應用對稱不確定性帝企鵝 -灰狼優化演算法於馬達故障診斷分析

Application of Symmetric Uncertainty and Emperor Penguin - Grey Wolf Optimization for Feature Selection in Motor Fault Classification

簡暐倫(WEI-LUN CHIEN)

指導教授 ： 許世哲 ； 李俊耀

中原大學/電機資訊學院/電機工程學系/碩士(2024年)

https://doi.org/10.6840/cycu202400219

摘要

本研究提出了一種基於機器學習的馬達故障診斷模型，與其他演算法
相比，此模型在提高適應值和減少運行時間方面展現了其優勢。模型結構可
分為三個主要階段：特徵擷取 (feature extraction)、特徵選取 (feature selection)和分類器 (classifier)。在特徵擷取階段，使用經驗模式分解 (empirical mode decomposition, EMD)、快速傅立葉變換 (fast Fourier transform, FFT)和多解析
度分析 (multi-resolution analysis, MRA)來識別重要特徵，此階段共篩選出 144個特徵。特徵選取階段結合了過濾器方法和包裝器方法，分別為對稱不確定
性方法 (symmetrical uncertainty, SU)、二 元 灰狼優化器 (binary grey wolf optimization, BGWO)和帝企鵝優化器 (emperor penguin optimizer, EPO)。最
後，使用支援向量機 (support vector machine, SVM)用於分類以產生適應值。
為了驗證模型的有效性和準確性，使用了馬達故障電流訊號資料集、凱斯西
儲大學基準資料集和機械故障預防技術基準資料集進行測試， 在馬達故障
電流訊號資料集中， 平均辨識率高達 並 在 ∞ dB條件下獲得的最小
平均運行時間 88.02秒。 此外，使用 凱斯西儲大學基準資料集和機械故障預
防技術基準資料集的 辨識率 分別為 99.54%和 99.52%。 與傳統演算法相比，
實驗結果表明對稱不確定性帝企鵝 -灰狼最佳化模型優於傳統模型表現。

關鍵字

特徵選取 ； 馬達故障診斷模型 ； 對稱不確定性方法 ； 二元灰狼優化器 ； 帝企鵝優化器

並列摘要

This research presents a model for diagnosing motor faults based on machine learning, demonstrating advantages over other algorithms in terms of both improved fitness values and reduced running time. The structure of the model involves three primary phases: feature extraction, feature selection, and classification. During the feature extraction phase, crucial features are identified using empirical mode decomposition, fast Fourier transform, and multiresolution analysis, resulting in a total of 144 features. The feature selection stage employs a new strategy combining symmetrical uncertainty in the filter approach with the binary grey wolf optimizer and emperor penguin optimizer in the wrapper approach. Finally, a support vector machine is used for classification to generate fitness values. To validate the model's effectiveness and accuracy, motor fault current signal datasets, Case Western Reserve university benchmark datasets, and mechanical failure prevention technology benchmark datasets are utilized. Comparative analysis with traditional algorithms reveals that symmetric uncertainty and emperor penguin - grey wolf optimization model outperforms traditional models in terms of performance.

並列關鍵字

Feature Selection ； Motor Fault Diagnosis ； Symmetrical Uncertainty ； Binary Gray Wolf Optimizer ； Emperor Penguin Optimizer

參考文獻

[1] Hoang, D.T., Kang, H.J, ‘‘A motor current signal-based bearing fault diagnosis using deep learning and information fusion,’’ IEEE Transactions on Instrumentation and Measurement, vol. 69, no. 6, pp. 3325–3333, Jun. 2020.

[2] Yuan, J., Jiang, H., Zhao, Q., Xu, C., Liu, H., Tian, Y, ‘‘Dual-mode noise-reconstructed EMD for weak feature extraction and fault diagnosis of rotating machinery,’’ IEEE Access, vol. 7, pp. 173541–173548, Nov. 2019.

[3] Frosini, L., Harlisca, C., Szabo, L, ‘‘Induction machine bearing fault detection by means of statistical processing of the Stray Flux Measurement,’’ IEEE Transactions on Industrial Electronics, vol. 62, no. 3, pp. 1846–1854, Oct. 2015.

[4] Singleton, R.K., Strangas, E.G., Aviyente, S, ‘‘The use of bearing currents and vibrations in lifetime estimation of bearings,’’ IEEE Transactions on Industrial Informatics, vol. 13, no. 3, pp. 1301–1309, Jun. 2017.

[5] Jiang, F., Ding, K., He, G., Lin, H., Chen, Z., Li, W, ‘‘Dual-attention-based multiscale convolutional neural network with stage division for remaining useful life prediction of rolling bearings,’’ IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1–10, Sep. 2022.