音洩線上檢測技術於壓力容器腐蝕研究

On-line Monitoring and Diagnosis of Pressure Vessels

現階段壓力容器安全檢查，多以停機檢查為主。停機檢查常需停機數月，耗費大量人力與費用。研究目標為利用音洩檢測技術，開發線上監視檢測數據，並智慧判讀是否腐蝕缺陷，亦即壓力容器不停機檢查系統。音洩檢測技術具有快速、早期破壞預測、不受複雜形狀影響、可在惡劣環境下檢測等優點。由音洩檢測數據特徵設定為類神經模糊系統的輸入層，腐蝕缺陷為系統的輸出層，兩者之間的關係，使用尤克里汀距離量測法以及連續前進尋根法進行最佳特徵選擇，後續使用相位陣列式超音波進行檢查驗證。此系統具有可靠度和安全性，可以減少停機檢查、增加效率、降低成本。

The current safety inspection of pressure vessels is the shutdown inspection (SDI). This method is time consuming, labor-intensive, and extremely expensive. Furthermore, the objective of this project is to develop an intelligent on-line real time corrosion defect inspection and diagnostic system, which is on stream inspection (OSI). Acoustic emission (AE) technique will be used for on-line real time data acquisition. This technique has the merits of quick inspection, early alarm, and inspection in the harsh environments. Feature extraction of the experimental data is input layer of the neuro-fuzzy system, the corrosion defect of pressure vessels is output layer. The relationship of between the both sides make optimization form Euclidean Distance Measure (EDM) and Sequential Forward Search Algorithm (SFSA), with phase array ultrasonic system for inspection. Finally, the developed intelligent on-line real time system has very high reliability and safety. This OSI system can reduce SDI, increase efficiency and reduce cost.