At present, there isn't a general way to model nonlinear electronic parts and the Backpropagation Network and the RBF network can approach any nonlinear and continuous function. This paper comes out based on the two facts.Firstly, this paper introduces some methods about modeling the nonlinear parts and discusses the advantages and disadvantages of the methods respectively. And then draw a conclusion that all of these methods can't be used as a general method.Secondly, the neural network, especially the BP network and the RBF network, is introduced.Finally, we get a new general method, which is based on the neural network, to model the non-linear electronic parts. And this paper introduces its general scheme and process. Comparisons of the merit and shortcoming between using these two modeling methods are presented.After a number of experiments, this scheme, approaching a function with neural network and describing the network with PSpice, is tested successfully. And then, we receive a general way for setting up the nonlinear electronic parts. This method will play an important role in circuit simulation. |