Developing Pollutant Emission Models for Diesel Public Vehicles in Tehran

Transcription

1 Developing Pollutant Emission Models for Diesel Public s in Tehran Mansour. Hadji Hosseinlou and Aidin.Massahi Abstract This paper presents diesel public vehicles models based on vehicle s speed for main public vehicles s including CO, NOx, PM 10, VOC, CH 4, HC and SO 2.The traffic emissions are calculated in three route s slope s for horizontal, downhill and uphill which are modeled based on an instantaneous emission model integrated with a macroscopic traffic simulation model. The emission models are based on empirical measurements. Public vehicle emissions relate to the vehicle, the instantaneous speed and vehicle s route slope s. Integrated emission models are applied for diesel public vehicles, which are classified into three category including urban buses, rural buses and minibuses so after presenting emission models, accuracy of them are evaluated by empirical measurements. Some models which are not good compatible with empirical data, have been changed by describing new emission form and new models validity are evaluated by empirical data. Consequently, by considering correlation coefficients can be evaluated new emission models are valid. Keywords Instantaneous emission, integrated emission model, correlation coefficient, speed. I. INTRODUCTION hate makes most people anxious about improvement Wtechnology is its destroying aspect about our environment, therefore nowadays humankind not only attempts to conduct our life towards more comfortable but also he tries to decrease environment pollution effects, which are brought about by technology. There is a strong relationship between environment extinction and technology. Had we realize the significance our environment destroying process, we would have deliberated on it much more meticulously. According to Stephen William Hacking s Theory [13], world become extinct by destroying human environment because humankind cannot repair his environment so we should be aware of our environment s. Air pollution plays a crucial role in destroying our environment, it is undeniable. Statistics indicate 3 billions of world s people were dead by air pollution effects in 2011[12], therefore air pollution are taken into account one of the most challenging problems in the world. Nowadays, cars are considered as the most important source of air pollution[8], so this paper presents cars air pollution emission models for F. A. Assistant professor, Civil Department, K.N.Tossi University of Technology, Tehran, Iran ( Mansour@Kntu.ac.ir). S. B. Post Graduate Student, Civil Department, K.N.Tossi University of Technology, Tehran, Iran ( A.massahi@sina.kntu.ac.ir). Tehran s diesel public vehicles, which are classified into urban buses, rural buses and minibuses. This paper submits air pollution emission models for 7 air s including CO, NO x, PM 10, VOC, CH 4, HC and SO 2 in three different street slope s including uphill, downhill, horizontal for diesel public vehicles. Moreover, paper calculates correlation coefficients for all of emission models. After that a new model form are developed for previous emission models that their correlation coefficients do not equal one and correlation coefficients are calculated for these new models. The paper is structured as follows. Section 1 contains an introduction of our research for creating emission models. Section 2 describes requirements and parameters, which are related to emission models. Section 3 presents the method for creating Tehran s emission models, Section 4, presents correlation coefficients for accuracy of this model, so driven data of this emission models are compared with achieving data which are measured by Tehran air quality control company[10], ultimately new form emission models are presented for models are not good compatible with actual data. Finally, Section 5 presents the conclusions. II. REQUIREMENTS AND PARAMETERS FOR MODEL Traffic generated air pollution is a great concern to the general public [3]. Motor vehicles emit nitrogen oxides (NO x ), carbon monoxide (CO), volatile organic compounds (VOC), particulate matter (PM 10 ), Methane (CH 4 ), sulfur dioxide (SO 2 ) and hydrocarbon (HC) which have constituted major sources of air pollution in large cities such as Tehran [14]. Traffic generates greenhouse gases, which consist of reaction hydrocarbons with nitrogen oxides, may contribute to global warming [8]. In this research, minibus, urban bus and rural bus emission rates are achieved for seven s in three-slope s: horizontal, uphill, downhill also these data are measured for 20 (Km/h), 40 (km/h), 60(km/h) and 80 (Km/h).Table 1 depicts urban bus emission rates in different speeds and slopes [10]. According to Table 1 data, emission models should be related to average speed of minibus, urban bus and rural bus and slope s. Speed is in the terms of road s, volume of cars in their routs link, also emission rates have a proportional relation with slope. It means emission rates in uphill is more than downhill and horizontal. In addition, emission rates have a none 472

2 linear relation with vehicle speed because emission rates increase at low speeds and high speeds, therefore emission rate are minimum between these speeds. It should be noted that emission rates in table 1 are obtained in standard temperature and pressure. III. METHOD FOR DEVELOPING EMISSION MODEL After obtaining emission data from Tehran s Air Quality Control Company [10] and by mentioning section two, we analyze different function forms for developing an acceptable form, therefore it has been deduced that polynomial functions are the best form for emission rates. According to table 1, emission rates are measured for 4 speeds 20, 40, 60 and 80 Km/h, thus air pollution emission model form consists of 4 parameters a, b, c, d, also by considering emission rate in high speeds and low speeds has been concluded, cars emit minimum emission rate between these speeds, therefore emission models are represented as second order polynomial functions. Model form is shown in below: 2 d ER( mg ) = a + bv + cv + d (1) km v Where ER is emission rate of s in terms of milligram per kilometer and a, b, c, d are models coefficients and V is speed in terms of kilometer per hour (Km/h). According to table1, emission rates are got for 20, 40, 60, 80 (km/h) in three different slope s. On the other hand, four parameters a, b, c, d have been applied for this model form. If we do not use any limited for this model form, emission models correlation coefficient (R 2 ) would be equal to 1, why that calibration process leads to solving a simultaneous system of four equations with four unknown variables and all coefficients could be uniquely calculated. This model form has been selected to determine emission rate because of follow reasons: 1-Pollutant Emission rates have a non-liner relation with speed and in the determined speed, it becomes to minimum because of using polynomial function. 2-Emission rate for zero speed will be infinite because diesel generate infinite emission rate that why they operate in fixed location. Because of this reason d parameter divide on speed, (d/v), so in low speed, emission rate approaches toward infinity. 3-Emission values are reported for four speeds, so estimating a model more than four parameters will not be reasonable because models correlation coefficients will be below 1. 4-In these models d parameter should not be negative ( because negative d leads to reduction emission rates for speeds below 20 Km/h and not only they are not compatible with actual emission rates between 0 to 20 Km/h but also there is little differences between achieving data by air quality control company and calculated data by these models in the range above, so whenever d would be negative, It is omitted from the model and model form is converted to follow model: d ) 2 ER ( mg ) = a + bv + cv (2) km It should be noted in this case correlation coefficient will be below one (R 2 <1) because it leads to solving simultaneous system of four equations with three unknown variables. Tables 2, 3, 4 depict s emission models respectively for urban bus, rural bus and minibus in different slope s, which are developed with Matlab software. IV. METHOD DEFINING OF NEW MODEL FORM FOR MODELS WITH CORRELATION COEFFICIENTS BELOW 1 Pollutants emission model correlation coefficients are shown in tables 2, 3 and 4, by observing these correlation coefficients can be concluded, whenever d parameter is omitted not only correlation coefficient is under one, but also models do not approach toward infinity, when speed approaches zero and these models are not compatible with actual cars emission rate. speed 20 Km/h 40 Km/h 60 Km/h 80 Km/h TABLE I PRIVATE CARS EMISSION RATES IN DIFFERENT SPEEDS AND SLOPES IN TERMS OF MILLIGRAM PER KILOMETER. [10] NOx CO PM 10 VOC SO 2 CH 4 HC Horizontal Uphill Downhill Horizontal Uphill Downhill Horizontal Uphill Downhill Horizontal Uphill Downhill

3 TABLE II POLLUTANT EMISSION MODELS COEFFICIENTS FOR URBAN BUS IN DIFFERENT SLOPE CONDITIONS a b c d R 2 Urban NOX Horizontal CO Horizontal PM10 Horizontal VOC Horizontal SO2 Horizontal CH4 Horizontal HC Horizontal NOX Uphill CO Uphill PM10 Uphill VOC Uphill E-10 1 SO2 Uphill CH4 Uphill HC Uphill NOX Downhill CO Downhill PM10 Downhill VOC Downhill SO2 Downhill CH4 Downhill HC Downhill Rural TABLE III POLLUTANT EMISSION MODELS COEFFICIENTS FOR RURAL BUS IN DIFFERENT SLOPE CONDITIONS a b c d R 2 NOX Horizontal CO Horizontal PM10 Horizontal VOC Horizontal SO2 Horizontal CH4 Horizontal HC Horizontal NOX Uphill CO Uphill PM10 Uphill VOC Uphill SO2 Uphill CH4 Uphill HC Uphill NOX Downhill CO Downhill PM10 Downhill VOC Downhill SO2 Downhill CH4 Downhill HC Downhill In figur 474

4 es1, 2, 3 and 4 some s emission models are polluted by using equations (1) and (2) for urban bus, rural bus and minibus. Figures 1 and 4 indicate emission models, which are described by equation (2) and figures 2, 3 demonstrate emission models, which are applied by equation (1). Results of emission rates in some of emission models which are created by equation(2), are not acceptable when speeds are increased above 80 km/h because not only emission rates do not increase but also these rates move toward zero, even in very high speeds this rate would be negative that is not logical. On the other hand, correlation coefficients are near 1 in the most of the cases. It can be conclude emission rate models with omitting d parameter can be acceptable in ranges between 20 kilometer per hour (Km/h), and 80 kilometer per hour (km/h), but emission rates cannot be acceptable in speeds below 20 kilometer per hour (Km/h) and above 80 kilometer per hour (km/h). By considering all above mentioned discussions, this research uses equation (2) for models with correlation coefficients below 1, so these models are calculated emission rates for V=10 Km/h. By using of five emission rates for speeds including 10, 20, 40, 60 and 80 Km/h, a new model form is described, which is expressed in equation 3: Fig.1 CH 4 emission model diagram was estimated by presented model for minibus in downhill slope in contrast with reported emission rate (r 2 =0.9963). Where ER is emission rate of s in terms of milligram per kilometer and a, b, c, d and e are model s coefficients and V is speed in terms of kilometer per hour (Km/h). mg 2 3 ER ( ) = a + bv + cv + dv + km e v (3) Minibus TABLE IV POLLUTANT EMISSION MODELS COEFFICIENTS FOR MINIBUS IN DIFFERENT SLOPE CONDITIONS a b c d R 2 NOX Horizontal CO Horizontal PM10 Horizontal VOC Horizontal SO2 Horizontal CH4 Horizontal HC Horizontal NOX Uphill CO Uphill PM10 Uphill VOC Uphill SO2 Uphill CH4 Uphill HC Uphill NOX Downhill CO Downhill PM10 Downhill VOC Downhill SO2 Downhill CH4 Downhill HC Downhill

5 Fig.2 PM 10 emission model diagram was estimated by presented model for rural bus in downhill slope in contrast with Fig.4 CO emission model diagram was estimated by presented model for minibus in uphill slope in contrast with reported emission rate (r 2 = ). 1-Pollutant Emission rates have a non-liner relation with speed and in the determined speed, it becomes to minimum because of using polynomial function. 2-Emission rate for zero speed will be infinite because diesel vehicles generate infinite emission rate that why they operate in fixed location. Because of this reason e parameter divide on speed, (d/v), so in very low speed, emission rate approaches toward infinity. 3-Emission values are reported for five speeds, so estimating a model more and less than five parameters will not be reasonable because models correlation coefficients will be below 1. Fig.3NO x emission model diagram was estimated by presented model for urban bus in horizontal slope in contrast with This model form has been selected for determine emission rate because of follow reasons: 4-In these models e parameter should not be negative ( because negative e leads to reduction emission rates for speeds below 10 Km/h and it is not compatible with actual data and vehicles emission rate theory. Nevertheless, fortunately all the new emission models have a positive e parameter so it can be concluded correlation coefficients (R 2 ) are equal to one for all models. e ) By solving a simultaneous system of five third order equations with five unknown variable, model s coefficients are obtained. TABLE V NEW POLLUTANT EMISSION MODELS COEFFICIENTS FOR URBAN BUS IN DIFFERENT SLOPE CONDITIONS a b c d e R 2 Urban CO Horizontal VOC Horizontal THC Horizontal CO Uphill THC Uphill CO Downhill VOC Downhill THC Downhill

6 In this new form, not only models correlation coefficients (R 2 ) is equal to 1 for previous models, which did not have correlation coefficients equaling to 1 but also e coefficient does not have any negative values in all models, so this form gives us acceptable values of emissions in all ranges between 20 until 80 km/h, below 20 km/h and above 80 km/h. It means in lower speeds than 20 km/h and higher speeds than 80 km/h, emission rates have been increased. Tables 5, 6 and 7 demonstrate models coefficients for s had correlation coefficients below 1 in pervious emission form. Figures 5, 6, 7and 8 some of new emission models are plotted. Considering these figures it can be concluded that these models are Valid in all ranges. For example by comparing figures 1 and 8 Can be concluded not only new emission model s rates for speeds containing 20, 40, 60 and 80 are exactly the same with measurement data for these speeds but also emission rates approach toward infinity in low speeds. Fig.5 CO emission model diagram was estimated by presented new model for minibus in horizontal slope in contrast with TABLE VI NEW POLLUTANT EMISSION MODELS COEFFICIENTS FOR RURAL BUS IN DIFFERENT SLOPE CONDITIONS a b c d e R 2 CO Horizontal Rural VOC Horizontal CH 4 Horizontal THC Horizontal CO Uphill VOC Uphill CH 4 Uphill THC Uphill VOC Downhill CH 4 Downhill THC Downhill TABLE VII NEW POLLUTANT EMISSION MODELS COEFFICIENTS FOR MINIBUS IN DIFFERENT SLOPE CONDITIONS a b c d e R 2 CO Horizontal Minibus SO 2 Horizontal CH 4 Horizontal THC Horizontal CO Uphill VOC Uphill SO 2 Uphill CH 4 Uphill THC Uphill CO downhill VOC downhill SO 2 downhill CH 4 downhill THC downhill

7 Fig.6 VOC emission model diagram was estimated by presented new model for rural bus in downhill slope in contrast with Fig.8 CH4 emission model diagram was estimated by new presented model for minibus in downhill slope in contrast with Presented models are first comprehensive models in Tehran, which can measure emission rates for main diesel vehicles s. These models can estimate dynamic emission rates for different s in urban streets, consequently they can estimate emission rates in urban networks. Moreover, the most of emission models, which were submitted for enormous cities in the world, cannot estimate emission rate for wide range of speeds.therefore the most of them are divided to several small ranges of different speeds but these presented models can estimate emission rates in wide range of speeds. Specially, they get us relatively precise values in low and high speeds. Fig. 7 HC emission model diagram was estimated by presented new model for urban bus in uphill slope in contrast with V. CONCLUSION In this report emission rates are presented for 7 s: CO, NO x, PM 10, VOC, SO 2, CH 4 and HC in three slope s: uphill, downhill, horizontal for three classes of diesel vehicles: urban bus, rural bus and minibus after that emission models have been evaluated and correlation coefficients are calculated for these models. Some models are not compatible with actual emission values so new emission form is presented for them and correlation coefficients are calculated for them. Ultimately, it demonstrates our models can present same emission rates rather than extracted data by air quality control company for speeds: 20, 40, 60, 80 Km/h. REFERENCES [1] Luc Int Panis, Steven Broekx, Modelling instantaneous traffic emission and the influence of traffic speed limits, Elsevier, Aug.2006, pp [2] LUO Qingyu, JUAN Zhicai, SUN Baofeng1, JIA Hongfeil Method Research on Measuring the External Costs of Urban Traffic Congestion, Elsevier, October 200, pp [3] Litman, T., Transportation Cost Analysis: Techniques, Estimates and Implications, Victoria Transportation Policy Institute, June [4] David J. Forkenbrock, External costs of intercity truck freight transportation,elsevier,1999. [5] Tanjima Pervin, Ulf-G Gerdtham and Carl Hampus Lyttkens Societal costs of air pollution-related health hazards: A review of methods and results, [6] Liping Xia, Yaping Shao, Modeling of traffic flow and air pollution emission with application to Hong Kong Island, Elsevier,2005, pp [7] Matthias Schmidt, Ralf-Peter Schafer, An integrated simulation system for traffic induced air pollution, Elsevier,1998, pp [8] Tehran s Air Quality Control Company site, ( [9] Iran statistic organization site,( [10] Tehran s Air Quality Control Company, s report, [11] Matlab (R 2008a) software help. [12] Hamshahri Newspaper, introduce with managing director of Tehran s Air Quality Control Company NO [13] Wikipedia, the free encyclopedia. [14] SHERIF ARIF The Energy Environment Review (EER) in The Islamic Republic of Iran (IRI) and in The Arab Republic of Egypt(ARE), (The World Bank)