Long term optimization of electrical system generation by convex programming

Abstract

This paper describes an application of convex programming for optimal long term planning of an electrical system.

After a short description of electrical system planning requirements (Section 1), the total costs function is established and the security constraints are expressed with a set of linear constraints, as are total capacity and annual rate limitations of various types of units (Section 2).

The resulting convex program is then solved with a feasible direction method (Section 3): each iteration, the locally best direction is computed by a process similar to the gradient projection method.

Additional comments can be found in Section 4.