Peramalan Beban Listrik Jangka Pendek di Jawa Timur dengan Menggunakan Metode Seasonal ARIMA dan Long Short-Term Memory

Irawan, Frendi Akbar (2024) Peramalan Beban Listrik Jangka Pendek di Jawa Timur dengan Menggunakan Metode Seasonal ARIMA dan Long Short-Term Memory. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003201072_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Energi listrik memiliki peran yang sangat penting dalam meningkatkan perekonomian suatu negara. Energi listrik merupakan tulang punggung dalam mendukung pertumbuhan industri, transportasi, komunikasi, dan sektor lainnya. Peningkatan permintaan listrik yang melebihi kapasitas pembangkit listrik yang ada menjadi permasalahan serius PT PLN dalam mendistribusikan tenaga listrik ke pelanggan. Jika listrik yang dibangkitkan lebih dari konsumsi pelanggan maka listrik akan terbuang dan daya pembangkit listrik perlu diturunkan. Apabila listrik yang dibangkitkan kurang dari konsumsi pelanggan maka dapat terjadi pemadaman listrik dan menunggu pengoperasian pembangkit cadangan. Penelitian ini bertujuan untuk melakukan peramalan beban listrik jangka pendek di Jawa Timur dengan menggunakan metode seasonal ARIMA dan long short-term memory (LSTM). Hasil analisis menunjukkan sMAPE out sample model terbaik dari model seasonal ARIMA yaitu SARIMA sebesar 0,91% untuk periode peramalan 168 tahap dan model DSARIMA sebesar 3,69% untuk periode peramalan 672 tahap. Model LSTM memiliki sMAPE out sample sebesar 5,74% untuk periode peramalan 168 tahap dan 8,16% untuk periode peramalan 672 tahap sehingga model seasonal ARIMA menjadi model terbaik untuk meramalkan beban listrik di Jawa Timur karena memiliki sMAPE out sample terendah.
========================================================================================================================
Electrical energy plays a very important role in improving the economy of a country. Electrical energy is the backbone in supporting the growth of industry, transportation, communication, and other sectors. The increase in electricity demand that exceeds the capacity of existing power plants is a serious problem for PT PLN in distributing electricity to customers. If the electricity generated is more than customer consumption, the electricity will be wasted, and the power of the power plant needs to be reduced. If the electricity generated is less than customer consumption, a power outage can occur and wait for the operation of the backup generator. This study aims to conduct short-term electricity load forecasting in East Java using the seasonal ARIMA and long short-term memory (LSTM) methods. The results of the analysis show the best out-sample sMAPE model from the seasonal ARIMA model, namely SARIMA, is 0.91% for a forecasting period of 168 stages and the DSARIMA model is 3.69% for a forecasting period of 672 stages. The LSTM model has an out-sample sMAPE of 5.74% for a forecasting period of 168 stages and 8.16% for a forecasting period of 672 stages so that the seasonal ARIMA model is the best model for forecasting electricity loads in East Java because it has the lowest out-sample sMAPE.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Beban Listrik, LSTM, Peramalan, Seasonal ARIMA, sMAPE, Electrical Load, Forecasting, LSTM, Seasonal ARIMA, sMAPE.
Subjects:	H Social Sciences > HA Statistics > HA30.3 Time-series analysis Q Science > QA Mathematics > QA280 Box-Jenkins forecasting Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Frendi Akbar Irawan
Date Deposited:	08 Aug 2024 06:17
Last Modified:	08 Aug 2024 06:17
URI:	http://repository.its.ac.id/id/eprint/114181

Actions (login required)

View Item