Investigating the Effects of Isolation Activities Following a Pipe Burst in a Water Distribution Network and Potential Mitigation Strategies

Abstract

The maintenance of water distribution networks’ integrity is crucial for ensuring high-quality potable water supply. Structural defects or failure within pipelines, such as leakage and bursts, can lead to physical water losses and contaminant intrusion. Despite the abundance of studies on water losses caused by pipe leakage and bursts, research on contaminant intrusion due to operational shutdown behaviours following pipe bursts is limited. There is also a dearth of investigation into the subsequent effects on water quality and the development of potential mitigation plans. In order to evaluate the effects of isolation activities following a pipe burst in a water distribution network and potential mitigation, computer programs EPANET and PODDS EPANET were used to develop simplified water distribution networks which in turn were utilised to analyse potential contamination risks. The findings in this study suggest that the implementation of valve shutdown operations to isolate pipe bursts in water distribution networks may lead to contaminant intrusion. This intrusion can occur due to the suction effect induced by negative pressure during the isolation activities. The prevention of contaminant intrusion during isolation activities is a crucial consideration for maintaining high potable water quality. To mitigate the intrusion of exterior contaminants into pipes during isolation activities, valve shutdown sequences were evaluated using hydraulic modelling in EPANET to determine the most effective operational shutdown sequence. Both fixed demand and pressure dependent demand modelling approaches were adopted to perform the hydraulic modelling study. The findings suggest that contaminant intrusion can be eliminated by applying the most effective valve shutdown sequence(s). The results also reveal that intrusion volume can be minimised by adopting optimal valve shutdown sequences. Nonetheless, it is inevitable for limitations to arise when utilising hydraulic modeling software to analyse common water distribution networks. The complexity and size of a water distribution network can simply increase by having more valves, pumps, pipes, etc. The validation of hydraulic models can be challenging when the real water distribution network is very complex. Therefore, the results presented in this study provide reasonable simulations of potential contaminant intrusion during the operational shutdown process following a pipe burst. The results assist utility operators to further understand contaminant intrusion caused by operational shutdown behaviours. They also provide an insight to utility operators to mitigate the risk. This investigation further examined the impact of the shutdown duration time of isolation valves, node elevations, and isolation block sizes on intrusion volumes. The results reveal that shortening the shutdown duration time of isolation valves reduces intrusion volumes. For example, in a pressure-dependent demand single pipe model which consists of two isolation valves, there is a potential to reduce the intrusion volume by 1m3 if the valve shutdown time can be reduced by 12 mins. The findings also suggest that changes in node elevations can increase intrusion volume in a hydraulic model where pressure dependent demand modelling approach is adopted. Water distribution network models characterized by a steeper topology between the reservoir and a burst pipe increase intrusion volumes. For instance, an 11 m increase of the elevation between the reservoir and a burst pipe can lead to additional 0.37 m3 of intrusion volume. The results further indicate a smaller isolation block will reduce demand within the block, thus lowering intrusion volumes. The impacts of the shutdown duration time of isolation valves and isolation block sizes on intrusion volumes were observed irrespective of the type of demand modelling approach employed. However, the results imply that adopting the pressure dependent demand modelling approach leads to 69% lower intrusion volumes due to the network demand's sensitivity to changes in nodal pressure. The research on simulated pipe bursts and operational shutdown activities also examined their subsequent effects on turbidity in water distribution network models. The results reveal that both pipe bursts and isolation activities significantly impact turbidity in water distribution pipes, particularly in pipes distributing substantial amounts of water in the network. The results also show that alterations in valve shutdown sequences closely relate to sensitive turbidity responses in water distribution pipes. For example, some turbidity spikes can reach approximately 100 NTU in conditioned water distribution models where the initial turbidity readings in pipelines are maintained below 0.4 NTU. However, the magnitude of the impact depends on the initial yield of the biofilm and the location of pipes within the water distribution network. In accordance with the above findings, the most effective valve shutdown sequence(s) should be implemented to minimise intrusion volumes during isolation activities and reduce the impacts on turbidity. To reduce intrusion volumes, water utility operators may shorten the shutdown duration time of isolation valves and decrease the isolation block sizes to exclude more demand nodes and reduce the demand within the block. To reduce the impacts of isolation activities on turbidity, they may avoid disturbing the primary water distribution pipes in a WDN.