2021/11/22 by Saumya Goyal, Goyal, Saumya, Om Damani +3
Engineering · Environmental Science · #FOS: Electrical engineering #Membrane Separation Technologies #Smart Grid Energy Management #Systems and Control (eess.SY) #Water Systems and Optimization #Water resources management and optimization #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2111.11865
openalex publication_date 2021/11/22 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Existing techniques for the cost optimization of water distribution networks\neither employ meta-heuristics, or try to develop problem-specific optimization\ntechniques. Instead, we exploit recent advances in generic NLP solvers and\nexplore a rich set of model refinement techniques. The networks that we study\ncontain a single source and multiple demand nodes with residual pressure\nconstraints. Indeterminism of flow values and flow direction in the network\nleads to non-linearity in these constraints making the optimization problem\nnon-convex. While the physical network is cyclic, flow through the network is\nnecessarily acyclic and thus enforces an acyclic orientation. We devise\ndifferent strategies of finding acyclic orientations and explore the benefit of\nenforcing such orientations explicitly as a constraint. Finally, we propose a\nparallel link formulation that models flow in each link as two separate flows\nwith opposing directions. This allows us to tackle numerical difficulties in\noptimization when flow in a link is near zero. We find that all our proposed\nformulations give results at par with least cost solutions obtained in the\nliterature on benchmark networks. We also introduce a suite of large test\nnetworks since existing benchmark networks are small in size, and find that the\nparallel link approach outperforms all other approaches on these bigger\nnetworks, resulting in a more tractable technique of cost optimization.\n