Optimal transition to network blueprints phase II

Within this project KWR is developing a method to automatically determine the transition from current structures to network blueprints, based on the Gondwana numerical optimisation, which takes into account the hydraulic and risk-driven aspects as well as the utility-specific decision-making mechanisms.

Water utilities are working actively on the development of network blueprints and on the gradual replacement of their distribution networks. This means that the mains are not to be replaced, on an individual basis, by new mains with the same properties. Rather, the utilities aspire to achieve an optimal network design after the large-scale replacement of old mains, spread over a period of one or several decades.

Phase I: Emphasis on the design

In 2017, within the Joint Research Programme with the water utilities, extensive research was carried out into numerical optimisation techniques that can assist in the design of network blueprints; initial steps were also taken towards the subsequent transition from the existing structures to the optimised network blueprints. In other words, the emphasis during the project’s first phase was mostly on the design of the network blueprints and less on the transition to them.

Phase II: The transition

The limited initial results provide an insight into what an optimal transition looks like when either disruption reduction, or hydraulic performance, are given precedence in the process. We also need insight into how the approach for the optimisation of the transition connects, or complements, the information from the decision-support systems that water utilities currently use in determining their replacement plans.

In this project we are elaborating a general approach which involves extending Gondwana, so as to include information from the decision-support systems in the optimisation process and to evaluate new target functions. We are also conducting a case study to properly test the added value of the developed approach, and to ensure a good connection with practice.

The model results will provide the basis for determining possible rules of thumb to feed decision-support models.