02FFL-183 Multi-Objective Optimization of Diesel Engine Emissions and Fuel Economy using Genetic Algorithms and Phenomenological Model T. Hiroyasu, M. Miki, J. Kamiura, S. Watanabe Doshisha University H. Hiroyasu Kinki University Copyright c  2002 Society of Automotive Engineers, Inc. ABSTRACT In this paper, the simulation of the multi-objective opti- mization problem of a diesel engine is performed using the phenomenological model of a diesel engine and the ge- netic algorithm. The target purpose functions are Specific fuel consumption, NOx, and Soot. The design variable is a shape of injection rate. In this research, we empha- size the following three topics by applying the optimiza- tion techniques to an emission problem of a diesel engine. Firstly, the multiple injections control the objectives. Sec- ondly, the multi-objective optimization is very useful in an emission problem. Finally, the phenomenological model has a great advantage for optimization. The developed system is illustrated with the simulation examples. INTRODUCTION Because of the merit of the durability and fuel efficiency, a diesel engine is loaded on from small to large vehi- cles. However, with increasing environmental concerns and legislated emissions standards, current engine re- search is focused on simultaneous reduction Soot and NOx during maintaining reasonable fuel economy. The combustion improvement especially can be achieved de- signing a good injection system and characteristics of spray combustion. To develop a good injection system, a parameter search to determine the influence an organization performance and an exhaust performance should be performed. However, when this parameter search is executed experimentally, the huge expense and huge time are needed. For this reason, the optimization of parameters by simulation on a computer is very useful. When the parameter is optimized by the simulation, the minimization of the fuel efficiency, the amounts of the ni- tric oxide (NOx), and the amounts of the soot have been interested in many engine designer[1, 2, 3]. Therefore, these NOx, Soot and fuel efficient become objective func- tions in optimization problems. There are some studies that solve this optimization problem[4, 5, 6, 7]. However, these problems are treated as single objective problems. Since there are trade-off relationships between the fuel ef- ficiency, NOx and Soot, it is natural to handle these prob- lems as Multi-objective Optimization Problems (MOPs). In this research, the minimization of fuel efficient, the amounts of NOx, and the amounts of Soot are simultane- ously performed by using the concept of multiple-purpose optimization. To perform optimization by simulations, an optimizer (it determines the next searching point) and an analyzer (it evaluates the searching point) are needed. The process of the combustion of the diesel engine is very complicated. At the same time, there are many require- ment items for the models such as injection character- istics, spray characteristics, air-fuel mixing, ignition, heat release rate, heat losses, exhaust emissions, and so on. Thus, it is almost impossible to build the model of diesel combustion with the numerical expressions. On the other hand, several types of the models of diesel combustion have been proposed[8]. Those are roughly divided into three categories; thermodynamic model, phenomenologi- cal model and detailed multidimensional model. The ther- modynamic model only predicts the heat release rate. In the phenomenological model, the prediction of equa- tion which is derived by the fundamental experiment is used. The detailed multidimensional model predicts sev- eral items by solving differential equations with small time steps. In this paper, the HIDECS which is based on the phenomenological model is used since it does not need a 1