Near Real Time Pump Optimization and Pressure Management

Abstract

Management of existing systems can be interpreted as sets of decisions to make regarding pumps and valves to create hydraulic conditions able to satisfy the demand without operational problems such as pressures lower or higher than the normative pressure values. However, among the large number of combinations, some of them manage to reduce energy consumption, by finding the best operating point for pumps, and also water losses, by finding the best operating point for pressure reducing valves (PRV). Several works may be found in the literature using recent and advanced optimization techniques to define pump and valve operation. However, the processing time to define operational rules is a limiting factor for real time decision-making. Taking into account the need to improve the models in terms of optimal rules to apply in near real-time operations, this work presents a hybrid model (simulator + optimizer) to find pump speeds and PRV set points, aiming at combining energy savings with pressure control while reducing water losses. PSO is applied as the main optimization algorithm, which can also work in cooperation with other bio-inspired concepts to deploy an effective and fast search algorithm. The results allow comparisons with other techniques and show the ability of PSO to find an optimal point of operation