A near-optimal solution method for coordinated operation planning problem of power- and heat-interchange networks using column generation-based decomposition

Abstract

A near-optimal solution method for coordinated operation-planning problems of power- and heat-interchange networks using column generation-based decomposition was developed to enhance computational efficiency and scalability. The coordinated operation-planning problem, based on a mixed-integer linear programming (MILP) approach, was decomposed into a master problem concerning power and heat interchanges and subproblems for energy-supply systems on the basis of the Dantzig-Wolfe reformulation. To determine a near-optimal solution, heuristic finalization was developed, in which a final MILP problem is solved after fixing part of binary variables based on the lower bound result obtained through two-stage iterative column generation. The developed method was then applied to the coordinated operation planning of power- and heat-interchange networks consisting of 5–100 cogeneration systems using a 150-kWe gas engine and a 45-kWe polymer electrolyte fuel cell on a winter representative day. In the case of using 100 cogeneration systems, 93% of the binary variables expressing the on/off status of the cogeneration units and heat-interchange pumps were fixed in the heuristic finalization. The near-optimal solution, which has a lower daily energy cost than in the conventional solution method, can be obtained without optimization termination due to limits of computation time and memory usage.