Abstract: The calibration of the electronic control diesel engine needs to consider comprehensively factors such as fuel consumption, emission and intake and exhaust temperature. However, with the implementation of the China Ⅴ emission regulations, the difficulty of calibration is increased exponentially. The traditional calibration method needs huge work and high cost. On the basis of model calibration, a point by point model calibration method based on important working points is proposed. Through cluster analysis and analysis of the actual emission characteristics, the important working conditions in the NEDC cycle are selected. Experimental design of the working points is completed by the space filling method, and engine test data is collected by the bench test. On the basis of the response model, the optimal combustion parameters combination at each working point is obtained by genetic algorithm. MAP mapping is completed according to the cubic polynomial fitting. MAP is optimized based on the universal characteristic test, and the vehicle drum emission test is carried out simultaneously. The results show that the MAP generated by the calibration model has good fuel economy and better emission performance, which can accurately predict the engine response parameters, with the fuel consumption 15% lower than that of artificial calibration.

Key words: diesel calibration, working points selection, design of experiment, response model, genetic algorithm