Simulation Study On Sliding Mode Control Of Idle Speed For Automotive Engine

Idle speed is important working condition of gasoline engine.And idle speed control is one of the most important control of gasoline engine control.The problem on how to use sliding mode control in the idle speed control of gasoline engine is discussed in this thesis.At first,some advance control strategies at home and abroad about idle speed control of gasoline engine are overviewed.And base theories of sliding mode control and neural network are introduced.Reductionistic transfer function model of idle speed is insufficient,so this paper selects high precision's gasoline engine mean value model of Denmark technology university. The error of this model is only 2~3% at whole operation fields.And this model has the same presicion when same gasoline engine uses difference intake manifold system and fuel injection system.Steady cooperating points of gasoline engine are obtained by simulation at different loads.These datas are used to off-line discipline BP neural network.Then BP neural network model is created under SIMULINK.So BP neural network is used to calculate steady cooperating point of idle speed in control simulation on-line.Then,small deflection method is used to linearize this mean value model,and simulation model of gasoline engine idle speed is established under SIMULINK.Based on these works,this paper designs three control methods to decrease or eliminate inhere chattering of sliding mode control:(a)sliding mode control based on the reaching law,(b)RBF neural sliding mode control,(c)equivalent sliding mode control based on RBF neural network.Controllers are designed.Simulation systems are created.Sparking advance angle and idle speed vavle duty cycle are selected ascontrol variables, (b) is simulated under SIMULINK, (c) is simulated under MATLAB. (a) is simulated under SIMULINK on different conditions: empty load , six types of typical load,air fuel ratio and sparking advance angle adding random noise.This paper summarizes simulation results and shows that sparking advance angle only appropriates to small load and idle speed vavle duty cycle appropriate to big load,and three control methods have good robustness , (b) and (c) have higher control precision than (a) and can eliminate chattering of sliding mode control.But (b) and (c) have some deficiencies,such as bad dynamic quality,long transitional time,big excessive regulation because adjusting weight value.center and width of basis function for RBF neural network is timeconsuming.So these should be further studied in the future work.At last,this paper selects sliding mode control based on reaching law.Stateflow is used to create idle speed control simulation system of gasoline engine under SIMULINK.lt realizes to automatically select sparking advance angle and idle speed vavle duty cycle as control variables of sliding mode control in terms of load's magnitude.The simulation results show that idle speed mean error is less than 0.6 rpm ,standard deviation equals to 1.721 rpm ,fluctuant amplitude is less than 1 rpm .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.There are 66 references,61 figures and 17 tables in this paper.