Comparison of Methods of Pump Scheduling in Water Supply Systems

In the domestic water supply industry, the reduction of pumping costs is a continuing objective. With the efficient scheduling of pumping operations, it is considered that 10% of the annual expenditure on energy and related costs may be saved. A typical cost function will include all of the expenditure caused by the pumping process and also consider the electrical cost of pumping taking into account the various electrical tariffs, as well as peak demand and pump switching costs. Using only fixed speed pumps, it is possible to use an efficient dynamic programming based method, provided that the storage reservoir levels are known. Other techniques that are showing fruitful results in optimisation are genetic programming and simulated annealing. This paper compares these methods and discusses which is more appropriate in this type of pump scheduling problem.