Discharge redistribution as a key process for heuristic optimization of energy production with pumps as turbines in a water distribution network

Water distribution networks often exhibit excess pressure that could lead to extensive leakage and infrastructure damages. While this problem can be mitigated with pressure reducing valves, the use of micro-turbines offers the additional benefit of harnessing the excess energy for electricity production. However, the efficient placement of turbines in a water distribution network constitutes a complicated optimization problem. The addition of a turbine in a water distribution network induces additional head losses and redistribution of the discharge within the network. This study considers the discharge redistribution as a key process for the maximization of power generation and presents a heuristic methodology based on nonlinear programming. Through an iterative process, pumps as turbines (PATs) are placed in pipes where the discharge has been increased due to previous placements of PATs elsewhere in the network. The suggested heuristic methodology is implemented in a synthetic network and the results are compared to the maximum power production from all possible combinations of PAT positionings in the network. Results show that the suggested methodology reduces considerably the number of combinations to be tested and it approaches satisfactorily the maximum possible power generation. In the synthetic network, the suggested methodology is able to predict almost the maximum possible power production with up to four PATs in the network and at least 87% of the maximum power production when five PATs are in the network. Finally, the suggested methodology is applied successfully to a real-world network, where it is able to identify the optimal location of one and two PATs.