液体导弹发动机马尔可夫链和RBF网络辨识方法研究

Identification Approach for Liquid Missile Engine Based on Markov Chain and RBF Neural Networks

分析了液体导弹发动机辨识方法 ,对基于马尔可夫链和RBF神经网络的液体导弹发动机的辨识原理和方法进行了较详细的叙述 ,依据某型液体导弹发动机实际测试数据 ,对其进行了仿真研究 ,并对仿真结果进行了比较分析。结果表明 ,基于动态系统马尔可夫链的液体导弹发动机的辨识方法优于基于RBF神经网络的辨识方法。

For improving identification approach of liquid missile engines, we deem it necessary to study the methods of identification approach. In this paper the two methods based on Markov chain and RBF neural networks are considered. We have derived eqs. (1) through (6) to describe the first order Markov model for the dynamic system, and then we describe the second order Markov model with eqs. (7) and (8). We also have derived eqs. (9) and (10) to describe the identification approach for aviation engine based on RBF neural networks. Figs.1 and 2 show the structure of first order Markov model, and Fig.3 shows a RBF neural networks. The real testing data of some liquid missile engine are used to analyse the two methods. Figs. 4,5,6,and 7 show the simulation results. The results illustrate that the identification approach on Markov model is more effective than that on RBF neural networks.