Heterogeneity of deterioration processes as a factor in statistical deterioration prediction of tunnel luminaires

A method for predicting statistical deterioration using visual inspection data for tunnel luminaires has been developed. The method uses the mixed Markov deterioration hazard model to first calculate the expected deterioration path and expected lifespan of luminaires on the basis of the statistical analysis of deterioration factors and then to estimate the deterioration path on the basis of the heterogeneity of the deterioration process of each luminaire or group of luminaires. The hierarchical Bayesian estimation method is suggested as a concrete model estimation method. To verify the effectiveness of the proposed method, empirical analysis was carried out using the visual inspection data for 10,584 luminaires located in 11 tunnels. The deterioration process depended significantly on the tunnel grade and the distance from the tunnel entrance. The expected lifespan of the luminaires was about 13.6 years and varied by about 4.3 years in accordance with the above factors. The expected lifespan of the luminaires varied from 6.9 to 17.5 years in accordance with the heterogeneity of each tunnel.