Optimized PI gains for dynamic voltage restorer control using admittance estimation strategy

This paper describes how to use particle swarm optimization to implement a neural network-based admittance estimation strategy for controlling a dynamic voltage restorer (DVR) with optimized controller gains (PSO). To improve DVR efficiency, robustness, and tracking capability, a control strategy based on source admittance components is used to determine fundamental admittance components from distorted supply voltages. Because of the algorithm's simple formulas, estimating conductance and susceptibility from load currents is quick. To minimize the gains of proportional integral controllers, particle swarm optimization (PSO) is used. This improves dynamic response, tracking capability, as well as transient and steady-state stability. When consumer loads change on a regular or irregular basis, this algorithm is very effective. The admittance control strategy for DVR is modeled and simulated using MATLAB/Simulink and implemented using hardware setup d-SPACE made MicroLab box DS1202/DS1302 for mitigating harmonics and voltage sag/swell (2 GHz dual-core real-time processor and user-programmable FPGA). The suggested DVR’s test results were satisfactory under both balanced and unbalanced load conditions.