基于PCA的传感器网络的故障诊断分析

Monitoring and Fault Diagnosis based on PCA for Sensor Networks

主成分分析属多元统计方法,正逐步成为控制领域中一种重要的数据处理方法,用于生产监测和质量控制。本文简要地介绍了PCA中两种常用的图形分析法——Q图和主元得分法,利用统计软件——SPSS对数据进行处理,简化了复杂的运算过程,并对其数据处理过程进行了说明。最后,通过空压机远程监控系统传感器网络的实例模型,运用SPSS软件,说明了这一数据处理方式的简便、有效性和缺陷。

Principle component analysis is based on multi-statistics method, and turning into an important way to proceed data for production monitoring and quality control. The two significant analysis method in PCA--Q graph analysis and Principle component score graph analysis, are presented briefly. It is convenient and efficiency that utilizing the software SPSS to proceed the data and analyze the system performance. It predigests the complex intrinsic data process and clarities the meaning. IN the last part of the article, the example of the air compressor long-distance monitor system is used to illustrate the convenience and efficiency of PCA utilizing SPSS, at the same time its lacuna is extruded.