Power generation scheduling through use of generalised network flow programming

The economic dispatch problem is important in both power system planning and operation. Although there have been major advances in defining and solving more complete optimal power flow (OPF) problems, there exists a need for constrained economic dispatch (CED) techniques which, though not as rigorous or as exact as an OPF, are fast enough to be used on desktop computers. The authors present a CED formulation that is based upon generalised network flow programming (GNFP). The methodology for applying GNFP to the CED problem is developed and illustrated in this paper using four examples: 5, 39, 264, and 515-bus systems. Starting with a solved loadflow case, execution times for the CED are shown to be about the same as for a Newton-Raphson loadflow, and memory requirements are approximately 75% of the loadflow. Salient features of the CED-GNFP method include the ability to minimise generation cost while meeting constraints on line flows and area interchange. Network location of loads and generators is automatically considered without need of penalty factors, and transmission line flow limits do not affect solution times significantly. The authors believe that, owing to its simplicity of application, robustness, and ease of implementation on desktop computers, the CED-GNFP procedure is an important adjunct to the group of techniques and algorithms currently available for solving the overall economic dispatch/OPF problem.