WCW26

Coagtech has been working with utilities in Western Canada to assess optimization of reservoir and pump station operation. For pressure zones where water is provided from a single pump station, optimization to meet system demands is relatively straightforward. However, for complex distribution networks where multiple pump stations provide water to a complex network with reservoirs, pressure reducing stations, and other pump stations the complexity increases significantly. Optimization of operation of these systems can reduce non-revenue water loss, reduce wear on pump station assets, reduce electricity usage, and improve reservoir cycling and water quality.


This presentation will discuss key considerations for optimizing reservoir and pump station operation in complex distribution networks and how machine learning is being used to support decision making at one utility. Specific challenges related to data analysis, hydraulic interactions of the distribution system, and overall system optimization will be shared. A hybrid machine learning methodology for modelling complex distribution systems will be discussed. Technology implementation approach to support operations in real time, challenges and benefits of technology adoption, and the central role of the operator will be highlighted.