A Short-Term Load Demand Forecasting based on the Method of LSTM

Bodur I., Celik E., ÖZTÜRK N.

10th IEEE International Conference on Renewable Energy Research and Applications, ICRERA 2021, İstanbul, Turkey, 26 - 29 September 2021, pp.171-174 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/icrera52334.2021.9598773
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.171-174
  • Keywords: short term load forecasting, long-short-term memory, recurrent neural network
  • Gazi University Affiliated: Yes


© 2021 IEEE.Electricity energy is produced from another energy source like fossil source such as oil, coil, natural gas renewable energy sources such as hydraulic, wind, solar. Their storage in high amount is a problematic issue. Therefore, the balance between the power generation and demanded power must be satisfied at all times. This is an obligation especially for companies that generate, transmit and distribute electrical energy. In this paper, a short term load demand forecasting based on a long short term memory (LSTM) is addressed, which may help planning operators for Turkish electricity market. The results of advocated approach were compared by the ones based on recurrent neural network. As a result, it is found that the proposed LSTM approach can predict especially daily and weekly demands with an accuracy more than 90%.