Scholarlink

Load forecasting of the State Electricity grid is an important activity. For the secure and smooth operation of the electrical grid, prediction of demand ahead of time is crucial. State Load Despatch Centre (SLDC) in India is entrusted with the responsibility of managing the state grid and has to carry out load forecasting work. The main objective of this paper is to forecast week-ahead demand using the machine learning model—Long Short-Term Memory (LSTM) for the Chhattisgarh State. To meet the objective, average data of 96 blocks in a day for the years Jan-2023 and Dec-2023 is gathered and processed. The processed data is then fed to the LSTM model for training purposes. The model predicts the load for the months from January 2024 to March 2024. The accuracy of the model is measured through metrics MAE and MAPE. The Mean Absolute Error (MAE) for the training and testing phase of the model is 36.21 MW and 45.97 MW, respectively. Mean Absolute Percentage Error (MAPE) for the training phase and testing phase of the model is 0.84% and 0.94%, respectively.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Weekly Demand Load Forecasting Using LSTM Model for Chhattisgarh State

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ketan Mishra,
  • Anamika Yadav,
  • Narendra D. Londhe

摘要

Load forecasting of the State Electricity grid is an important activity. For the secure and smooth operation of the electrical grid, prediction of demand ahead of time is crucial. State Load Despatch Centre (SLDC) in India is entrusted with the responsibility of managing the state grid and has to carry out load forecasting work. The main objective of this paper is to forecast week-ahead demand using the machine learning model—Long Short-Term Memory (LSTM) for the Chhattisgarh State. To meet the objective, average data of 96 blocks in a day for the years Jan-2023 and Dec-2023 is gathered and processed. The processed data is then fed to the LSTM model for training purposes. The model predicts the load for the months from January 2024 to March 2024. The accuracy of the model is measured through metrics MAE and MAPE. The Mean Absolute Error (MAE) for the training and testing phase of the model is 36.21 MW and 45.97 MW, respectively. Mean Absolute Percentage Error (MAPE) for the training phase and testing phase of the model is 0.84% and 0.94%, respectively.