GPU-Based Power Flow Method a Multi-Objective Power Optimization Model for Reconfiguration Problem in Radial Distribution Networks

GPU-Based Power Flow Method a Multi-Objective Power Optimization Model for Reconfiguration Problem in Radial Distribution Networks

Hiba Yahyaoui, Abdelkader Dekdouk, Saoussen Krichen
Copyright: © 2018 |Pages: 12
DOI: 10.4018/IJEOE.2018100103
(Individual Articles)
No Current Special Offers


This article addresses the distribution network reconfiguration problem (DNRP) and the power flow method. The studied DNRP operates on standard configurations of electrical networks. The main objectives handled are the minimization of power loss, the number of switching operations and the deviations of bus voltages from their rated values. Metaheuristic approaches based on Greedy Iterated Local Search where proposed to solve the DNRP. A benchmarking testbed on standard systems well illustrates the incentive behind using GrILS for solving the DNRP. In addition, the proposed approaches and the power flow method where implemented on GPU architecture. The GPU implementation shows its effectiveness against the CPU in terms of time consuming specially for large-scale bus systems.
Article Preview

1. Introduction

Power system distribution delivers power utility to numerous consumers as residential, industrial and commercial customers. Hence, with the recent challenges of power production and consumption, a reliable and economic customer satisfaction is becoming more and more crucial in a modern electric power grid. The continuous increasing of power demands and the high load density in the urban areas make the power distribution management more complex, particularly when operating on the existing power grid.

In this paper, we treat the problem of power distribution network reconfiguration on the top of a power flow analysis. This consists in finding the right configuration of power distribution that mainly minimizes the power loss. Indeed, many researches have been carried out to solve optimally the network reconfiguration problems using approximate methods (like metaheuristics) as this problem is a well-known NP-problem. Actually, metaheuristics have been effectively adapted on a variety of problems with highly complicated interactions between decision variables and objectives.

Several engineering problems have been successfully solved by metaheuristics and yield satisfactory solutions within acceptable computation time. However, with the increase of clients’ demands and the complexity of the offered services, the problems are becoming more and more complex and computationally demanding, hence the need of high performance computing machines like parallel machines and multicore processor machines are deemed more than necessary. Indeed, the conventional Central Processing Units (CPUs) have attained their physical evolution limit, hence, scientists focus more on an alternative infrastructure that deals with many cores to carry out computationally expensive tasks. In practice, the potential of the CPU is combined with the GPU, which acts as a co-processor that executes many threads in parallel.

Complete Article List

Search this Journal:
Volume 12: 1 Issue (2024): Forthcoming, Available for Pre-Order
Volume 11: 4 Issues (2022): 2 Released, 2 Forthcoming
Volume 10: 4 Issues (2021)
Volume 9: 4 Issues (2020)
Volume 8: 4 Issues (2019)
Volume 7: 4 Issues (2018)
Volume 6: 4 Issues (2017)
Volume 5: 4 Issues (2016)
Volume 4: 4 Issues (2015)
Volume 3: 4 Issues (2014)
Volume 2: 4 Issues (2013)
Volume 1: 4 Issues (2012)
View Complete Journal Contents Listing