Risk-informed maintenance optimization of railway turnout components: A 20-year empirical reliability framework with imperfect maintenance

This study presents a risk-informed reliability framework integrating failure mechanism analysis with maintenance cost rate optimization. Using 20 years of Seoul Metro data (2004–2024), the stochastic degradation of four components was characterized: stock rail, tongue rail, lead rail, and crossing. Wear-dominant components (stock, tongue rails) followed Weibull distributions, while fatigue-dominant components (lead rail, crossing) followed lognormal distributions, consistent with multiplicative damage in rolling contact fatigue. A cost rate model incorporating structural failure risk and imperfect maintenance via the Kijima Type II virtual age model was developed. Optimal replacement intervals ranged from 3.4 years for the lead rail to 10.1 years for the tongue rail. A six-month preventive grinding schedule extended service life to the 15-year limit and reduced daily cost by approximately 36–41%, offering built environment asset managers a risk-informed decision-support tool to balance economy with structural integrity.