A simultaneous perturbation approach for solving economic dispatch problems with emission, storage, and network constraints

Abstract

Power system optimization problems are commonly posed as constrained optimization problems and formulated as linear or nonlinear programming problems that incorporate a Lagrange multiplier for each equality constraint and binding inequality constraint. Lagrangian relaxation solution algorithms attempt to iteratively find a solution by temporarily relaxing the Lagrange multipliers and hence temporarily decoupling the optimization across time periods. A simultaneous perturbation technique for economic dispatch problems with inter-temporal constraints has been previously proposed and shown to have good convergence properties while solving for all time periods simultaneously. However, this technique was previously illustrated only by coordinating a hydro unit and a single composite steam unit without a network model.