基於多通道注意機制之可解釋機器健康狀態預測

Explainable Machine Health State Prediction through Multi-Channel Attention

詹玉立(Yu-Li Chan)

指導教授 ： 帥宏翰

國立交通大學/電機學院/電機學院電信學程/碩士(2020年)

摘要

工業4.0的核心為製造業之智慧化，其中一項重要的工作為異常偵測。雖然現有方法在異常偵測上有很高正確率，然而判斷的依據卻無法提供可解釋性，大幅降低了模型改善與人機合作的可能性。本論文旨在對機器故障偵測提供可解釋性，以可解釋之分類方法實現軸承的健康狀態監視和預知保養。明確來說，我們先將來自多個傳感器收集而來的振動信號經由短時傅立葉變換可視化，並利用多通道方式監控，結合具有Squeeze-and-Excitation模塊與自我注意力機制的卷積神經網路對其整體的退化狀態進行評估，一旦進入早期退化階段，可藉由注意力權重分佈得知導致異常發生的組件來源。實驗顯示，在軸承從正常運行至損壞的實驗數據中，此方法達到了很高的準確率，並在解釋其預測結果的能力上優於其他軸承偵測的最新方法。

關鍵字

機器健康狀態預測 ； 通道注意力 ； 自我注意力 ； 可解釋性

並列摘要

The core technology of Industry 4.0 is to enable the intelligence of manufacturing. One of the important tasks is anomaly detection. Although existing anomaly detection methods have achieved high accuracy, the basis of judgments cannot provide explainability, which greatly reduces the possibility for improving the model or facilitating human-machine cooperation. Therefore, in this thesis, the goal is to provide the explainability for machine fault detection and realize the health monitoring and prognosis of the bearings simultaneously. Specifically, vibration signals from multiple sensors are visualized through short-time Fourier transform. Afterward, the features of frequency-domain data are extracted by the Squeeze-and-Excitation block and self-attention mechanism to assess the degradation of whole system. Once the process enters the early degradation, we can identify the source of components that causes the abnormality through the attention weight distribution. Experimental results show that the proposed approach achieves high accuracy in run-to-failure tests. Moreover, the proposed approach shows a better ability to explain the predicted results than the state-of-the-art bearing detection methods.

並列關鍵字

machine health state prediction ； channel attention ； self-attention ； explainability

參考文獻

[1] Hui Wang et al. “A New Intelligent Bearing Fault Diagnosis Method Using SDP Representation and SE-CNN”. In: IEEE Transactions on Instrumentation and Measurement (2019).

[2] Xiaolong Li et al. “Research on Fault Diagnosis Algorithm Based on Structure Optimization for Convolutional Neural Network”. In: IEEE 8th Joint International Information Technology and Artificial Intelligence Conference (2019).

[3] Ramin Hasani, Guodong Wang, and Radu Grosu. “A Machine Learning Suite for Machine Components’ Health-Monitoring”. In: AAAI Conference on Innovative Applications of Artificial Intelligence (2019).

[4] Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. Neural Machine Translation by Jointly Learning to Align and Translate. 2014. eprint: arXiv:1409.0473.

[5] Yao Qin et al. “A Dual-Stage Attention-Based Recurrent Neural Network for Time Series Prediction”. In: International Joint Conference on Artificial Intelligence (2017).