Current and future use of systems analysis in water distribution network design

Computer use in the design of water distribution networks was inititated through the use of network analysis techniques to determine system performance in terms of heads and flows. The last fifteen years, however, have seen the introduction of systems analysis optimization techniques to the range of computer models available for network design purposes. These optimization models differ markedly from the ‘traditional’ network analysis models in that they ‘design’ systems for specified loading conditions rather than just analysing the performance of predetermined systems under given loading conditions. Cost was the primary or only objective in almost all these early optimization models. Water distribution network design has, however, a number of other important objectives, such as maximizing reliability. Issues related to reliability concern include probability of component failure, probability of actual demands being greater than design values, and the system redundancy inherent within the layout of the network. The joint characteristics of these aspects of reliability are not well defined, and as a result no acceptable statement of overall distribution network reliability is yet available. The implications of these issues on each other and the design process as a whole are examined in detail. It is asserted that future work into the application of systems analysis techniques in general, and optimization models in particular, to the design of water distribution should be directed at the resolution of these issues. It is also claimed that the role of computer graphics in such models is not simply as a medium for displaying input parameters and output results such as head contours and flows in a clear graphical form. A more appropriate application of computer graphics is as a means for displaying parameters and characteristics of the reliability and redundancy such that the designer can ‘convert’ the available but inexact measures of these issues into practical statements of system performance.