電車線系統年度維護任務產生之決策支援模型研發

Abstract

現今軌道運輸需求高漲，在傳統鐵路大量電氣化以及高鐵的興建之下，電車 線系統的維護顯得愈發重要。然而近年來電車線事故頻繁發生，如何在有限資源 中合理安排電車線維護任務，同時兼顧鐵路系統的營運可靠度，為鐵路營運上相 當重要的研究課題。本研究提出了一套最佳化架構，協助規劃者以整合性的維護 策略，在考慮預測性維護與週期性維護下產生優良之電車線夜間年度維護計畫。 此外考量到面對未預期事件之彈性以及問題複雜度，亦發展出對應之規則式啟發 式演算法，協助決策者在面對未預期事故時，能在短時間內產生一個近似最佳的 電車線年度維護計畫。在案例分析中，本研究利用實際的鐵路系統資料，比較現 行維護計畫及結果之優劣。分析的結果說明本研究所提出之決策支援模型能夠有 效協助維護規劃者產生優良的電車線年度維護計畫，讓鐵路業者能夠進行更有效 益的電車線維護，建構更穩健的財務結構以及安全的運輸環境。

A well-maintained catenary system is crucial to safe and efficient railway operation. However, the nature of nonredundancy in the catenary system is the reason for the expenditure on catenary maintenance that often results in a high proportion of maintenance expenditure. Therefore, proper allocation of resources in catenary maintenance according to reliability and cost is an important task in the annual maintenance planning process. Past research on catenary maintenance planning considers only the variation in maintenance intervals in periodic maintenance (PM) policy. Our study develops decision-support models including optimization framework and heuristic algorithm to implement integrated maintenance scheme, which integrates both predictive maintenance (PdM) policy and periodic maintenance policy in the annual maintenance planning problem. Results from the case study demonstrate that the developed models can successfully identify appropriate maintenance tasks and then assign them into the annual maintenance plan in a short time. Adopting the developed models can assist planners to identify the appropriate maintenance activities, create an excellent annual maintenance plan, and ensure the required reliability in the catenary system simultaneously and efficiently.