Design to optimize the location, number, and performance of dynamic voltage restorers using artificial neural networks

Abstract

The need for electrical energy always increases from year to year. This means that the distribution system in the electric power system needs to pay attention to its level of stability and reliability. A low level of stability can cause disruption and result in losses. The system's stability and reliability can be increased by installing custom power devices (CPD) equipment such as a dynamic voltage restorer (DVR). In this research, the location, number, and performance of DVRs are optimized using an artificial neural network based on the voltage stability of the distribution network in the Sibolga Penyulang SB02 area. Based on the research results, buses 2, 12, 24, 27, and 35 are the best places to install DVRs, and the system will have five DVRs installed. A three-phase short circuit simulation was used to determine how feeder stability was impacted by DVR performance. Then, the voltage falls to 0.1770 p.u. during a disturbance and then rises to 0.8073 p.u., which is within the typical voltage limit of > 0.9 p.u. It means that DVRs restored the voltage fully to the acceptable threshold.