查詢結果分析
相關文獻
- 類神經網路應用於船舶推進軸系故障診斷之研究
- 線切割放電加工故障診斷專家系統之研發
- A Study of Power Transformer Fault Diagnosis Using Neural Network
- 利用模糊集合與類神經網路診斷中央空調系統之故障
- 變壓器故障氣體診斷法之開發
- 應用類神經網路作空調系統故障診斷
- On-Line Distribution System Fault Diagnosis Using Decision Tree Based Neural Network
- 馬達故障診斷之模糊類神經網路
- 分立模糊類神經網路於多重故障診斷的整合評量方法
- 再生能源雙向市電併聯轉換器類神經故障診斷系統之判讀--以反流器為例
頁籤選單縮合
題 名 | 類神經網路應用於船舶推進軸系故障診斷之研究=Intelligent Fault Diagnosis for Marine Propulsion Shaft System by Artificial Neural Network |
---|---|
作 者 | 郭興家; 吳立仁; 陳俊宏; | 書刊名 | 中國造船暨輪機工程學刊 |
卷 期 | 18:4 1999.11[民88.11] |
頁 次 | 頁25-35 |
分類號 | 444.7 |
關鍵詞 | 類神經網路; 功率頻譜密度; 故障診斷; Neural network; Power spectrum density; Fault diagnosis; |
語 文 | 中文(Chinese) |
中文摘要 | 本文探討類神經網路倒傳遞演算法應用於船舶推進軸系故障診斷,研究以推進系 統中之軸承系為觀測點,取其振動加速度信號,以求取功率頻譜密度,經分析提取振動特徵 ,將特徵參數輸入類神經網路。利用倒傳遞演算法結合模糊理論來達到最佳控制學習參數的 目的,由實際軸系故障之信號為訓練樣本,訓練網路中神經元間之權值,當推進系故障發生 時,可診斷出故障類別。經實測結果顯示,在螺槳轉速範圍為100~500RPM下,取48組學 習樣本與48組測試樣本,以模糊控制器做類神經網路動態學習參數調適學習,其網路測試 之平均正確率可達81.25%。在軸承鬆動,螺槳轉動葉片不平衡,基座鬆動及正常狀況等四 項診斷項目中,以軸承鬆動之正確診斷率91.7%為最高。 |
英文摘要 | In the paper, the Back Propagation Neural Network (BPNN) is applied to the faultier diagnosis for ship propulsive system. There exist a typical vibration signal that can be obtained as training pattern for each type of fault. We collect these typical vibration signals as training data for the neural network to train the weights of network. The input of the neural network is the characteristic parameters from the Power Spectrum Density (PSD of the vibro-acceleration signals measured in the shaft bearing system. On the other hand, by applying the combination of BPNN and fuzzy algorithm, we can get an optimal propagation neural network and an opitmal learning coefficient for the neural network can be obtained. Once faults happen, we can collect the vibration signals as input of the BPNN. The BPNN will recall and judge what type of faults they belong to. During the experiments, we chose four testing types, normal condition, loose bearing, unbalanced proppeder and loose base. Forty-eight samples have been taken in the range from 100 to 500 RPM of the propeller rotating speed. The diagnosis results, indicate 81.25% correctness in average can be reached and the best result is 91.7% correctness for the case of loose bearing. |
本系統中英文摘要資訊取自各篇刊載內容。