Real-time Whole-cost optimization of water production and distribution

Optimizing water production and distribution in near real-time can result in significant savings in energy and chemical costs. This paper presents a novel, generic optimization framework, based on a single-solution meta-heuristic optimization algorithm called modified hybrid discrete dynamically dimensioned search (MHD-DDS). The optimization framework finds optimal settings for all stations in the network and optimal frequencies for all variable-speed driven pumps (VSP) for the 24 hours following the optimization run start. Tampere water supply system was used as a large-scale casestudy, and the optimization was able to reduce production and distribution costs by almost 20 % while ensuring better quality of service (QoS) than before.