Water pump station scheduling optimization using an improved genetic algorithm approach

Abstract

The problem of energy and water wasting can be found all over the world. Water pump stations are mainly operated empirically, under the experience of the employees. Such behavior generates waste of energy and high operation costs, what are completely undesired. The system of the pump station can be modeled. The system is responsible to govern the pumps rules and reactions. Each pump has a specific curve and reacts differently to changes in the system. The best optimal scheduling point of operation of the pumps is the main focus of the present work, based on the shaft power of the pumps. The operation constraints and the water pump best efficiency zone should always be ensured in order to save energy, avoid shortages, and do not waste water. However, some typical mathematical techniques often are unable to find the optimal solution in a practical approach, due to the complexity of the constraints and difficulties of modelling the system. Therefore, aiming to solve such situation, an improved genetic algorithm method is presented. The method is designed with adaptive crossover and mutation operators, allowing a faster convergence and enhancing the chance of finding the global optimal solution. Several tests were carried out in a practical manner in a water pump station located in Shanghai, China. The method showed satisfactory results and proper implementation performance.