AI技術應用於產線異常震動之偵測 = AI technology Applied to Production Line Abnormal Vibration Detection / 李蔚廷.
- 作者： 李蔚廷
- 其他題名：
- AI technology Applied to Production Line Abnormal Vibration Detection
- 主題： 嵌入式系統 震動感測器 AI模型預測異常震動. , Embedded System Vibration Sensor AI Model Abnormal Vibrations.
- URL：
電子資源
- 一般註：指導教授: 陳明宏, 王得貴. 學年度: 112.
- 書目註：參考書目： 葉.
-
讀者標籤：
- 系統號： 005181963 | 機讀編目格式
館藏資訊
摘要註
現今許多工廠因機台的故障，造成大量成本損失，如能在故障前予以保養就可以減少損失。本研究以嵌入式系統結合震動感測器收集機台馬達相關震動資訊，將資料以一般數據和震動數據分析其門檻值，再去判斷是否產生震動，並提出了一般震動判斷、多軸權重震動判斷和單一軸加重權重三種震動判斷。此外，本文加入RNN、CNN和LSTM三種AI模型預測異常震動，最後再比較人工及AI模型分析的結果，找出最佳的方法來判斷是否發生異常震動。經過最佳的異常震動判斷，希望未來可以加入預防保養的系統中，達到預防保養最好的效果，並減少因機器故障所造成的成本損失。. In modern manufacturing, machinery breakdowns lead to significant financial losses for factories. Preventative maintenance before such breakdowns can help mitigate these costs. This study employs an embedded system and vibration sensors to gather vibration data from motors on the machine. The collected data is analyzed to establish threshold values, which are used to assess whether abnormal vibrations are occurring. Three vibration detection methods are proposed; they are, general vibration detection, multi-axis weighted vibration detection, and single-axis weighted vibration detection. Furthermore, three AI models, RNN, CNN, and LSTM, are utilized to detect these abnormal vibrations. The study compares the effectiveness of manual analysis and AI-driven analysis to identify the most accurate method for detecting abnormal vibrations. By identifying the best approach, the goal is to develop a preventive maintenance system that can minimize machinery breakdowns and reduce associated costs..