Comparative evaluation of three distinct energy optimization tools applied to real water network (Monroe)

Abstract Pump station is the biggest energy consumer in a water distribution system (WDS). A large amount of money is expended to provide energy for pumps. The environmental footprint associated with these excess energy demands is a source of concern. By implementing an optimum pump schedule that needs a minimum amount of energy to provide enough pressure and flow for water system, operational cost will be reduced and water system will be more environmentally friendly. Researchers are trying to find practical tools and methods to optimize pump operation. In this research, Pollutant Emission Pump Station Optimization (PEPSO), Darwin Scheduler (DS) and another approach that uses Markov Decision Processes (MDP) have been used as three different tools for optimizing pump operation of WDS of Monroe, MI, USA. In all three methods pumping optimizations have been done based on reducing energy usage, at the end results of running these three tools have been compared. The comparison results show that pump operation that has been taken from MDP algorithm has the best result in terms of energy usage and the number of pump switches, while pump operation taken from DS can be more effective at volume stored in tanks. The simulations showed PEPSO to be considerably faster than the other two evaluated methods in arriving at the optimum solution.