摘 要

隨著現(xiàn)代工業(yè)的發(fā)展，大型旋轉(zhuǎn)機(jī)械對國民經(jīng)濟(jì)越來越重要，是許多大型企業(yè)的重點(diǎn)關(guān)鍵設(shè)備，因此，對大型旋轉(zhuǎn)機(jī)械的故障診斷的重要性日益顯著，而選擇合適的診斷方法對于診斷結(jié)果就顯得極為重要。隨著計算機(jī)技術(shù)的普及，基于神經(jīng)網(wǎng)絡(luò)的故障智能診斷顯示出極大的優(yōu)勢。量子神經(jīng)網(wǎng)絡(luò)將量子力學(xué)的思想引入到神經(jīng)網(wǎng)絡(luò)之中，以其強(qiáng)大的量子并行運(yùn)算和聯(lián)想能力非常適合旋轉(zhuǎn)機(jī)械的故障診斷，具有可行性和優(yōu)越性。
本文首先系統(tǒng)介紹了旋轉(zhuǎn)機(jī)械故障診斷技術(shù)的重要性和國內(nèi)外發(fā)展現(xiàn)狀，并針對旋轉(zhuǎn)機(jī)械振動的特點(diǎn)，進(jìn)行了典型的故障特征分析。然后研究了BP神經(jīng)網(wǎng)絡(luò)和量子神經(jīng)網(wǎng)絡(luò)（QNN）的基本原理、模型機(jī)構(gòu)和算法，比較了兩種網(wǎng)絡(luò)的優(yōu)缺點(diǎn)，選擇了以多層激勵函數(shù)的量子神經(jīng)網(wǎng)絡(luò)作為旋轉(zhuǎn)機(jī)械故障診斷的方法。最后，在MATLAB開發(fā)環(huán)境下建立了QNN算法程序，并對某熱電廠汽輪機(jī)的故障數(shù)據(jù)進(jìn)行了測試和仿真，和人工神經(jīng)網(wǎng)絡(luò)的診斷相比較，結(jié)果表明量子神經(jīng)網(wǎng)絡(luò)具有更加良好的識別性能和準(zhǔn)確率，具有很好的應(yīng)用前景。


關(guān)鍵詞：量子神經(jīng)網(wǎng)絡(luò)，故障診斷，故障特征，QNN算法















ABSTRACT

With the development of modern industry, large rotating machinery is more and more important on national economy, which is the key equipments of many large companies，therefore, fault diagnosis of rotating machinery is more and more significant. So choosing fit fault diagnosis method seems to be more important to the result. Along with the computer technology popularization, the breakdown intelligent diagnosis based on Quantum Neural Network demonstrates the enormous superiority. Quantum Neural Network introduces the thought of Quantum Mechanics into the Neural Network, with its strong ability of parallel computation and association, it is very suitable to diagnose the fault of rotating machinery, and it’s feasible and superior.
First，this article system introduction rotating machinery fault diagnosis technology importance and domestic and foreign development present situation, and in view of the characteristic of rotating machinery fault feature, introduces the characters of rotating machinery faults and the mechanism of rotating machinery. Then, the basis principle、the structure and algorithm of the Artificial Neural Network and Quantum Neural Network are presented, comparing their respective good and bad points, a rotating mechanical fault diagnosis method based on Quantum Neural Network based on multilevel transfer function. At last, this article setups the QNN algorithm program in the operating platform of MATLAB.Compared with Artificial Neural Network, recognition performance and accuracy of the fault diagnosis system based on Artificial Neural Network was better tested and simulated by fault feature vector, which come from the steam turbine of one thermal power plant, the result indicates the effectiveness of this method.


Key words： Quantum Neural Network, fault diagnosis, fault feature, QNN algorithm,






目 錄
中 文 摘 要 I
ABSTRACT II
1 緒論 1
1.1旋轉(zhuǎn)機(jī)械故障診斷研究的目的和意義 1
1.2國內(nèi)外旋轉(zhuǎn)機(jī)械故障診斷技術(shù)的研究現(xiàn)狀和發(fā)展方向 2
1.2.1國外研究現(xiàn)狀 2
1.2.2國內(nèi)研究現(xiàn)狀 3
1.2.3發(fā)展趨勢 3
1.3量子神經(jīng)網(wǎng)絡(luò)在故障診斷中的應(yīng)用 4
1.4本文研究的主要內(nèi)容 5
2 大型旋轉(zhuǎn)機(jī)械典型故障分析和故障診斷原理 6
2.1轉(zhuǎn)子故障 6
2.1.1轉(zhuǎn)子不平衡 6
2.1.2轉(zhuǎn)子不對中 7
2.1.3轉(zhuǎn)子碰摩故障 8
2.1.4轉(zhuǎn)子彎曲故障 9
2.2軸承的故障機(jī)理 9
2.2.1滑動軸承的故障機(jī)理 9
2.2.2滾動軸承的故障 10
2.3旋轉(zhuǎn)機(jī)械故障診斷原理 12
3人工神經(jīng)網(wǎng)絡(luò) 15
3.1 人工神經(jīng)網(wǎng)絡(luò)的基本原理 15
3.1.1人工神經(jīng)網(wǎng)絡(luò)的概念和神經(jīng)元模型 15
3.1.2人工神經(jīng)元傳遞函數(shù)的類型 15
3.1.3人工神經(jīng)網(wǎng)絡(luò)的聯(lián)接 16
3.2 BP神經(jīng)網(wǎng)絡(luò) 18
3.2.1 BP神經(jīng)網(wǎng)絡(luò)模型與結(jié)構(gòu) 18
3.2.2 BP算法 18
4 量子神經(jīng)網(wǎng)絡(luò)及其在旋轉(zhuǎn)機(jī)械故障診斷中的應(yīng)用 23
4.1 QNN的量子并行處理能力和QNN的優(yōu)勢 23
4.2幾種QNN模型 24
4.2.1.多層激勵函數(shù)的量子神經(jīng)網(wǎng)絡(luò) 24
4.2.2 Qubit 神經(jīng)元模型 24
4.2.3 多宇宙的量子神經(jīng)網(wǎng)絡(luò)模型 25
4.3基于多層傳遞函數(shù)的量子神經(jīng)網(wǎng)絡(luò) 26
4.3.1多層傳遞函數(shù)的神經(jīng)元 26
4.3.2多層傳遞函數(shù)的量子神經(jīng)網(wǎng)絡(luò) 27
4.3.3 QNN訓(xùn)練算法 28
4.4 量子神經(jīng)網(wǎng)絡(luò)在旋轉(zhuǎn)機(jī)械故障診斷中的應(yīng)用 28
4.4.1旋轉(zhuǎn)機(jī)械故障診斷網(wǎng)絡(luò)建模 28
4.4.2神經(jīng)網(wǎng)絡(luò)的訓(xùn)練 29
4.4.3診斷模型在汽輪機(jī)故障診斷中的應(yīng)用 32
4.5旋轉(zhuǎn)機(jī)械故障診斷系統(tǒng)開發(fā) 34
4.5.1旋轉(zhuǎn)機(jī)械故障診斷過程 34
4.5.2故障診斷系統(tǒng)介紹 34
結(jié) 論 37
致 謝 38
參 考 文 獻(xiàn) 39
附錄A：MATLAB部分程序代碼 41