Expected security cost optimal power flow using parallel and distributed computation

Abstract

Expected security cost optimal power flow (ESCOPF), a form of optimal power flow (OPF), addresses security as an economic cost rather than as a constraint. The objective of this study is to solve the expected security-cost optimal power flow (ESCOPF) problem using a parallel and distributed computation technique. Previous work by our research group at OSU successfully solved the integrated version of the ESCOPF problem on a single machine but pointed out that, due to combining loadflow equations and operating limit constraints for each post-contingency state, the size of the problem becomes very large. Therefore, a decomposed solution method was needed so that the computational time could be reduced. As a result we would be able to use this approach for solving the practical power system problems.

This work presents a decoupled solution method by which the ESCOPF problem can be decomposed into a main problem and many subproblems (one for each contingency). The decoupled ESCOPF has been solved using a Beowulf Linux cluster. MatlabMPI has been used for inter-processor communication. A decoupled ESCOPF method, with reduced computational complexity, has been developed and implemented. This method helps to reduce the computation time and therefore can solve problems associated with larger power systems. As a result of the improved computational efficiency, we are now one step closer to the real-time implementation of OPF and its variants.