An Application of the Harmony-Search Multi-Objective (HSMO) Optimization Algorithm for the Solution of Pump Scheduling Problem. (2016)

• Record Type: Journal Article
• Authors: De Paola, Francesco; Fontana, Nicola; Giugni, Maurizio; Marini, Gustavo; Pugliese, Francesco
• Is Part Of: Procedia engineering. Volume 162(2016)
• Page Start: 494
• Page End: 502
• Publication Date: 2016
• Subjects: Water Networks -- Optimization techniques -- Pump scheduling -- Meta-heuristic algorithms.; Engineering -- Congresses; Engineering -- Periodicals; Engineering; Conference proceedings; Periodicals; 620.005
• DOI: 10.1016/j.proeng.2016.11.093
• Languages: English
• ISSNs: 1877-7058

Abstract

Abstract: In hydraulic systems, water is often pumped to reach higher elevations, so as to ensure the minimum required pressure and guarantee adequate service level. However, pumps cannot be instantly activated and people do not consume the resource in uniform mode throughout the day. To avoid direct pumping, water can be stored in tanks at a higher elevation, so that it can be supplied whenever there is a higher demand. Because of the significant costs required for pumping, energy-saving in water supply systems is one of the most challenging issues to ensure optimal management of water systems. Careful scheduling of pumping operations may lead not only to energy savings, but alsoto prevent damages, as consequence of reduction of operation times and switches. By means of computer simulation, an optimal schedule of pumps can be achieved using optimization algorithms. In this paper, a harmony-search multi-objective (HSMO) optimization approach is adapted to the pump scheduling problem. The model interfaces with the popular hydraulic solver, EPANET 2.0, to check the hydraulic constraints and to evaluate the performances of the selected schedules. Penalties are introduced in the objective function in case of violation of the hydraulic constraints. The model is applied to a case study, showing that the results are comparable with those of competitive meta-heuristic algorithms (e.g. Genetic Algorithms) and pointing out the suitability of the HSMO algorithm for pumping optimization.