Reservoir Targets for Urban Water Supply Systems

Abstract

Computer simulation models are widely used for planning multiple reservoir water supply systems. These models are used to investigate different options such as water supply augmentation and to establish long-term operational strategies. The driving force for these simulation models is the operating rules. As water resources become more fully utilized, it is necessary to operate the systems efficiently by using optimal operating rules. These operating rules may also be used as input to the simulation models describing the systems. A generic methodology is developed, using stochastic dynamic programming (SDP), to determine the operating rules in terms of reservoir targets for urban water supply reservoir systems. The SDP explicitly accounts for the stochastic nature of streamflow to the water supply system by considering a theoretical probability distribution for streamflow, which is a major advantage of the method. Moreover, the method theoretically produces the global optimal solution. The methodology is applied to the Melbourne urban water supply system in Australia to produce the reservoir targets. The conclusion made is that the SDP provides an objective method of deriving the reservoir targets for urban water supply reservoir systems, thereby providing satisfactory and acceptable operating rules.