Study On Prediction Of Gasoline Intake Air Flow Based On CMAC Neural Network

With the world’s energy crisis and environmental pollution becoming worse and worse, there are more harsh law to restrict emissions from vehicle,the development of the automotive industry has faced more challenges, but the car is more significant for on people’s lives and the social and economic development, it also provides greater opportunities. Therefore, improving efficiency of fuel combustion and the cleaning performance of catalytic converters on gasoline has become an important research direction for thee researchers who researching the field of an internal combustion engine, there also is a huge economic and social benefits.Air-fuel ratio is a crucial factor for improving performance of gasoline engines,it affects purification effect in the cylinder and the cleaning efficiency of catalytic converters. Therefore, the more precise control of air-fuel ratio,the better condition where gasoline engines work in, in order to improve vehicle power, and reduce the emission of pollutants and the cost on driving. However, in transient conditions, as the dynamic effects in the intake process and the air flow sensor response delay characteristics and error, result in making the intake air flow measured values is not collect, lead to considerable error, the quantity of fuel injection that calculated by the intake flow will be large errors, it result in greater shocks in air-fuel ratio. Therefore,in order to enhance the effect of air-fuel ratio control, it is necessary to ensure that the measured intake air flow with sufficient accuracy.After the analysis of the intake air flow measurement error-depth, propose forecasting method of the CMAC neural network, described its structure, algorithms and works in detail. Then, making some improvements in its algorithm and architecture. Through a complex function tests the reliability and validity of the improved model. Then,in transient conditions, using an improved CMAC neural network to predict the intake air flow rate, establishing the intake flow projection model based on CMAC neural network in Matlab simulation environment, While combine the intake air flow forecasting model with the mean model. Finally, make the whole simulation experiment under two representative transient conditions.It come to some conclusions through simulation experiments: the gasoline engine intake air flow forecasting model established in this paper is able to accurately predict the intake air flow, with fine accuracy, and high efficiency, reliability, which enhance the effectiveness of the air-fuel ratio control.