Prediction of the future condition of a water distribution network using a Markov based approach: a case study of Kampala water
Prediction of the future condition of the pipe network continues to trouble water utility managers because of the complexities in determining pipe conditions amidst poor data situations. The available approaches in literature for prediction of pipe condition are skewed towards modeling the deterioration of large diameter pipes such as transmission mains without a failure history or which have been recently repaired. This is mainly because the available pipe condition assessment techniques are very costly and justified only for major transmission water mains. Moreover to date approaches for prediction of future condition of a water distribution network based on the current pipe condition are lacking. Predicting the future condition of the pipe network is required as a proactive strategy to determine the maintenance and repair requirements to enable network managers to draw the necessary priority based main replacement schedules. This thesis presents a Markov based approach for predicting the future condition of a Water Distribution Network and a method for main replacement optimisation. The approaches are illustrated on City Centre, a case study in Kampala Water, Uganda as a proof of concept. Based on case study results, the Markov model was found to be simple with an acceptable percentage accuracy of predictions of 88.4%. The Markov model approach showed both conceptually and through statistical analyses, to be an appropriate model for predicting pipe conditions in a water network. It is a good alternative for contemporary approaches that are currently being employed in developing countries including Uganda such as replacing when the break rate becomes very frequent, when the pipe becomes very old or when continuous maintenance becomes uneconomical. These contemporary approaches do not produce a practical output because they do not take into consideration the criticality of the pipes being replaced. Basing on the condition prediction models, a main replacement strategy that can enable network managers to make optimum main replacement schedules and budget forecasts was developed. The developed approach can help water utility managers optimize main replacement and maintenance decisions amidst budget limitations whilst taking into consideration both current and future states of the pipe network. This newly developed proactive approach can greatly reduce on the number of breaks especially in critical service areas as well as the severity of their impact. The research has contributed towards providing a new approach for predicting the future condition of a water distribution pipe network.