Real-Time Distributed Economic Dispatch Adapted to General Convex Cost Functions: A Secant Approximation-Based Method

Abstract

The increasing penetration of distributed energy resources (DERs) in power systems has aroused interest in distributed economic dispatch (DED). While quadratic cost functions are usually adopted, those of DERs are not necessarily quadratic in practice. Existing distributed algorithms perform unsatisfying convergence rates under nonquadratic cost functions. Therefore, this article proposes a secant approximation-based method (SAM) for general convex cost functions to achieve efficient convergence. The marginal cost functions are represented by linear functions in particular price intervals. With a dynamic ratio consensus (DRC) approach, the DERs with linear marginal cost functions can rapidly reach price consensus. The approximated prices can, in turn, determine new price intervals, which are narrower and more accurate.Case studies based on real-world data verify the improved convergence rate under general convex cost functions and exhibit growing performance in larger-scale cases. The proposed algorithm can be applied to DED in systems with high penetration of DERs and massive agents with general convex cost functions.