Design of 3-Phase Load Dividing Automatic System with Naive Bayes Method

Abstract

The electric power distribution system is a process for distributing electricity from a 150 kV electric power transmission system to consumers, both 20 kV consumers and 380/220 V consumers. The problem that often occurs in the electric power distribution system at PT.X is load imbalance due to the different consumption power on each phase, so a neutral current appears. The current flowing in the neutral phase will cause power losses. Based on these problems, a load balancing device was made for a three-phase network with the standard SPLN No. 17 of 2014, namely the magnitude of the value of the current imbalance between the phases may not exceed 25%. This tool is also equipped with an overcurrent protection system and a tapping phase limitation system to support the performance of the load-balancing device on a three-phase network. The method used is the Naive Bayes method. The naïve Bayes method calculates the current parameters in the R, S, T, and N phases. The classification results will be used as a reference for activating the tapping phase system and the protection system. This system uses Internet of Things technology to monitor neutral current, current and voltage values in phases R, S and T, and power loss values. Based on the test results, when there is an imbalance of 30.20%, a neutral current of 1.20A appears, causing a power loss of 0.17W. After the tapping phase is carried out, the percentage of unbalance is reduced to 12.20%, the neutral current is 0.45A, and the power losses are 0.02W.