神經(jīng)網(wǎng)絡(luò)理論在機(jī)械故障診斷技術(shù)中的應(yīng)用


摘 要

隨著設(shè)備復(fù)雜程度的提高和自動(dòng)化水平的提高，設(shè)備故障診斷就顯得尤為重要，而選擇合適的診斷方法對(duì)于診斷結(jié)果是否精確就顯得更為重要。隨著計(jì)算機(jī)技術(shù)的普及，故障的智能診斷顯示出極大的優(yōu)勢(shì)。而神經(jīng)網(wǎng)絡(luò)技術(shù)以其信息的并行分布式處理在設(shè)備故障診斷中更顯其優(yōu)越性。本文就人工神經(jīng)網(wǎng)絡(luò)理論在機(jī)械故障診斷技術(shù)中的應(yīng)用進(jìn)行了系統(tǒng)而深入地研究。
首先，本文在系統(tǒng)介紹機(jī)械故障診斷技術(shù)的重要性和國(guó)內(nèi)外發(fā)展現(xiàn)狀、趨勢(shì)的基礎(chǔ)上，指出了人工神經(jīng)網(wǎng)絡(luò)理論應(yīng)用于機(jī)械故障診斷技術(shù)具有極大的應(yīng)用價(jià)值和發(fā)展?jié)摿ΑＭ瑫r(shí)也指出，我國(guó)目前人工神經(jīng)網(wǎng)絡(luò)理論在機(jī)械故障診斷領(lǐng)域的應(yīng)用還不完善，仍處于發(fā)展階段。
其次，在本論文中，對(duì)旋轉(zhuǎn)機(jī)械的故障特點(diǎn)知故障機(jī)理進(jìn)行了分析。介紹了神經(jīng)網(wǎng)絡(luò)的基本知識(shí)，并在此基礎(chǔ)上闡述了人工神經(jīng)網(wǎng)絡(luò)理論應(yīng)用于機(jī)械故障診斷技術(shù)研究的理論意義，指出人工神經(jīng)網(wǎng)絡(luò)獨(dú)有的結(jié)構(gòu)和特點(diǎn)為機(jī)械故障診斷技術(shù)注入了新的生命力，從而很有可能使機(jī)械故障診斷技術(shù)走出困境。

關(guān)鍵詞：旋轉(zhuǎn)機(jī)械 故障診斷 人工神經(jīng)網(wǎng)絡(luò)


Abstract

With the development of equipment and the improvement of automat level，the subject of equipment fault diagnosis is more important，and the selection of suitable diagnosis method to the results of diagnosis is very improved. With the prevalence of computer technology，the intelligence fault diagnosis system is more important the neural networks with the information proceeded juxtaposed have many advantages. The thesis focuses on the application of artificial neural network (ANN) theory in the mechanical fault diagnosis technology.
First，on the basis of systematically introducing the importance and the internal and external developing current situation、trend，the thesis points out the application of ANN theory in mechanical fault diagnosis has great applying value and developing potential. At the same time，the thesis also indicates that this application is still imperfect and is in developing period.
Next，in this article，the development of fault diagnosis technology, the fault diagnosis technology of rotating machinery and the application of neural network in the field of fault diagnosis have been discussed. On the basis of systematically introducing the importance and the internal and external developing current situation、trend，the thesis points out the application of ANN theory in mechanical fault diagnosis has great applying value and developing potential. At the same time，the thesis also indicates that this application is still imperfect and is in developing period.

Key words: rotating machinery fault diagnosis artificial neural networks





目 錄
任務(wù)書
摘要
1 緒論…………………………………………………………………………………1
1.1機(jī)械故障診斷技術(shù)的研究的意義………………………………………………..1
1.2國(guó)內(nèi)外機(jī)械故障診斷技術(shù)發(fā)展現(xiàn)狀……………………………………………..2
1.3機(jī)械故障診斷技術(shù)發(fā)展趨勢(shì)……………………………………………………..3
1.4基于神經(jīng)網(wǎng)絡(luò)的故障診斷研究與應(yīng)用現(xiàn)狀……………………………………..3
1.4.1基于BP神經(jīng)網(wǎng)絡(luò)的故障診斷…………………………………………………4
1.4.2存在的問題……………………………………………………………………...4
1.5本課題的主要工作………………………………………………………………..5

2 旋轉(zhuǎn)機(jī)械主要故障機(jī)理分析………………………………………………………6
2.1旋轉(zhuǎn)機(jī)械故障的來源及主要原因………………………………………………..6
2.1.1在設(shè)計(jì)和制造過程中引起的故障……………………………………………...6
2.1.2由于安裝維修不當(dāng)引起的故障………………………………………………...6
2.1.3運(yùn)行操作不當(dāng)引起的故障……………………………………………………...6
2.1.4由于機(jī)器劣化造成的故障……………………………………………………...6
2.2旋轉(zhuǎn)機(jī)械的振動(dòng)基本特性………………………………………………………..6
2.2.1轉(zhuǎn)子渦動(dòng)………………………………………………………………………...7
2.2.2 臨界轉(zhuǎn)速………………………………………………………………………..8
2.3幾種常見故障的故障機(jī)理分析…………………………………………………13
2.3.1轉(zhuǎn)子不平衡…………………………………………………………………….13
2.3.2不對(duì)中………………………………………………………………………….15
2.3.3油膜渦動(dòng)和油膜振蕩………………………………………………………….20

3 人工神經(jīng)網(wǎng)絡(luò)的基本原理及其學(xué)習(xí)算法………………………………………..21
3.1 引言……………………………………………………………………………...21
3.2 人工神經(jīng)網(wǎng)絡(luò)的生物原型—生物神經(jīng)元……………………………………...21
3.3人工神經(jīng)網(wǎng)絡(luò)模型………………………………………………………………22
3.3.1人工神經(jīng)網(wǎng)絡(luò)的基本處理單元—人工神經(jīng)元（節(jié)點(diǎn)）…………………….22
3.3.2激勵(lì)函數(shù)……………………………………………………………………….24
3.3.3人工神經(jīng)網(wǎng)絡(luò)模型…………………………………………………………….25
3.4人工神經(jīng)網(wǎng)絡(luò)的學(xué)習(xí)（訓(xùn)練）…………………………………………………26
3.4.1有指導(dǎo)的訓(xùn)練………………………………………………………………….26
3.4.2無指導(dǎo)的訓(xùn)練………………………………………………………………….26
3.4.3　誤差反向傳播學(xué)習(xí)算法（BP算法）………………………………………27
3.5 人工神經(jīng)網(wǎng)絡(luò)的特點(diǎn)…………………………………………………………...30

4人工神經(jīng)網(wǎng)絡(luò)在旋轉(zhuǎn)機(jī)械故障診斷技術(shù)中的應(yīng)用………………………….…..32
4.1引言………………………………………………………………………………32
4.2 故障診斷的人工神經(jīng)網(wǎng)絡(luò)模式識(shí)別方法…………………………………...…32
4.3 人工神經(jīng)網(wǎng)絡(luò)在旋轉(zhuǎn)機(jī)械故障診斷中的應(yīng)用研究………………………...…33
4.3.1故障診斷神經(jīng)網(wǎng)絡(luò)模型結(jié)構(gòu)………………………………………………….33
4.3.2 機(jī)械故障診斷神經(jīng)網(wǎng)絡(luò)的訓(xùn)練參數(shù)…………………………………………34
4.3.3 故障診斷神經(jīng)網(wǎng)絡(luò)的訓(xùn)練……………………………………………………35
參考文獻(xiàn)……………………………………………………………………………..37
致謝…………………………………………………………………………………..39



參考文獻(xiàn)
【1】周繼成，周清山，韓飄揚(yáng)等．人工神經(jīng)網(wǎng)絡(luò)――第六代計(jì)算機(jī)的實(shí)現(xiàn)．科學(xué)普及出版社
【2】黃文虎，夏松波，劉瑞巖等．設(shè)備故障診斷原理、技術(shù)及應(yīng)用．科學(xué)出版社
【3】 閻平凡，張長(zhǎng)水．人工神經(jīng)網(wǎng)絡(luò)與模擬進(jìn)化計(jì)算．清華大學(xué)出版社