摘   要随着人民生活水平的提高，为满足人民生活，生产业和制造业都在飞速的发展以满足人民的需求。人们对生活品质追求的不断提高，这就要求生产的精度不断增加，因此工业生产系统的复杂程度也越来越高，为了保证生产过程的安全有效，减少故障的发生率，及时检测出故障是最重要的。在过程检测中采集大量的数据，对这些数据利用算法进行处理得到生产系统的状态，从而及时的了解系统的状态，对出现的故障进行处理。因此，故障检测的地位日趋显著。

本文是利用主元分析（PCA）方法对化工过程田纳西-伊斯曼过程（TEP）进行故障检测。TEP是典型的工业生产过程，它包含多种变量，这些变量具有相关性。

本文首先介绍多元统计方法，主元分析；再介绍了基于主元分析的故障检测，利用SPE 和 统计量来了解系统的状态，得到SPE 和 统计图，贡献图来进行故障诊断的方法；接着介绍基于主元分析的过程故障检测,介绍了TE过程的工作流程；最后利用仿真研究，通过21组预设定故障对TE过程进行仿真研究。69577

该论文有图5幅，表4个，参考文献25篇。

毕业论文关键词：TE过程  主元分析  故障检测

Principal Component Analysis for Tennessee - Eastman Process Fault Detection

Abstract

With the development of people's living standard, in order to meet the people's life, production and manufacturing are developing rapidly to meet the needs of the people. With the improvement of people's the quality of life , this requires the production precision increases. Therefore, the complexity of the system in industrial production is also getting more higher. It is so important for guaranteeing the safety of the production process and reducing the fault incidence to timely detect the fault . In the process of testing, we collect a large amount of data, we use the algorithm to process the state of the production system, so as to know the state of the system and to deal with the failure of the system.

This article uses principal component analysis (PCA) method for the chemical process of Tennessee Eastman process (TEP) for fault detection. TEP is a typical industrial production process, it contains a variety of variables and these variables have a correlation.

Firstly,this essay describes the multivariate statistical methods, principal component analysis and then introduces the fault detection based on principal component analysis and statistics to understand the state of the system to give SPE and T² charts, contribution plot ; then introduced fault detection based on process fault detection,the working process  of TE introduces ; Finally, simulation studies, by the 21 pre-set faults on TE process simulation.

This article has 5 figures, 4 tables and 25 references.

Key words:Tennessee Eastman process  Principal component analysis  Fault detection

目   录

摘   要 I

Abstract II

目   录 III

图清单 IV

表清单 IV

1. 概述 1

1.1课题研究背景及意义 1

1.2过程故障检测的方法分类 2

1.3故障检测技术在国内外的发展史，研究现状以及发展趋势 2

1.4设计内容 4

2．主元分析 6

2.1主元分析的基本概念