類神經網路應用於系統感測器失效偵測之研究

Research on Neural Networks Applied to Sensors Fault Detection

朱致名(Zhi-Ming Chu)

指導教授 ： 陳以明

元智大學/工程學院/工業工程與管理學系/碩士(2006年)

摘要

現代科技產業中量測系統技術被普遍使用，尤其當感測器失效偵測及隔離是避免裝備損壞、確保系統性能與可靠度的重要關鍵。本研究探討類神經網路在感測器失效偵測之應用，類神經網路具有分散式平行處理、學習能力的優點，可運用來估測感測器運作狀態，處理即時性的失效與判別。本文提出失效偵測模式並以F-16飛機操控系統感測器為研究對象，以飛機正常飛行及各種系統感測器失效狀況模擬，運用訓練過的類神經網路進行狀態估測值與感測器量測值之殘差計算，再配合失效偵測模式中的偵測邏輯及臨限值設定與判斷，達到感測器之失效偵測。最後以三種不同失效類型驗證分析，以驗證其偵錯能力。

關鍵字

類神經網路 ； 感測器 ； 失效偵測 ； 殘差 ； 臨限值

並列摘要

The paper investigates the application of sensor fault detection design in neural networks. The sensor technology is a key point for avoiding break down with equipment and ensuring the running and reliability of system in industry for sensor fault detection and isolation. Neural networks(NNs) used to provide analytical redundancy to estimate working status for sensor and to treat real fault detection and judgment in detection logic, because they have the following advantages: robustness against unmodeled dynamics, capability of handing nonlinear dynamics, modular and systematic design, and potential for unanticipated failures. It’s show that program to construct flying condition via F-16 aircraft model and to collect regular and fault information as well. The developed back propagation NNs were trained flying information, and then, to fulfill detection by residuals from comparing the NNs output with the sensor output. The faults decision employed logic detection and threshold value setting in the proposed fault detection mode. Finally, the paper verify it’s detects ability among three various fault type analysis.

並列關鍵字

Neural networks ； Sensors ； Fault detection ； Redundancy ； Threshold

參考文獻

[1] P.M. Frank, “Enhancement of robustness in observer-based fault detection”, Int. J. Control ,59, 955–981 (1994).

[2] Willsky, A.S.“A Survey of Design Methods for failure Detection in Dynamic System”, Automatica 12:60, pp601-611,1976.

[4] R. Isermann and P. Balle, “Trends in the Application of Model-Based Fault Detection and Diagnosis of Technical Processes”, Control Eng. Practice, vol.5, no.5, pp.709-719, May 1997.

[5] J. J. Gertler, “Survey of Model-based Failure Detection and Isolation in Complex Plants,” IEEE Control Systems Magazine, pp.3-10, Dec. 1998.

[6] G. Betta and A. Pietrosanto, “Instrument Fault Detection and Isolation：State of the Art and New Research Trends,” IEEE Instrumentation and Measurement Technology Conference, St Paul, Minnesota, USA, pp.483-489, May 1998.