本文選題：鋁電解　切入點：小波變換　出處：《北方工業(yè)大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

【摘要】：工業(yè)鋁電解過程在電解槽內(nèi)進行,電解槽的運行狀況直接影響電流效率和能量消耗,掌握電解槽的運行狀況至關(guān)重要。鋁電解槽的運行狀況與陰極狀況有關(guān),鋁電解槽陰極軟母線處的電壓是反映槽況的重要參數(shù)之一。所以通過分析鋁電解槽陰極軟母線的電壓研究陰極狀況,對陰極狀況做出準確的診斷具有重要意義。采用小波變換對采集到的鋁電解槽陰極軟母線的電壓信號進行分析。分別從時域和頻域兩個方面對采集的電壓信號進行分析。采用了能量對比實驗法確定小波分解層數(shù),對電壓信號進行小波多尺度分解后,得到含噪聲較少含主成分較多的第三層近似信號。時域上,先對第三層近似信號采用無量綱特征參數(shù)提取法得到特征參數(shù),再對提取的特征參數(shù)應(yīng)用K-mean聚類分析的方法,得到區(qū)別明顯的有故障的一類數(shù)據(jù)和無故障的一類數(shù)據(jù)。但當(dāng)診斷時間間隔縮小后驗證此診斷方法的適用性時,故障識別率不能滿足實際需要。頻域上,對第三層近似信號的包絡(luò)譜提取包絡(luò)譜最大值特征參數(shù),對包絡(luò)譜最大值通過統(tǒng)計分析法得到的劃分有無故障的臨界值。用臨界值對數(shù)據(jù)進行分類分析,將數(shù)據(jù)分為有故障的一類和無故障的一類,此方法通過了適用性的驗證。最后,比較了時域和頻域上兩種診斷方法的優(yōu)缺點,發(fā)現(xiàn)頻域上對信號包絡(luò)譜的最大值進行分類分析的診斷方法更優(yōu)。故將此故障診斷方法應(yīng)用到鋁電解槽陰極狀況診斷系統(tǒng)的設(shè)計中。診斷系統(tǒng)用MySQL存儲采集到的數(shù)據(jù)并供Matlab調(diào)用;Matlab實現(xiàn)了導(dǎo)入MySQL中數(shù)據(jù)的功能、查看數(shù)據(jù)功能、繪制曲線功能、診斷功能。
[Abstract]:The process of industrial aluminum electrolysis is carried out in the electrolytic cell. The operation condition of the electrolytic cell directly affects the current efficiency and energy consumption. It is very important to master the operation state of the electrolytic cell. The operation condition of the aluminum reduction cell is related to the cathode condition. The voltage at the soft bus of the cathode of the aluminum reduction cell is one of the important parameters to reflect the condition of the cell, so the cathodic condition is studied by analyzing the voltage of the soft bus of the cathode in the aluminum reduction cell. It is of great significance to make accurate diagnosis of cathode condition. Wavelet transform is used to analyze the voltage signal of the soft bus in the cathode of aluminum reduction cell. The voltage signals collected in time domain and frequency domain are analyzed respectively. The energy contrast experiment method is used to determine the number of wavelet decomposition layers. After wavelet multiscale decomposition of voltage signal, the third level approximate signal with less noise and more principal components is obtained. In time domain, the dimensionless feature parameter extraction method is used to extract the third layer approximate signal. Then K-mean clustering analysis is applied to the extracted feature parameters to obtain a class of data with obvious fault and a class of data without fault. However, when the time interval of diagnosis is reduced, the applicability of this method is verified. The fault identification rate can not meet the actual needs. In frequency domain, the envelope spectrum of the third level approximate signal is extracted from the maximum characteristic parameters of the envelope spectrum. The critical value of the maximum value of envelope spectrum is obtained by statistical analysis. The data is classified and analyzed by using the critical value, and the data is divided into two categories: the one with fault and the one without fault. The method is proved to be applicable. The advantages and disadvantages of the two diagnostic methods in time domain and frequency domain are compared. It is found that the method of classifying and analyzing the maximum value of signal envelope spectrum in frequency domain is better. Therefore, the fault diagnosis method is applied to the design of cathode condition diagnosis system for aluminum reduction cell. The diagnosis system is collected by MySQL storage. The function of importing data into MySQL is realized by Matlab calling Matlab. View data function, draw curve function, diagnose function.
【學(xué)位授予單位】：北方工業(yè)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TQ151

【參考文獻】

相關(guān)期刊論文 前10條

1 馬憲民;張興;張永強;;基于支持向量機與粗糙集的隔爆電動機故障診斷[J];工礦自動化;2017年02期

2 曾水平;崔福偉;鄒愛笑;;300kA預(yù)焙陽極鋁電解槽陰極狀況的診斷[J];輕金屬;2017年01期

3 王臻;李承;陳旭;李惠章;王蕾;;短時數(shù)據(jù)下利用雙HTLS參數(shù)估計的異步電機轉(zhuǎn)子斷條故障診斷[J];電力自動化設(shè)備;2017年01期

4 李輝;王毅;楊曉萍;賈嶸;羅興，

本文編號：1634735