基于RBF神经网络的压气机特性仿真

为克服传统仿真方法误差较大的问题，提出了一种基于RBF（多变量插值的径向基函数）神经网络的压气机特性仿真方法。利用RBF神经网络能逼近任意非线性系统的特点，对压气机特性进行了拟合。试验结果表明，此方法具有精度高，收敛速度快等优点，可广泛运用于发动机数值仿真及控制模拟等领域。

Compressor characteristic simulation based on RBF neural network

To overcome the weakness of traditional method, a simulation method of compressor in aeroengine based on RBF neural network was developed. According to the character of RBF neural network that it can approach any nonlinear system, compressor characteristic was simulated. The result implies that this method has high precision and fast convergence pace and it can be applied to many fields such as aeroengine numeral simulation, control simulation and so on.