SISTEM PREDIKSI PENGGUNAAN LISTRIK PELANGGAN DI PT.PLN (PERSERO) RAYON LAMONGAN AREA BOJONEGORO DENGAN METODE TRIPLE EXPONENTIAL SMOOTHING (BROWN)

Abstract

Electricity is one of the means of fulfilling the needs of human life which is very important in this era. Excessive use of electricity will have an impact on the high use of electricity kWh. The process of recording kWh on the customer meter is carried out by officers from PLN who routinely visit the customer's homes once a month. The meter recording clerk cannot record when the customer's house cannot be reached resulting in empty customer kWh data. Prediction System Using Electricity Customers at PT. PLN Lamongan aims to determine the amount of electricity usage kWh of the customer for the next period. This research uses the Triple Exponential Smoothing method (Brown). The calculation is done on 10 different customers with 24 data, namely the use of electric kWh per period from January 2015 to December 2016 with 9 different alpha values, namely alpha 0.1 - 0.9 and uses a reference of 3 months, 6 months and 12 months before. Prediction results will be compared with the actual data of kWh to determine the failure value or error value in predictions using mean absolute deviaton (MAD) and mean absolute percentage error (MAPE). From the third average forecasting test analysis, it produces an average MAPE value of 3 months reference with an average value of 2.922%, 6 months reference with an average value of 3.092% and a 12-month reference with an average value of 4.175%. The smallest MAPE, which is a test using a 6-month reference, produces a value of 1.886% with alpha 0.1.