廢氣甚至被視為空氣污染的主要來源。因此，相關主管機關如環保署便實施車輛定檢計畫，希望車輛污染排放能合乎標準，進而有效地減少空氣污染問題。 本研究回顧車輛污染相關文獻，大多根據車輛定檢資料來找出具有高污染性之車輛；但其僅限於對車輛特性之探討(即直接影響)，例如車齡、排氣量、燃油類型、手/自排車、廠牌、汽缸數與觸媒轉換器等，其他間接影響之變數如駕駛人之社經背景、使用行為與主要用車區域等亦會影響到碳氫化合物(Hydrocarbon, HC)與一氧化碳(Carbon Monoxide, CO)之排放。因此，本研究即分別針對汽、機車之污染進行探討，主要分為兩部分：(1)根據汽、機車之定檢資料，找出機動車輛HC與CO排放之直接影響(direct effect)因素(2)納入問卷資料，進一步分析污染排放之間接影響(indirect effect)因子。 在直接影響因素之探討方面，透過聯立迴歸(Simultaneous Regression)分析發現，汽車之車齡、排氣量、汽缸數、車重、行駛里程與廠牌對HC與CO排放均有顯著的影響;而機車方面之關鍵影響因子則為車齡、排氣量、行駛里程、二/四行程與廠牌。另外，為了得到駕駛人相關資訊，本研究進行全國家戶機動車輛問卷調查，共郵寄90000份，回收之有效份數為5981份。藉由車牌號碼之配對，將定檢與問卷資料串聯，並利用結構方程模式(Structural Equation Modeling, SEM)來分析車主社經背景、主要用車區域、車輛使用行為、車輛基本特性與污染排放濃度(HC、CO)之因果關係。結果顯示，車主社經背景與主要用車區域對車輛基本特性有顯著的影響，車輛基本特性又會進一步影響車輛使用行為;在污染排放濃度上，則會受到車輛使用行為與車輛基本特性之影響而有所不同。另外，在總影響方面，汽、機車之車輛基本特性構面，其對污染排放濃度之總影響均為-0.76；而汽車之車主社經背景構面對污染排放濃度之總影響為0.24，機車則為0.08。最後，根據以上分析結果，提出具體之車輛污染管制策略，以期望能減少對環境的衝擊。

The blossom of economic growth in Taiwan associated with the continuous construction of highway infrastructures for convenient movements of people and freights has inevitably brought a rapid growth of private vehicles over the past decades. In 1990, for instance, there were only 2.3 million cars and 7.1 million motorcycles registered; in 2007, the numbers of registered cars and motorcycles have increased to 6.7 millions and 14.0 millions, respectively, which were almost three times and double of the 1990 figures. The trend toward greater use of private vehicles has not only created ubiquitous congestion on the urban roads and intercity highways, but also resulted in excessive emissions that have broken the ecological balance. In metropolitan areas, the poor air quality has been mainly resulted from the mobile emissions of these vehicles. Thus, the Environment Protection Authorities at the central and local levels have endeavored to vehicle inspection and maintenance programs (I/M) and other control initiatives aiming to effectively restrain the air pollution from these mobile polluting sources. Many studies even attempted to identify the high-polluting vehicles based on the I/M database. However, only vehicle-related factors (i.e., direct factors) such as vehicle age, engine size, type of fuels, automatic/manual, brand, number of cylinders, catalytic converter, and gas mileage were recorded in the database. Other indirect factors such as drivers’ demographics, usage of vehicles, type of regions may also cause these vehicles high-polluting. To propose effective control strategies, a clear understanding of who, where and why these high-emitters are still in use is imperative. The objectives of this paper are twofold: (1) to identify the key direct factors significantly affecting the emissions of hydrocarbon (HC) and carbon monoxide (CO) of cars and motorcycles, according to the I/M database; (2) to further identify the key indirect factors making these vehicles so high-polluting. The direct factors significantly affecting the emissions of HC and CO can be identified by simultaneously regressing HC and CO, respectively, on the characteristics of car and motorcycle. It is found that the key direct factors significantly affecting car emissions are age, engine size, number of cylinders, weight, mileage traveled, and brand. In contrast, the key direct factors for motorcycle emissions are age, engine size, mileage traveled, 2-stroke/4-stroke, and brand. The profiles of some high-polluting vehicles are clearly identified, accordingly. Furthermore, this study also conducted a nationwide post-mail questionnaire survey on some 90 thousand car and motorcycle owners, with 5981 valid samples returned. The I/M dataset and the returned questionnaire dataset are matched via the vehicle license-plate number so as to scrutinize who, where, and why the identified high-emitters are still in use. Structural equations modeling (SEM) was then employed to test the hypothesized cause-effect relationships among five constructs: driver demographics, vehicle usage, regions, vehicle characteristics, and emissions. The results show that the constructs of driver demographics and regions have significant effects on the vehicle characteristics. The construct of vehicle characteristics has then significant influence on the construct of vehicle usage. Moreover, the direct factors causing the high emissions levels of HC and CO are the two constructs: vehicle usage and vehicle characteristics. In terms of total effect to the construct of emissions, the construct of vehicle characteristics has the highest total effect of -0.76 for both cars and motorcycles, followed by the construct of driver demographics with total effect of 0.24 for cars and 0.08 for motorcycles. Based on the direct/indirect factors found as well as their relationships identified, some relevant emission control strategies are proposed and discussed.