基于改进ICA的工业过程故障诊断研究

如今的工业过程系统结构复杂设备精密度高,随之而来的就是系统的高故障性,所以如何准确地检测到故障的发生已经成为一大难题。该文基于提高故障诊断性能的目的,提出一种DPCA-ICA的故障诊断方法。这种方法先采用DPCA对数据进行降维和去噪处理,得到能最大反映系统信息的低维数据,然后再通过ICA方法提取独立元,进行故障诊断。仿真结果表明,改进后的ICA故障诊断方法不仅具有比传统PCA方法更低的故障误报率,并且对一些PCA难以检测的故障也有很好的诊断效果。

Based on the improved ICA fault diagnosis of industrial processes

Nowadays , The structure of industrial process system is complex and the equipment is precise .So the possibility of fault is high .How to detect the fault exactly has become a major problem . Based on the purpose of improving the performance of fault diagnosis , we propose a DPCA-ICA method. The method first use DPCA for data dimensionality reduction and denoising processing to obtain the information which is largest low-dimensional and can reflect the system＆#39;s feature to a large extent , then we extract independent element through ICA , and use it for diagnosis .Simulation results show that the improved ICA method has a lower failure rate of false positives than the traditional PCA method , and have a good effect on detecting the complex fault .