OPTIMAL COORDINATION OF OVERCURRENT RELAYS USING LINEAR PROGRAMMING H.E.A. Talaat A.Y. Abdelaziz A.I. Nosseir A.A. Hajjar Department of Electric Power and Machines Electric Power Department Faculty of Engineerin, Ain Shams Univ., Faculty of Mech. and Elec. Eng., Cairo, Egypt Tishreen University, Latakia, Syria ABSTRACT The problem of optimal coordination of directional overcurrent relays in interconnected power networks is addressed. The paper introduces a new problem formulation and presents the justification of this formulation using mathematical proof and computer simulation. The objective function to be minimized is defined as the sum of the time dial settings of all the relays. The inequality constraints guarantee the coordination margin for each primary/ backup relay pair having a fault very close to the primary relay. Using this formulation the size of the optimization problem is significantly reduced. Linear programming with Active set strategy two-phase method is applied to minimize the operating times of the relays. The application of the proposed approach to two example networks proves its effectiveness to set optimally the relays without the occurrence of any miscoordination and in minimum operating time as compared to other techniques. Keywords: interconnected networks, directional overcurrent relaying, optimal coordination, linear programming. 1. INTRODUCTION Directional overcurrent relaying is commonly used in power system protection as a primary protection in distribution and subtransmission systems and as a secondary protection in transmission systems. The main problem that arises with this type of protection, is the difficulty in performing the relays coordination, especially in the multiloop, multisource networks [1]. Since the sixties, a great effort has been devoted for solving this problem using computer simulation. Three approaches have been used for the setting of directional overcurrent relays: trial and error approach [2], topological analysis approach [3,4], and optimization approach [5-8]. Traditionally, the trial and error method has been