Ковзний режим другого порядку в задачі слідкування за вихідною напругою матричного перетворювача

Abstract

The object of the study is a three-phase-three-phase matrix converter based on fully-controllable switches with an additional LC filter at the output connected to a non-autonomous three-phase supply grid of limited power. The mathematical model of this system, which is described by ordinary differential equations, is constructed. The synthesis of the control law by the forced introduction of the first-order two-dimensional sliding mode in the contours of monitoring the components of the spatial vector of the output voltage and one-dimensional - into the control circuit of the reactive component of the input current of the matrix converter is executed. Using the vector control strategy allows to provide a sinusoidal form of output voltage, minimizing the impact of perturbations on it, and the ability to control the reactive component of the input current. Conditions of the existence of a sliding mode with the use of the known method of equivalent control are obtained. In the formation of the sliding surface to tracking the two coordinates of the output voltage applied second-order sliding mode. For this purpose, the two-component operator variable and an additional two-component vector of error are introduced in the differential equations that describe the replacement-scheme. The intro-duction of the second-order sliding mode has reduced the number of electrical quantities that require direct measurement. This strategy eliminates the need to use an accurate current sensors, which reduces the cost of the matrix converter and facilitates practical implementation. The features of the matrix converter are investigated with this type of sliding mode in the system with real parameters of the electrical circuit. The results of a simulation carried out taking into account the constraints that affect the ability to implement the selected control strategy are presented. The possibility of reducing the effect of load changes on the form of the output voltage has been considered, taking into account the constraints specific to real systems, and has been verified using digital simulation.