System of monitoring parameters of distribution grid with corrective control based on neural network

Abstract

Background. As a rule, the control of compensating devices is carried out in the automatic control system with sensors of network parameters and control system included in a specific node of the electrical network. However, the general state of the electrical network in terms of reactive power flows is not considered. At present, static VAR compensators are mostly spread. They are designed on the principle of an indirect compensation system, which has several disadvantages. In this regard, to optimize reactive power flows and maintain the specified voltage values in the network nodes with an abruptly variable nature of reactive power consumption, it is necessary to stabilize the required network parameters and minimize the loss of electrical energy due to the flow of reactive power. Materials and methods. To improve the energy efficiency of corrective devices, it is proposed to use static VAR compensators based on magnetic valve elements. To generate control actions, an artificial neural network (ANN) module is introduced into the monitoring to predict the capacities of consumers. Such a neural network is based on an electrical network model described by the combined matrix method. The main processor generates control signals for corrective devices. Results. The authors have proposed to generate control signals for corrective devices by processing information received from remote voltage sensors and current sensors of the distribution grid. The proposed system for monitoring the distribution grid makes it possible to stabilize the required parameters of the network for consumers, to minimize the loss of electrical energy due to the flow of reactive power. Conclusions. The block of neural networks minimizes the emergency situations and accidents. The use of the static VAR compensators based on magnetic valve elements will additionally improve the energy efficiency of the distribution network monitoring system. The use of matrix analysis of network parameters in the distribution network of monitoring system to generate control signals for corrective devices allows optimizing networks in such a way as to minimize reactive power losses to select and install reactive power compensation devices and control them. The use of SVC based on magnetic valve elements as a corrective device improves the efficiency of reactive power compensation in networks with an abruptly variable nature of electrical energy consumption.