Prediction of Critical Clearing Time of Java-Bali 500 kv Power System Under Multiple Bus Load Changes Using Neural Network Based Transient Stability Model

A transient stability model based on back propagation neural network is used to analyze transient stability of Java-Bali electricity system, especially in calculating the critical clearing time. The real and the active load changes on each bus that shows the real load pattern of the system used as neural network input, while the target is the Critical Clearing Time (CCT). By using the load pattern as input, it is hoped that the robustness of the proposed method against load changes at multiple bus can be achieved. Data of target critical clearing time used for the training was calculated from the concept of One Machine Infinite Bus (OMIB), by reducing the multi-machine system using a combination of methods of Equal Area Criterion (EAC) through the Trapezoidal method and the Runge-Kutta 4th order method. To analyze transient stability, a three phase ground fault was conducted at one bus and assumed not changed during the simulation. The proposed method will be implemented at Java-Bali 500 kv power system. The simulation results show the calculation of critical clearing time from the proposed method has a minimum error of 0.0016% and a maximum error of 0.0419% compared with CCT by OMIB.