Scholarlink

<p>The forecasting of electrical load plays vital role for the successful operation and management as well as maintaining a balance between supply and demand in modern power systems. This paper presents a Stacked Long Short-Term Memory (SLSTM) neural network model particularly developed for short-term load forecasting in the Ontario real-time electricity market. The SLSTM model uses historical hourly load data together with weather and time parameters to represent complicated temporal dynamics through a multi-layered structure. Comprehensive simulations are performed, including seasonal analysis, varying forecast horizons, and benchmarking against Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Auto-Regressive Integrated Moving Average (ARIMA) models. The SLSTM always gets better results, as shown by the Mean Absolute Percentage Error (MAPE) and the Root Mean Squared Error (RMSE). Shapley Additive explanations (SHAP) analysis helps us understand important factors like temperature and the day of the week. The results confirm that the Stacked LSTM (SLSTM) is a strong and useful technique to forecast load in complex energy markets.</p>

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

A stacked long short-term memory model for electrical load forecasting in the Ontario electricity market

  • 期刊论文

  • 发表日期：

    2025年12月1日

查看全文
  • Sultana Parween,
  • Md Irfan Ahmed,
  • Harsh Wardhan Pandey,
  • Md Nasim Ansari

摘要

The forecasting of electrical load plays vital role for the successful operation and management as well as maintaining a balance between supply and demand in modern power systems. This paper presents a Stacked Long Short-Term Memory (SLSTM) neural network model particularly developed for short-term load forecasting in the Ontario real-time electricity market. The SLSTM model uses historical hourly load data together with weather and time parameters to represent complicated temporal dynamics through a multi-layered structure. Comprehensive simulations are performed, including seasonal analysis, varying forecast horizons, and benchmarking against Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Auto-Regressive Integrated Moving Average (ARIMA) models. The SLSTM always gets better results, as shown by the Mean Absolute Percentage Error (MAPE) and the Root Mean Squared Error (RMSE). Shapley Additive explanations (SHAP) analysis helps us understand important factors like temperature and the day of the week. The results confirm that the Stacked LSTM (SLSTM) is a strong and useful technique to forecast load in complex energy markets.