Research On Model-based Calibration Of Common Rail Deisel Engine

Due to common rail deisel engine’s functionality of controlling injection pressure,injection quantity,injection timing and injection rate independently and flexibly,reducing emssions and noise,it has become one of the hot spots in development of electronic controlled deisel engine. While with long work cycle and large workload,traditional calibration methods have been difficult to adapt to rapid development of deisel engine.Model-based calibration technique is a good choice.Model-based calibration process mainly includes determination engine inputs and outputs and levels, Design of Experiment,engine bench test according to Design of Experiment, establishing engine model based on bench test results,optimizing control map offline based on engine model.For the main part of the model-based calibration technology,the main work of this paper is as follows.In order to improve the efficiency of test bench experiment,test data from different test equipment was intergrated and a platform was further developed with functions of automatic scanning calibration and key parameters recording.The factorial experiment design,space-filling design and optimal design were re-searched.Results show that sapce-filling design is more match with engine bench test.According to different sensitivities of various control parameters,a two-stage response model was established.A polynomials one-stage model with main injecton advance angle as local input and torque,exhaust temperature and fuel consumption as outputs,a RBF two-stage model with engine speed,main injection volume and rail pressure as inputs was established. The two-stage model was proved with high accuracy and good generalization ability through experiment.The main injection advance angle was obtained by using genetic algorithm based on the two-stage model.Evaluation result on the MAP via engine bench test shows advantages of Model-based calibration technique.