Application of Bayesian estimation method with Markov hazard model to improve deterioration forecasts for infrastructure asset management

Abstract

The heart of asset management systems for road infrastructure is the deterioration forecasting model. It provides the most fundamental information for better asset management. So far, there are many practices to build a reliable forecasting model using inspection data in conjunction with statistical theories. In many applications, however, an inadequate stock of inspection data or difficulty in applying sophisticated statistical methods have often been serious obstacles. As a solution, this paper suggests applying the Bayesian estimation method with a multi-state exponential hazard Markov chain model for simple and reliable deterioration forecasting for infrastructure. The main contents of this paper are an introduction of the model's framework combining Markov chain, hazard theory, and Bayesian estimation method, and a demonstration of its practical application with an empirical study. The empirical study was conducted with time-series inspection data of pavement from the Korean National Highways. The estimation results from the suggested method would be useful for improving the current pavement maintenance strategy for Korean National Highways. However, the most important message of this paper is that the framework of the Bayesian Markov hazard model could be the best model to use for other civil infrastructure that has gradual changes in condition. The great advantages from the Bayesian estimation method may facilitate development of customized asset management systems.

Development of the Road Pavement Deterioration Model Based on the Deep Learning Method

“…To solve such problems, some researchers have proposed using a Bayesian estimation method. This has been found to solve problems concerning a small sample quantity and convergence of the estimation function [18][19][20].…”

A stochastic model for the estimation of renewal periods of sharply curved metro rails

“…Examples of the Markov models used to characterize the deterioration of highway pavements have been presented in literature. 14–17 Madanat et al., 18 Madanat and Ibrahim, 19 Morcous, 20 Tsuda et al., 21 Li et al., 22 Wellalage et al., 23 and Hasan et al. 24 employed Markov chain models to predict the future condition of bridge components, systems, and networks.…”

Renewal Management Framework for Urban Rail Transit Assets

“…Baik et al (2006) modelled the deterioration of sewer pipe by using a Markov chain model in which the pipe condition was represented by five discrete states and the transition probabilities between the states were estimated using an ordered probit model-based methodology for developing appropriate proactive maintenance and renewal strategies. Han et al (2014) classified the highway pavement condition into five states by using three performance indicators: cracking, rutting, and the international roughness index. Pavement deterioration was characterised using a Markov hazard model and Bayesian estimation.…”