Fuzzy clustering means algorithm analysis for power demand prediction at PT PLN Lhokseumawe

Abstract

Indonesian National Electricity Company (PT PLN) as the main electric power provider in Lhokseumawe City. In fulfilling the need of electricity supply for the whole requirement, which upscale gradually. The proper forecasting method need to be premeditated. The area that was grouped based on the total of power consists of the four sub districts, namely Banda Sakti, Blang Mangat, Muara Dua and Muara Satu. In this study the fuzzy clustering mean (FCM) Classification was applied in determining the power demand of each area and categorized into a cluster respectively. The data clustering divided into six variable and five classifications of power of customer. Based on clustering step that applied revealed for four different classification of power requirement for future demand, the house hold electricity consumption measured for current consumption 9.588.466 Kw/H and forecast 10.037.248 Kw/H, for Business cluster classification measured 10.107.845 Kw/H and forecast 10.566.854 Kw/H, for industry the power measured 9.195.027 Kw/H and the forecasting revealed 9.638.804 Kw/H, and the last analysis was applied in general cluster classification based on measurement was recorded 9.729.048 Kw/H and forecasted result 10.198.282 Kw/H. this method has shown the better result in term of forecasting method by employing the cluster system in determining future power consumption requirement for the area of Lhokseumawe District.