6 - Long-term Investment Modeling of Water Distribution Infrastructure

For long-term, cost-efficient asset management, balancing risk versus cost requires models linking a portfolio's asset degradation with owner investment. Modeling degradation of water distribution system assets against investment allows for projection of the system's condition over time, and an understanding of how different levels of consistent funding may impact the system long-term. Previous research has developed deterioration models through Markov Chain, neural networks, and other methods for a variety of asset types, but they do not examine how their projected condition translates into a portfolio's value, and the investment needed to maintain the system's condition. In this study we build a Markov Chain deterioration model drawing from the categorical conditions of over 70,000 water distribution pipes in the U.S. Air Force inventory. The probability of transitioning condition states from green to amber to red is solved via non-linear optimization, into eight sets of Markov Chain transition matrices. These models are compared and the best model selected. Additionally, we estimate the cost of pipe replacement based on assumed pipe depth, diameter, and material. We then combine the probability transition matrice(s) of the best Markov Chain model with costs to highlight the backlog of maintenance across the asset portfolio. We run this analysis multiple times in a Monte Carlo simulation and add further scenarios of funding to assess changes in overall system health over time. Our expected results will evaluate funding strategies for effectiveness in managing the aging Air Force water distribution network assets. The goal of the research is to provide a framework for assessing long-term investment and promote replacement activities.