Evaluation of methodologies for shunt VAr planning

Abstract

Two methods are developed and evaluated for shunt VAr planning. These methods are capable of sizing and siting shunt VAr at optimum cost and with greater reliability. The methods utilize the generalized Benders decomposition (GBD) technique to decompose the VAr problem into two subproblems. The first level of the problem solves an optimal power flow problem by scheduling control variables including existing VAr sources. The two methods for the second-level discrete allocation process use a branch-and-bound enumeration technique and a mixed-integer linear programming technique. The constraints include security while minimizing the total investment cost. Tests on the IEEE 14-bus system have shown both algorithms to be robust and fast.<>