Peramalan Beban Listrik Konsumen Jangka Pendek Menggunakan Metode Customer Baseline Load - Backpropagation Neural Network (CBL-BPNN)

Short-Term Load Forecasting Using Customer Baseline Load - Backpropagation Neural Network (CBL-BPNN) Methods

Energi listrik merupakan kebutuhan pokok masyarakat. Sehingga kebutuhan energi listrik dimasa yang akan datang harus bisa diramalkan untuk memenuhi permintaan energi listrik yang digunakan oleh masyarakat. Untuk meramalkan beban listrik tersebut diperlukan peramalan daya yang dikonsumsi oleh pelanggan. Metode yang digunakan untuk meramalkan beban listrik konsumen pada penelitian ini terdiri dari  Customer Baseline Load (CBL) dan Backpropagation Neural network (BPNN). Penelitian ini bertujuan untuk mengetahui konsumsi energi listrik yang digunakan oleh pelanggan wilayah kecamatan Karangpilang Surabaya barat dengan jenis tarif Bisnis dan Industri satu minggu kedepan (peramalan jangka pendek).Hasil penelitian peramalan beban listrik konsumen jangka pendek menggunakan metode hybrid Customer Baseline Load – Backpropagation Neural Network (CBL-BPNN) didpatkan hasil untuk jenis tarif Bisnis sebesar, MSE = 0.00009465  dan MAPE = 0.000269%. Sedangkan untuk jenis tarif Industri sebesar, MSE = 0.000096684 dan MAPE = 0.000245%.  Sehingga dengan menggunakan metode hybrid Customer Baseline Load – Backpropagation Neural Network (CBL-BPNN) didapatkan peramalan yang lebih akurat selama satu minggu kedepan.

Kata Kunci : Customer Baseline Load, Backpropagation Neural Network, Beban listrik, peramalan jangka pendek

Electrical energy is one of the primary needs of society. In order to meet the demand for said energy used by society, the ability to forecast the energy consumed is needed. Furthermore, the forecast of energy used by customers is needed to forecasting said electrical load.The method used to forecast the amount of consumer electricity load in this research comprises of Customer Baseline Load (CBL) and Backpropagation Neural network (BPNN). This researches purpose is to find out the electrical energy consumed by customers in the sub-district of Karangpilang, West Surabaya with the fare type of Business and industry for a week ahead (short-term forecasting).In this research of short-term consumer electrical load forecast using hybrid Customer Baseline Load – Backpropagation Neural Network (CBL-BPNN) method obtained the result for Business-type fare in the amount of MSE = 0.00009465  and MAPE = 0.000269%. for Business-type fare in the amount of MSE = 0.000096684 and MAPE = 0.000245%.   By using the hybrid Customer Baseline Load – Backpropagation Neural Network (CBL-BPNN) method, obtaining the forecasting for the week ahead is more accurate than ever.

Keywords: Customer Baseline Load, Backpropagation Neural Network, Electricity load, short-term Forecasting