Analysis of the water distribution main replacement conundrum in Durban.
The optimisation of the decision of when to replace water distribution mains is a complex task. There are numerous drivers in the decision making process (informed by financial data, performance data and water quality data) and hundreds of variables and performance indicators that can be considered when trying to reach an optimised decision. Most of the assets under consideration are buried and the internal and external pipe conditions are not easily assessable, forcing the Utility to rely on the available direct and indirect variables from which conclusions on the reliability of the mains are to be inferred. The cost of mains replacement is relatively low but the assessment cost, if carried out can be relatively high. The total value of a metropolitan distribution network typically runs into billions of rands but the impact of an individual pipe failure is generally low. The distribution network is comprised of many different pipe materials and components, of different pressure classes, made by different manufacturers, installed by numerous contractors with different skill levels under differing quality control regimens over many years. To add to this complexity, various parts of the network are operated at different static pressures and varying velocities. Some sections of the network are isolated more often than others and at times there can be large pressure surges that the network is subjected to by either the Utility or Consumer. These pressure surges are known to have a marked detrimental effect on the network. False markers also exist that can give rise to totally incorrect decisions and therefore performance data cannot be accepted at face value and needs to be scrutinised and cleansed to increase its reliability prior to being utilised in decision making process. This important step has been missed by much of the research carried out to date. In the Durban context, a further complication is caused by consumers tampering with the water mains and also not reporting leaks. This has a negative effect on the performance of the water main that can cause it to be flagged for replacement, but its replacement will not result in an increase in performance if the social issues are not resolved first. The aim of this research is to make recommendations on the methodology to be employed to improve network performance and thereby delay the point at which the water mains are to be replaced for as long as possible. These recommended activities will be carried out to remove false markers and improve upon the quality and reliability of the data available on the network performance. A further output is to make recommendations regarding the minimum data that can be reasonably collected and analysed in order to determine an optimised result. The recommendation of which mains should be targeted for replacement should result in the highest benefit for the utility as well as the consumers. By implication, this will result lowest long term capital and operational expenditure and thus the lowest long term tariffs charged to the consumers whilst complying with the water quality criteria and service level targets.