The present control method for combustion parameters such as spark ignition timing of a car petrol engine is feedforward control, which is called the 'table looking-up' method. It is interesting to replace the conventional table looking-up method with the neural networks method, which is a new technique for explaining such nonlinear characteristics as combustion parameters, and to investigate its advantages and disadvantages. Reported here is control of combustion parameters by neural networks, and especially the learning characteristics of the networks when the control is adopted as on-line control, i.e. adaptive control. As a result, it was found that neural networks are more suitable for on-line control for car use than the table looking-up method. (A).
Abstract