Estimating Transition Probabilities in Markov Chain-Based Deterioration Models for Management of Wastewater Systems

Abstract

Accurate prediction of the current and future conditions of wastewater systems using available assessment data is crucial for developing appropriate proactive maintenance and rehabilitation strategies for an aging wastewater collection and conveyance system. This paper proposes a method to estimate the transition probabilities of different condition states in Markov chain-based deterioration models for wastewater systems using an ordered probit model. The proposed model is applied and evaluated using the condition data of sewer pipes managed by the City of San Diego's Metropolitan Wastewater Department. The developed model presents some advantages in estimating transition probabilities over the approaches developed in the past, including the nonlinear optimization-based approach, in terms of versatility in the implementation, precision of the estimated data, and appropriateness of the assumptions in the model. The paper concludes that the ordered probit model approach is a statistically sound and robust method; however, in order to gain greater accuracy in deterioration modeling, periodic assessment of the wastewater systems with more data types is desirable.