Clustering Data in Power Management System Using k-Means Clustering Algorithm

Abstract

Electricity is a source of current that cannot be released from life because it is needed as a means of production and helps solve problems in daily life. Most users use electricity without realizing the amount of electricity used in that period, it can make electricity usage soar because there is no control of electricity usage. The problem of the amount of electricity usage also occurs in campus buildings, logistics staff cannot control the use of electricity because there is no history of electricity usage in certain buildings. To solve this problem, an IOT-based KWH electricity usage monitoring system was built. Furthermore, this application has a data clustering calculation using the K-Means algorithm which aims to classify campus area data according to its electricity usage whether it enters areas that use large, normal or low loads. By using information from the data clustering, logistics employees can make a policy to make electricity savings. This system has three main parts, namely the hardware system, IoT server, and website monitoring application. In this research focuses on making website monitoring and clustering data applications. From the results of tests conducted by the K-Means algorithm has the highest accuracy value of 83.3%.