Modeling of Static Exhaust Emission Performance for Automotive Gasoline Engines Based on Neural Network

The static exhaust emission performance of automotive gasoline engines is influenced by a few factors such as engine's crankshaft rotate speed, load, air/fuel ration and spark timing, etc, the relationship between those factors and the engine exhaust emission performance is nonlinear. A BP neural network was developed to describe the relationships, and the neural network model was trained by engine test data. This model can be used for forecasting exhaust emission and controlling A/F, spark timing to decrease exhaust emission.