- 學位論文
時域反射與小波轉換應用於線束故障偵測之研究
Application of time domain reflection and wavelet transform for wiring-fault detection
摘要
現代商用飛機維修乃根據美國聯邦航空署(FAA) 『維修指導委員會』使用其所討論編制之「維修標準」，再加上各航空公司經驗，依照飛機系統、結構等不同區域分析所編製之「建議維修計畫」作為航空公司維修飛機最基本的指導方針與標準。 然而此維修標準並未考慮到飛機線束維修需視為一特別之系統，結果使得維修人員只有在為了其他目的與理由在執行該區域檢查維修時，才會在該區域一起做線束目視檢查。 換言之，依照目前飛機維修之標準與法規，並不會對線束做特別檢查除非該線束相關界面功能產生功能失常。直到最近幾年，線束故障所導致之飛行安全問題，已引起美國太空總署(NASA)、美國運輸安全委員會(NTSB)與FAA等等單位高度重視，開始研究飛機線束老化、故障、偵測相關技術與測試儀器等議題。 本研究報告乃針對上述議題收集相關研究資料，研究線束偵測技術最近發展，包含介質-高電壓表、時域反射儀、駐波比、駐波反射儀、超音波、網路分析儀、相增益分析儀、紅外線檢測、阻抗比、電容比、射頻-電磁放射-反相幾何、X光檢查、類神經網路分析等技術，試圖找出較佳之線束缺陷瑕疵偵測技術以及測試訊號處理與分析技術趨勢。其中經過改良之時域反射儀已有多項加強功能可以協助許多故障類別之線束偵測，但維修人員必須具有良好電路基礎與多年經驗才能達到一定判斷能力。為了改善這種缺失，本文導入小波轉換技術應用在故障點所反射回來之信訊號分析上，經過Daubechies母波轉換分解後，可以自動準確的判斷出故障位置，不需要靠資深維修人員靠反射電壓波型比對經驗判斷，即可得到良好效果。 本文獲得主要結果為： (1)以時域反射儀所累積經驗與技術，比較能夠鑑別線束故障之種類與位置，但是卻必須依賴工程師多年累積之經驗，以所謂的經驗比對法來判斷線束故障種類與位置，使得飛機線束維修工作充滿了誤差與不確定性，浪費許多工時成本。 (2)對於線束老化偵測問題，目前尚無有效相關測試儀器能夠提供具體辨識功能。 (3)使用目前比較先進之時域反射儀(TDR)以及其偵測結果，進一步應用小波轉換技術，可以改善線束故障點之確認，也就是改善故障位置(location)之精確度，且不需依賴工程師經驗之人工判斷，即可加速維修人員在飛機狹小有限空間，精確找到故障位置並排除之，節省飛機停機時間，增加飛行安全。
並列摘要
Most modern commercial airplanes are delivered with a recommended maintenance plan based on an airplane zonal analysis using the Maintenance Steering Group evaluation process for the maintenance criteria. However, the maintenance criteria and previous evaluation processes did not consider the airplane wiring as a specific system. Typically, result in that airplane wiring was examined visually when a maintenance technician was in the general area for other reason. Since year 1998, due to wiring-fault may result in aircraft serious accident, the Government of United State related department, such as NASA, NTSB and FAA pay a lot of attention and efforts, include set up a committee (ATSRAC ) especially to evaluate and investigate the issues consists of airplane aging wiring, wiring system fault detection and associated test equipment. The latest technical reports, papers and developments associated the above issues are collected from industrial instrument area, comprises DWV, TDR, SWR, Ultrasound, network analyzer, phase gain analyzer, infrared inspection, resistance ratio, capacitance ratio, Reverse Geometry X-Ray Inspection, Neural Networks Analysis…etc. The above application of techniques using on those instrument can help us have superior understanding the trend and development on wiring-fault detection. Among others, many functions have improved on the TDR, which instrument not only can figure out the fault distance but also can diagnostic the different type faults of wiring system, however, the wiring-fault justify exactly shall reliance upon good knowledge of electrical background engineer and his many years working experience. This report introduce the wavelet transform techniques, and using Daubechies’ mother wavelet to decomposition a reflection voltage wave from fault point on the wiring. The result of above analysis is that an exact wiring-fault location detecting can automatic justify without depending on experienced electrical engineer. The consequence of this study learned that (1) TDR operation & justification for wiring-fault location shall reliance on the contrast & compared method of electrical engineer many years working experience, but sometimes, engineer still can’t make very certain results, and waste a lot time for wiring-fault trouble shooting, (2) There is no appropriate test instrument, functionality and other better way to figure out the trend of decaying on aging wires issues, (3) The wavelet transform techniques, using Daubechies’ mother wavelet to decomposition a reflection voltage wave from wiring-fault point is a effective method, which algorithm can expedite wiring-fault trouble shooting process and save much aircraft ground time and increase flight safety.