基于改进BPNN的电子化指挥系统故障诊断

Fault diagnosis for electronic command system based on improved BPNN

基于BP神经网络和案例推理模型,建立导弹电子化指挥系统故障诊断专家系统的基本结构,并给出一个具体的应用实例。首先对导弹电子化指挥系统的故障机理进行分析,在此基础上研究基于BP神经网络和案例推理相结合模型的故障诊断技术的应用,确定了基于改进BP神经网络的故障诊断专家系统的设计思想和实现方法,并建立了基于BP神经网络和案例推理结合模型的导弹测控系统故障诊断专家系统,弥补了单一故障诊断推理模型的不足。通过对某型导弹电子化指挥系统中语音系统的故障诊断实例,说明了该故障诊断专家系统的可行性。

The paper sets up the fundamental structure of a fault diagnosis expert system for the mobile missile electronic command system (MMECS) based on combining the BP neural network (BPNN) and the case-based reasoning (CBR) model and puts forward a concrete implementation simultaneously. Firstly, the fault mechanism for the MMECS is analyzed in great detail. Secondly, the application of the fault diagnosis technique based on BP neural network reason model is studied. Finally, the design concept and the corresponding realization method of the system are presented. A fault diagnosis expert system for the MMECS based on BP neural network and CBR reason model is developed. The developed expert system make up for the shortcomings in the conventional single inference model. The developed expert system is applied to the voice interrupt in the MMECS and proved to be feasible by the diagnosis results.