Convex Relaxation of Combined Heat and Power Dispatch

Combined heat and power dispatch are utilized for coordinated scheduling of electric power systems and district heating systems via interactions through combined heat and power plants and heat pumps. Nonlinear heating flow models generally make the problem hard to solve. Existing algorithms only guarantee a local optima with a heavy computational burden. We propose a convex relaxation of the combined heat and power dispatch model to efficiently derive the scheduling strategies. A near globally optimal solution can be found to reduce considerable operational costs with affordable computing time. We test our algorithm in an integrated energy system composed of a 30-bus power distribution system and a 33-node district heating network.