Power Flow Using Thread Programming

Abstract

Power flow problem which is one of the most important applications in power system is solved using OpenMP Application Program Interface (API) a thread based programming technique for the shared memory programming environments. The parallel implementation of power flow is tested on a shared memory machine with cache-coherent non-uniform memory access (ccNUMA) architecture. Results are presented for various power test systems. The thread programming based power flow program decreases the runtime substantially. OpenMP lets a user concentrate on his/her own work rather than the parallel programming details. OpenMP is a set of compiler directives and requires minimal knowledge regarding parallel programming. It can easily be used not only on the shared memory systems but also on the systems with hyper threading and multi-core technologies as well. Thus speeding up any kind of power system related study is now viable without having to have expensive parallel computers and extensive experience on theirs programming. Since todays' and future personal computers and servers will mostly be based on the multi-core technologies the proposed method is an alternative to hard-coded parallel programming. © 2011 IEEE.