SA Control Using Neural Network Based On MFB

Abstract: Sparkle advance angle control is one of the most important control of gasoline engine control; SA control accurately involves dynamical and economic of engine, is also a challenging subject. Aiming to this subject,the author develops RBF(radial basis function) neural network to solve the problem of SA control in gasoline engine; And then the author emulates it on-line under SIMULINK softwork platform on mean engine model.The author begins its paper with influences of SA to engine,with the necessity and importance of SA control;synthetically analyze the control strategy of MFB and its background, researches on SA control that have been done both domestically and abroad.on the basis of it,the author adopts neural network to be the controller of SA.Furthermore,the author analyzes these neural networks that have been used in control area respectively and get the result that RBF neural network is a very good choice for SA control of gasoline engine.Knowledge which is related to neural network is introduced in this paper.The gasoline engine model which was brought up by professor Elbert Hendricks and his colleague who works in Denmark technology university is adopted in this paper.This engine model has very high precision and its standard error is only 2～3% in whole working range for three different engine models. It has the similarly precision level to the same engine adopting different manifold and fuel injection system. So it is adopted widely.After engine model and its control strategy have been established, the author compiled S_function of RBF NN controller;so the SIMULINK emulation on-line model based on S_function is accomplished. During its simulation,the controller's adaptability is tested by adding random noise,changing load parameter;the result shows RBF NN has a good dynamical character. after that,the former knowledge about RTW is introduced based on SIMULINK model.In the end, the achievements and some problems in the research are summarized in the dissertation, and some suggestions for the future study are provided.