Adaptive virtual inertia controller based on machine learning for superconducting magnetic energy storage for dynamic response enhanced

Abstract

The goal of this paper was to create an adaptive virtual inertia controller (VIC) for superconducting magnetic energy storage (SMES). An adaptive virtual inertia controller is designed using an extreme learning machine (ELM). The test system is a 25-bus interconnected Java Indonesian power grid. Time domain simulation is used to evaluate the effectiveness of the proposed controller method. To simulate the case study, the MATLAB/Simulink environment is used. According to the simulation results, an extreme learning machine can be used to make the virtual inertia controller adaptable to system variation. It has also been discovered that designing virtual inertia based on an extreme learning machine not only makes the VIC adaptive to any change in the system but also provides better dynamics performance when compared to other scenarios (the overshoot value of adaptive VIC is less than -5×10-5).