Savenko O. Improvement of VAR Compensation Devices and their Control Methods

The dissertation shows the possibility of reducing the rated power of the voltage source inverter (VSI), on the basis of which the active part of the hybrid VAR compensator is made, due to the development of new methods for calculating the capacitance of the capacitor bank. The power of the active part is not more than 20 % of the power of the compensator as a whole. Approaches to determining the capacitance of stages have been developed, which allow to significantly reduce the number of capacitors in the capacitor bank. To avoid problems arising during the operation of the capacitor bank in non-sinusoidal voltage, control systems have been developed allowing for flexible configuration of operating mode: between fundamental reactive power compensation only (the compensator current is sinusoidal, therefore, provides «insulation» of capacitor bank from higher harmonics) and limitedly non-active power compensation, including distortion power (overvoltages and exceeding the permissible currents of the capacitor bank are not allowed). An analysis of the operation of the hybrid VAR compensator with the proposed control systems with four- and three-wire connection to the grid confirmed their effectiveness. Control systems for the transformer connection of a hybrid VAR compensator to the grid have been developed, which allows the use of standard three-pin batteries. A control system for two-phase transformer connection has been developed, which allows to improve the technical-economic and mass-dimensional parameters of the device. A method for stabilizing the DC link voltage has been developed, which makes it possible to reduce the time of the transient process during a sudden change in the load power up to 2 voltage periods and to minimize the capacitance of the DC link capacitors of the active compensator. In addition, an adaptive regulator control system is proposed to ensure a constant switching frequency of the inverter keys, which recalculates the value of the interface filter inductance in real time, thereby adapting to its changes. The efficiency of the proposed algorithms is tested and proved by mathematical and simulation modeling, and on an experimental sample.