Peramalan Beban Listrik Jangka Pendek di Jawa Timur Menggunakan Model Hybrid Long Short Term Memory Artificial Neural Network (LSTM-ANN)

Jurnal Mahasiswa TEUB
Vol. 12 No. 5 (2024)

This research focuses on developing a hybrid Long Short-Term Memory-Artificial Neural Network (LSTM-ANN) model to short-term electricity load forecasting in East Java. The model is tested and compared with LSTM and ANN models to evaluate its performance. This research uses historical electricity load data from PLN UP2B East Java and PJM Interconnection East Region. The results show that the hybrid LSTM-ANN model provides the best performance in forecasting short-term electricity load, especially for predicting multiple time steps ahead. The complexity of the architecture and the ability to utilize sequential data are the key advantages of this model. Thus, the hybrid LSTM-ANN model can be an effective tool to improve the efficiency and reliability of the electricity system in East Java. Keywords— forecasting, electricity load, LSTM, ANN, Hybrid model.