Optimization of Renewal Strategy for Railway Track Based on Remaining Useful Life Prediction

Track geometry condition is an essential part of railway track performance, thus scheduling reasonable maintenance time can improve transportation efficiency and reduce maintenance cost, while ensuring operation safety. In this paper, a probabilistic framework for optimum maintenance strategy combined inspection and maintenance postponement interval is presented based on Remaining Useful Life (RUL) prediction under uncertainty. This optimum probabilistic framework is based on five objectives considering extended service life, total life-cycle cost, inspection delay, maintenance delay and infrastructure failure probability. When the information related to the RUL of infrastructure is obtained from deterioration model, the multi-objective optimization process provides a series of trade-off solutions. Through decision-making process, the best solution is selected, indicating the optimum inspection application time and maintenance postponement interval. The proposed framework is applied to a railway line subject to geometry degradation. As a result, the proposed method can assist railway agencies in achieving acceptable track geometry condition level with less track possession time and efficiently allocating budgets and resources.