摘要
针对目前单一诊断方法的缺陷,提出了一种基于主元分析方法(PCA)和神经网络的组合故障诊断方法。利用多PCA来监测雷达状态,在雷达状态异常时再通过分析变量贡献率来确定故障部位,最终采用神经网络进行故障诊断。仿真实验了该方法的有效性。
To overcome the shortcomings of single diagnosis method, a combined fault diagnosis approach based on principle component analysis （PCA） and neural network （NN） is presented. The approach use PCA to monitor radar state and identify fault location by variable contributions. At last, neural network is utilized for fault diagnosis. Case study of the proposed method is presented for radar, and the feasibility of method is analyzed.
出处
《科学技术与工程》
2009年第16期4836-4839,共4页
Science Technology and Engineering
关键词
故障诊断
雷达
主元分析
神经网络
fault diagnosis radar principle component analysis neural network
