基於多尺度方均根、多尺度熵、費雪法與倒傳遞網路之軸承錯誤診斷系統

Abstract

機械的狀態監控對於產業日漸重要，因為產業需要提高機械的可靠性，並且減少機械故障可能造成的經濟損失，以提升產業競爭力。在監測旋轉機械狀態的領域中，使用振動訊號進行分析是相當普遍的。透過比較機械運行時，正常與故障情況下產生的訊號，將可能檢測出軸承缺陷錯誤的類型。一般情況下，軸承錯誤診斷流程可以分為三個主要步驟：特徵擷取、特徵選取與錯誤情況的分類。在本論文中，我們提出了一個錯誤檢測的演算法，以區分不同種類的軸承故障。首先，收集振動訊號，並使用不同方法擷取其特徵，如多尺度熵(Multiscale Entropy, MSE)和多尺度方均根(Multiscale Root Mean Square, MSRMS)演算法。其次，使用費雪法(Fisher Score, FS)選取最佳的特徵。最後，使用最佳的特徵與倒傳遞網路 (Backpropagation Neural Network, BPN)來建立錯誤狀態分類的模型。在我們的模擬中，使用了凱斯西儲大學(CWRU)的軸承振動訊號資料。實驗結果表明，此錯誤診斷的流程應用於軸承訊號具有相當高的辨識率。

Machine condition monitoring is gaining importance in industry because of the need to increase reliability and to decrease the possibility of production loss due to machine breakdown. The use of vibration signals is quite common in the field of condition monitoring of rotating machinery. By comparing the signals of a machine running in normal and faulty conditions, detection of fault types of bearing defects is possible. Generally, a bearing fault diagnosis process can be decomposed into three major steps: feature extraction, feature selection and fault condition classification. In this dissertation, we propose a fault detection algorithm to distinguish different types of bearing fault. Firstly, the features of vibration signals collected from different conditions were extracted by multiscale entropy (MSE) and multiscale root mean square (MSRMS) algorithm. Secondly, the optimal feature set is selected by Fisher score. Thirdly, the optimal feature set and backpropagation neural network (BPN) was used to build the model of fault classifier. In our simulations, the vibration signal datasets of bearing from Case Western Reserve University (CWRU) are utilized. Experimental results demonstrate that the proposed algorithm provides a high accuracy of prediction on the test data.