Modeling and Forecasting Engine Air Induction under Transient Condition Based on Neural Network

Precise measurement of the air induction flow is the basis of accurate control of air fuel ratio for gasoline engines. However during transient conditions, the serious fluctuation of air induction state and the lagging response of the airflow sensor seriously affect the accuracy of air fuel ratio control. In this paper, the characteristic of air induction flow under transient condition is analyzed and a method of airflow forecast under transient conditions based on transient parameter information and BP neural network is presented. Meanwhile, the topological structure of BP neural network is also established and the model is trained and simulated by using experiment data in the acceleration and deceleration condition of gasoline engine. The results show that this method can accurately forecast the engine induction airflow under transient condition and can eliminate the lagging characteristic of the airflow sensor.