PrincipalComponentAnalysisMechanicalFaultPrediction.PDF

Proceedings of Annual Conference of China Institute of Communications Principal Component Analysis：Mechanical Fault Prediction Based on Data-Driven Technique 1 2 Luhui LIN , Jie MA School of Automation, Beijing Information Science Technology University, Beijing, China E-mail:163.com; mjbeijing@163.com Abstract: This paper describes the technique based on data-driven. This method does not require a priori knowledge of the process, solely on the basis of process data collected, and can explore the hidden information in the data. And principal component analysis (PCA) is one of the analysis methods of the data-driven. This paper aims to propose a fault prediction method based on PCA which is a multivariate statistical analysis method. It does not need to understand the structure and principles of the system in-depth and can be based on the sensor measurement data of the system running process entirely; and the PCA algorithm, which is simple and easy to implement, can also reduce the noise interference and simplify the data processing. Keywords: principal component analysis (PCA); data-driven; fault prediction 主元分析：基于数据驱动的机电设备故障预测方法 林禄辉 1，马 洁 2 北京信息科技大学自动化学院，北京，中国，100192 E-mail:163.com; mjbeijing@163.com 摘 要: 本论文介绍的是基于数据驱动的方法，其不需要过程的先验知识，仅仅以采集的过程数据为 基础，而且还能够发掘数据中的隐藏信息；主元分析方法(Principal Component Analysis, PCA)就是其中 的一种分析方法。旨在提出基于 PCA 这种多元统计分析的故障预测的方法，这种方法可以不需要对系 统的结构和原理有深入的了解，完全基于系统运行过程中传感器的测量数据