A perspective on reducing uncertainties in water distribution network operation using clusters

In water distribution system operation problems, a solution is often, if not always, valid for a small variety of forecasted scenarios. The number of unknown variables, such as consumer consumptions, pipe smoothness, leaks or infections makes the forecasting of the conditions in which the network operates almost impossible. On the other end, most of the solution procedures take use of deterministic models such as water and energy balance or demand patterns - and even stochastic models cannot predict the conditions for which the problem in hand is defined. This procedure of using exact solution models for inexact problems is reconsidered while a methodology of using clusters in water distribution systems is demonstrated on selected regions, utilizing simple examples as an effective tool for bridging the gap.