- 上一篇：铁路计轴器检测系统的设计+电路图
- 下一篇：PCA的TE过程故障检测系统设计
摘 要在工业生产过程中，生产设备损耗和故障会对产品质量造成影响，严重时甚至会导致重大生产安全事故。传统依靠人力的监测方法已经不能够适用于目前复杂的生产过程。目前各种化工过程的故障检测大多数采用基于主元分析（Principal Component Analysis，简称PCA）的故障检测方法。但基于主元分析的故障检测方法是线性方法，它并不能够在非线性系统中有效。
为了跟上日趋复杂的生产过程故障检测的要求，本文的研究是基于核主角知识，依据TE过程的过程数据，同时在阅读了大量参考文献的基础上，对基于核主角（Kernel Principal Angle ，简称KPA）的故障检测方法进行研究。69576
本文将通过以下几方面进行研究，（1）对传统的主元分析故障检测方法进行简单描述（2）讲述核主角故障检测方法的主要内容：①构建特征子空间、②核主角测量。（3）通过MATLAB对TE过程故障检测仿真实验证明基于核主角的故障检测方法能够有效地应用在非线性系统中。
该论文有图8个，表4个，参考文献33个
毕业论文关键词：核主角 特征子空间 故障检测 主元分析
Development of Tennessee Eastman Process Fault Diagnosis System Based on Kernel Principal Angles
Abstract In the process of industrial production,if the production equipment loss and failure,will affect the quality of the products, it even can lead to major production safety accidents seriously. Traditional methods depend on human monitoring could not apply to the current complex industrial production process in today.most of the process of chemical industrial use the fault detection methods based on Principal Component Analysis,(abbreviation :PCA) . But the fault detection method based on principal component analysis is a linear method, it is not necessarily that it can effective use in the nonlinear systems.
In order to keep up with the requirement of increasingly complex production process monitoring, this article based on Kernel Principal Angle,( abbreviation KPA) fault detection methods were studied which based on Kernel Principal Angles ,the leading role knowledge, on the basis of the Tennessee Eastman process - the process of data, and on the basis of reading a large number of references,
This article will through the following aspects to introduce the fault detection method, (1) the traditional principal component analysis of the fault detection methods are simply described. (2) the KPA main fault detection method is the main content of: ① building feature subspace, ② KPA measurement. (3)Through MATLAB simulation experiment proves that the fault detection method based on KPA can effectively used in nonlinear system by TE fault procession
This paper has 8 pictures, 4 tables, 33 references
Key words: Kernel Principal Angles Feature Subspace Fault Detection Principal Component Analysis
目录
摘 要 I
Abstract .II
目 录 ..III
图清单 . ..V
表清单 V
1 概述 1
1.1 课题研究背景及意义