Development of an automation model for leakage reduction and energy harnessing through optimised placement and setting of pressure reducing valves and pump as turbine placement in water distribution networks.

Abstract

Water leakage is of paramount interest in South Africa. Due to the increasing population density access to drinking water will be scarcer, as much of the water produced is lost through leakages. South Africa’s non-revenue water (NRW) accounts for approximately 30% of the total water supply, with leakages accounting for more than 70% of NRW in a network. Water leakage losses are defined as water lost during transportation to the consumer from the water source. Reducing leakage losses in existing networks is a major task for engineers, as replacing pipes is costly. There is a direct relationship between the pressure in a water distribution network (WDN) and the network’s leakage losses. Network pressure is affected by many factors such as consumer demand, nodal elevation differences and network characteristics. The easiest and most cost-efficient method of pressure reduction in existing WDNs is through efficient pressure management. Developing methods to reduce the excess network pressure will reduce the leakage losses in the network. There is currently little research that incorporates optimising pressure-reducing valve (PRV) placement to regulate network pressure and reduce leakage losses. Furthermore, the use of pumps operating as turbines (PATs) to harness energy in WDNs has not been fully explored. Literature gaps encompass optimising and automating these processes and creating an algorithm applicable to complex WDNs. This dissertation presents a new optimisation model aimed at minimising network leakage losses within a WDN by determining the optimal placement and setting of additional PRVs. Energy generation from excess pressure is a secondary benefit and by-product of pressure reduction and can be harnessed by replacing PRVs with PATs. Replacing PRVs with PATs to harness renewable energy from the WDN is also incorporated into the new optimisation model. Automation of the model has been achieved using MATLAB and EPANET. The objectives of this study are to create a mathematical model to optimise the placement and setting of additional PRVs within a WDN, following all hydraulic, mathematical, and linear constraints. The model must determine which PATs can feasibly replace PRVs within a WDN to generate electricity and the PAT’s potential power output and potential gross margin. The mathematical model must be automated using MATLAB and EPANET. The newly created objective function is made up of two stages to account for the multiple objectives. The model has been applied to two real-world networks: the Cornubia Integrated Human Settlement Development Phase 2A- Zone 1 network and the Pat Marshal Housing Project. The proposed methodology’s ability to reduce network pressures, reduce associated leakage losses, and generate energy has been demonstrated. The network leakage losses minimisation accounts for the hydraulic characteristics and constraints of each WDN in the form of mass continuity equation constraints at the nodes, momentum balance constraints along the pipes, network constraints, and pressure reducing equipment constraints. Five successful additional PRVs have been optimally placed and set within the Cornubia WDN, achieving a 12.8011% leakage rate reduction. Two successful additional PRVs have been optimally placed and set within the Pat Marshal WDN, achieving a 27.907% leakage rate reduction. The model has determined that nine of the 36 PRVs in the Cornubia WDN can be feasibly replaced by PATs to harness 135.48 kW of energy, equivalent to a Mini-hydropower plant, with a 948.67 MWh per annum output. One PAT can feasibly replace a PRV in the Pat Marshal WDN to harness 24.55 kW of energy, equivalent to a Micro-hydropower plant, with a 131.6 MWh per annum output. The energy produced could be sold to realise gross margins of approximately R920 211.17 and R127 662.30 per annum for the Cornubia and Pat Marshal WDNs, respectively. The results obtained from the newly proposed method are positive, significantly reducing the leakage rate in both networks. The excess pressure from the networks can be efficiently harnessed to generate renewable energy to be utilised or sold.