基于神经网络的非线性电子器件的建模方法

Neural Network Method for Modeling Nonlinear Electron Devices

针对当前电路仿真中复杂非线性电子器件的建模问题,提出了先由神经网络学习器件的输入输出特性,再将得出的网络结构用Pspice中的电路描述语言描述的建立非线性电子器件模型的新方法,并通过实验验证了这种方法的有效性。这一方法可作为一种通用方法,用于其它器件的建模。

Aiming at the modeling problem of complicated nonlinear electron devices in circuit simulation, a method, which describes the architecture of a neural network in Pspice is proposed. The architecture of the neural network can be achieved by learning the input and output data of the device which is intended to be modeled. The efficiency of this method is proved by experiments. This method can be used as a general way to model other devices.