篇名 | 馬達故障診斷之模糊類神經網路 |
---|---|
卷期 | 33:3 |
並列篇名 | Motor Fault Diagnosis by Using Fuzzy Neural Network |
作者 | 陳冠宇 、 王俊傑 、 趙安民 、 康淵 、 朱明輝 |
頁次 | 329-338 |
關鍵字 | 貼近度 、 類神經網路 、 模糊類神經網路 、 馬達故障診斷 、 Similarity 、 Neural network 、 Fuzzy neural network 、 Motor fault diagnosis |
出刊日期 | 200509 |
本文使用隸屬度函數與類神經網路邁構模糊類經網路應用於馬達故障診斷,根據實際馬達故障檢修資料,建立撤障類型與頻譜特徵關係,作為貼近度診斷與類神經網路學習的依據。量測馬達的振動信號,經快速傅立葉轉換為信為頻譜,提取頻特徵並經隸屬度函數分級後,再以類神經網路推理完成診斷。本文以兩個故障馬達實例,利用模糊類神經網路進行診斷,並與貼近度診斷、類神經網路診斷相互比較,探應模糊類神經網路對於多重混合性撤障的診斷能力。
This study proposed a method that using membership faction and neural network for motor diagnosis. The relationship between faults and frequency symptoms built up by expert experiences and overhaul information for motors. Using this relationship to diagnosing by the similarity method and training with the neural network. The frequency symptoms extracted by measured signals and graduated by membership function, and diagnosis by the neural network. In this paper, diagnosis by using fuzzy neural network in two cases and compared with similarity method and neural network to prove the detect ability of the neural network for multiple faults.