Radial basis function and particle swarm optimization to predict range extender engine performance

Abstract

Range extender engine is one of the potential technologies to develop in future. However, this technology still has performance and emission problem. To solve this problem, a new technology model and optimization method are needed. Therefore, for this purpose, the radial basis function and particle swarm optimization are used in the investigation. In this study, two types of radial basis function (RBF), Cauchy and Gaussian, were used to construct the prediction model of fuel consumption and range extender engine emission. Both RBF types were compared with one another to decide on which one is the best to predict the model. By using data from a two-cylinder 999 cc gasoline engine, generator, battery, electric motor and other vehicle components, a range extender electric vehicle (REEV) model simulation was built. Based on this simulation result, some output data will be taken as training set data to build the prediction model of fuel consumption and emission and some output data will be taken to test this prediction model in MATLAB. Moreover, particle swarm optimization (PSO) was used to calculate some control parameters of range extender engine to optimize fuel consumption and emission based on the best model. The result shows that the radial basis function is successfully used to develop the prediction model of some range extender engine control parameters. The Cauchy type radial basis function has better accuracy than Gaussian type radial basis function. Moreover, based on the model, PSO method is able to calculate control parameters efficiently to optimize evaluation item based on the model.