Interior-Point NLP Approach Application for Evaluation of Optimum Relay Parameters for a Mixed-Conductor Power Distribution System
摘要
Mixed overhead line and underground cable AC distribution systems impose additional restrictions on relay coordination problems due to the highly excessive current drawn by underground cables owing to their high capacitance. In electricity distribution networks, overcurrent relays (OCRs) are commonly used to develop main and backup protection systems. These OCRs have two relay settings parameters: plug setting (PS) and time multiplier setting (TMS). With the enhancement in the number of relays in large distribution systems, a perfectly matched operation of protective relaying is essential to avoid undesirable tripping of the normal sections of the electricity distribution system. As both parameters are assumed to be variables, finding accurate relay settings becomes a nonlinear optimization problem. This paper implements an interior point algorithm (IP-NLP) approach. With this method, the ideal result is obtained in two phases. A feasible optimal solution is first attained, and then the initial optimal solution is refined in the second phase. It is observed that the PS varies significantly between 0.0944 and 0.8391, while TMS is constant. This is because the PS depends on the current during a fault as detected at the relay's position, while the TMS is independent of the current during a fault as seen by the relay. A mixed conductor 13-bus benchmark distribution system is considered to formulate the constrained NLP.