Application of ant colony optimisation in distribution transformer sizing

Abstract

This study proposes an optimisation method for transformer sizing in power system using ant colony optimisation and a verification of the process by MATLAB software. The aim is to address the issue of transformer sizing which is a major challenge affecting its effective performance, longevity, huge capital cost and power loss. This method accounts for the constraints imposed by the load capacity and the thermal overload that the transformer serves throughout its lifetime. The objective function to be minimised includes the transformer capital cost as well as the energy loss cost. In this paper, the Optimal Transformer Sizing (OTS) problem which is fundamentally the basic routine for the location of transformer was addressed by means of the heuristic Ant System Method using the Elitist strategy, called Elitist Any System (EAS). EAS belong to the family of Ant Colony Optimisation (ACO) algorithm. ACO when appropriately applied determines the least cost path, taking into consideration the various essential factors including transformer bid price, growth rate, inflation rate, peak load, thermal deviation and energy loss cost. The study demonstrated a significant saving in capital cost using this approach as evidenced from the changes to the transformer following the initial installed capacity of 190kVA to 320kVA in the second stage and then finally to 630kVA in the third stage which effectively supported the remaining period under consideration. This finding is in contrast to the traditional simplified sizing strategy usually adopted by utilities companies.

Keywords: ant colony, optimization, transformer sizing, distribution transformer