Selection of deterioration factors in RC slab using Markov deterioration hazard model

The deterioration process of infrastructure has the various uncertainties. Therefore, the deterioration prediction of infrastructure was difficult conventionally. However, by the development of the Markov deterioration hazard model, the technology of statistical deterioration prediction was improved dramatically. By Markov deterioration hazard model, the deterioration prediction in consideration of various deterioration factors is possible. In this study, the authors focus on the deterioration of RC slabs in urban expressway and clarify its deterioration process by Markov deterioration hazard model. The detailed inspection data is distinctive of this study, and the RC slabs are inspected not per span but per panel. The 15 kinds of structural and environmental conditions are utilized as candidates for the characteristic variables. Through the comparison of the t-value and AIC, the authors found that the characteristic variable "whether or not construction joint" has the most explanatory power statistically. By estimating the Markov deterioration hazard model, Markov transition probability to represent the deterioration process can be obtained. Then, employing Monte Carlo simulation, life expectancies of each RC slab can be calculated per span. Finally, the authors quantitatively represent the difference of life expectancies of RC slabs per span corresponding to the change of RC slab's composition such as the number of panels in one span, the number of panels of construction joint and the change of other characteristic variables.