【摘要】：齒輪作為現(xiàn)代機械設(shè)備中使用極其廣泛的連接和傳動零部件，故障發(fā)生概率比較高，一旦發(fā)生故障會造成停機停產(chǎn)，甚至威脅到人員安全。目前齒輪故障診斷方式比較多，但是傳統(tǒng)方法在進行診斷時未將齒輪振動信號的非線性因素考慮在內(nèi)。為了彌補這方面的局限，本文在總結(jié)和借鑒前人研究成果的基礎(chǔ)上，采用分形理論這一非線性分支學(xué)科，將其與具有非線性特點的神經(jīng)網(wǎng)絡(luò)相結(jié)合進行齒輪的故障診斷。 本文首先介紹了齒輪故障診斷的研究意義和發(fā)展現(xiàn)狀，分析了傳統(tǒng)故障診斷方法的不足，在此基礎(chǔ)上提出了本文的研究內(nèi)容和研究方法。 針對齒輪故障，對典型故障進行了研究介紹，分析了齒輪運行時的振動類型，然后選取三種常見典型故障從有限元的角度分析了振動特性，隨后在自制平臺上進行了齒輪故障模擬試驗，采集得到齒輪振動信號。 根據(jù)分形理論，對振動信號依次進行了分形濾波、分形維數(shù)求解等處理，然后將振動信號的分形維數(shù)和統(tǒng)計參數(shù)作為BP神經(jīng)網(wǎng)絡(luò)的輸入，進行模式識別，實現(xiàn)了分形理論和神經(jīng)網(wǎng)絡(luò)的聯(lián)合診斷，，并通過了理論方程和試驗數(shù)據(jù)的雙重驗證，且故障診斷準確率較高，證明了本文所選方法的有效性和正確性。 利用Visual Basic和MATLAB兩大軟件的優(yōu)勢進行混合編程，開發(fā)出齒輪故障診斷系統(tǒng)，將本文之前部分內(nèi)容進行了整合，界面友好、操作簡單，使本文實現(xiàn)了由理論研究到實踐應(yīng)用的邁進。
[Abstract]:Gear is widely used in modern machinery and equipment, and the probability of failure is relatively high. Once the fault occurs, it will cause shutdown and even threaten the safety of personnel. At present, there are many ways of gear fault diagnosis, but the traditional method does not take the nonlinear factor of gear vibration signal into account. In order to make up for this limitation, this paper uses fractal theory, a nonlinear branch of science, to diagnose gear faults by combining it with neural network with nonlinear characteristics on the basis of summarizing and referring to the previous research results. This paper first introduces the research significance and development status of gear fault diagnosis, analyzes the shortcomings of traditional fault diagnosis methods, and then puts forward the research contents and methods of this paper. In this paper, the typical faults of gears are studied and introduced, and the vibration types of gears are analyzed. Then, three common typical faults are selected to analyze the vibration characteristics from the point of view of finite element method. Then the gear fault simulation test was carried out on the self-made platform, and the gear vibration signals were collected. According to fractal theory, the vibration signal is processed by fractal filtering and fractal dimension solution, and then the fractal dimension and statistical parameters of vibration signal are used as the input of BP neural network for pattern recognition. The joint diagnosis of fractal theory and neural network is realized, and the double verification of theoretical equation and experimental data is carried out, and the accuracy of fault diagnosis is high, which proves the validity and correctness of the proposed method. By using the advantages of Visual Basic and MATLAB, a gear fault diagnosis system is developed. The former parts of this paper are integrated, the interface is friendly and the operation is simple, which makes this paper realize the progress from theoretical research to practical application.
【學(xué)位授予單位】：燕山大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2012
【分類號】：TH165.3

【參考文獻】

相關(guān)期刊論文 前10條

1 孫博玲;分形維數(shù)(Fractal dimension)及其測量方法[J];東北林業(yè)大學(xué)學(xué)報;2004年03期

2 歐陽明;李玉清;曾海霞;王洪亮;;VB與MATLAB混合編程在工程中的應(yīng)用[J];電子技術(shù);2009年07期

3 劉小峰;彭永金;李慧;;諧波小波解調(diào)法在齒輪箱故障診斷中的應(yīng)用[J];重慶大學(xué)學(xué)報;2011年01期

4 朱榮福,葉念渝;一種模糊自動控制分形濾波器[J];華中科技大學(xué)學(xué)報;2001年12期

5 董辛e

本文編號：2180558