Rule Based Algorithm for Meter Placement and ANN Based Bus Voltage Estimation in Radial Power Distribution System

Abstract

The single most important parameter for assessing the health of a power distribution system is the bus voltages. Depending on the information of the bus voltages, different corrective measures such as optimal capacitor switching, feeder reconfiguration, load balancing, etc. are undertaken. Presently, under the ambit of distribution automation (DA) system, the bus voltages are either directly measured or estimated (using state estimation technique) from the data collected by the sensors and remote terminal units (RTU) placed in the system. Because of the radial structure of the power distribution system, a relatively large number of RTUs and sensors are required to be placed for a good quality of voltage estimation. To reduce the number of sensors and RTUs required, artificial neural networks (ANN) can be used effectively, without compromising the accuracy of bus voltage estimation. One such scheme for bus voltage estimation for radial distribution feeders is presented in this article. It is shown that instead of using a single ANN, if multiple ANNs are used, accuracy of the estimation increases considerably. However, the estimation accuracy depends on the locations of meters placed and the number of ANNs used. To determine the location of meters and the number of ANNs required for a given estimation accuracy, simple rule based algorithms are proposed in this article.