Modified firefly algorithm for directional overcurrent relay coordination in power system protection / Muhamad Hatta Hussain

Nowadays, modern power systems are interconnected and protected by Directional Overcurrent Relays (DOCRs). Therefore, the protection coordination scheme poses great challenge especially in multi-source networks due to network topology and microprocessor relay. Thus, it becomes very difficult to set the sequence of relay operations for various faults in power system. Moreover, it leads to miscoordination between the primary and backup relay pairs with the occurrence of multi-directional fault current in network. The optimization techniques eliminate the need to find the breakpoints which is applied for setting the relays in coordination process since four decades ago. Most optimization techniques especially the conventional mathematical based optimization is found to face inaccurate and stuck at local minimum phenomenon while it was computationally burdensome. Thus, a reliable optimization technique such as Nature Inspired Metaheuristic Algorithms (NIMA) is crucial to address this issue. This thesis presents Modified Firefly Algorithm for Directional Overcurrent Relay Coordination in Power System Protection. The objectives of the studies are to develop a new optimization technique termed as Modified Firefly Algorithm (MFA) for minimizing the relay operating time, to develop a Multi-Objective Modified Firefly Algorithm (MOMFA) for minimizing both the total relay operating time and relay coordination time and to develop an integrated optimal predictor termed as Modified Firefly Algorithm-Artificial Neural Network (MFA-ANN) for accurate prediction of relay operating time. A new parameter is added into traditional FA in order to speed up the convergence process. All the developed techniques have been validated on the IEEE 8-Bus, WSCC 9-bus and IEEE-14 Bus Transmission Systems. The Electric Transient and Analysis Program (ETAP) was used as the simulation tool, while Matrices Laboratory (MATLAB) was utilized to implement all the algorithms in this study. Subsequently, a new Objective Function is proposed which considers penalty function, total relay operating time and relay coordination time. Comparative studies have been conducted with respect to Multi-Objective Modified Firefly Algorithm (MOMFA), Multi-Objective Artificial Bees Colony (MOABC) and Multi-Objective Particle Swarm Optimization (MOPSO). Results revealed that the MOMFA outperformed MOABC and MOPSO in terms of preventing miscoordination occurrence within less iteration and computational time. In terms of prediction time-current characteristics, the proposed MFA-ANN model has shown the reduction in Root Mean Square Error (RMSE) values which improved the correlation coefficient of the relay operating time. The proposed MFA-ANN model managed to achieve 0% RMSE value. Thus, the protection engineers can benefit from the result in this study. They can utilize the results for the setting, type and coordination protective device to ensure reliable design in terms of security, selectivity, stability and speed.