摘要
故障预测技术是电子装备预测与健康管理(PHM)领域的核心内容,对电子装备关键部件实施有效的预测是保证系统正常运行的关键。首先将灰色理论和人工神经网络算法相结合,构建灰色神经网络模型并对其进行分析;然后在此基础上通过附加动量变学习速率法对灰色神经网络的权值更新策略进行改进,提出一种基于改进灰色神经网络的故障预测模型;最后以某型脉冲测量雷达中频接收组合中的压控振荡器为例,以采集的原始频率数据为基础进行仿真验证。预测结果表明,将该预测方法应用于电子装备PHM是行之有效的,可有效提高故障预测精度。
Fault prediction technology is the core content of electronic equipment PHM, carrying out effective prediction on the key components of electronic equipment is the guarantee of system running in normal operation. Firstly, this paper built and analyzed the general grey neural network model by combining grey theory and artificial neural network. Then improved the weight updating strategy of grey neural network by the method of additional momentum and variable learning rate, and put forward a fault prediction method based on improved grey neural network model. Finally, it took a voltage controlled oscillator （VCO） of the intermediate frequency combination in a certain pulse instrumentation radar as an example, and the collected original frequency data as the basis to simulate. The results show that applying the prediction method to electronic equipment PHM can effectively improve the fault prediction accuracy.
出处
《计算机应用研究》
CSCD
北大核心
2013年第12期3625-3628,共4页
Application Research of Computers
基金
国家"十一五"装备预先研究项目(51317030104)
国家自然科学基金资助项目(60771063)
关键词
故障预测
预测与健康管理
灰色神经网络模型
附加动量变学习速率法
改进灰色神经网络
fault prediction
prognostic and health management （PHM）
grey neural network model
additional momentum and variable learning rate method
improved grey neural network
