Scholarlink

<p>In the Brazilian power system, challenges are inherent in forecasting thermal generation dispatch, which plays a critical role in ensuring the efficient operation of electrical power systems. Traditional forecasting methods often struggle to capture the dynamic and nonlinear nature of such systems, leading to inaccuracies and inefficiencies. Here, a hybrid group method of data handling (HGMDH) is proposed based on the combination of denoising techniques and a prediction model. By integrating denoising techniques with the prediction model, the proposed approach aims to overcome the limitations of individual methods, enhancing forecasting accuracy and adaptability to different operating conditions. The effectiveness of the HGMDH approach is evaluated through simulations and comparisons with the Christiano–Fitzgerald filter considering a maximum period of oscillations equal to nine. With a mean absolute percentage error of 0.00988, the HGMDH demonstrates superior performance in predicting thermal generation dispatch in electrical power systems. This innovative hybrid approach presents a promising avenue for improving the efficiency and reliability of time-series forecasting in complex energy systems.</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

Hybrid group method of data handling for time-series forecasting of thermal generation dispatch in electrical power systems

  • 期刊论文

  • 发表日期：

    2025年6月28日

查看全文
  • William Gouvêa Buratto,
  • Rafael Ninno Muniz,
  • Rodolfo Cardoso,
  • Ademir Nied,
  • Carlos Tavares da Costa Jr,
  • Gabriel Villarrubia Gonzalez

摘要

In the Brazilian power system, challenges are inherent in forecasting thermal generation dispatch, which plays a critical role in ensuring the efficient operation of electrical power systems. Traditional forecasting methods often struggle to capture the dynamic and nonlinear nature of such systems, leading to inaccuracies and inefficiencies. Here, a hybrid group method of data handling (HGMDH) is proposed based on the combination of denoising techniques and a prediction model. By integrating denoising techniques with the prediction model, the proposed approach aims to overcome the limitations of individual methods, enhancing forecasting accuracy and adaptability to different operating conditions. The effectiveness of the HGMDH approach is evaluated through simulations and comparisons with the Christiano–Fitzgerald filter considering a maximum period of oscillations equal to nine. With a mean absolute percentage error of 0.00988, the HGMDH demonstrates superior performance in predicting thermal generation dispatch in electrical power systems. This innovative hybrid approach presents a promising avenue for improving the efficiency and reliability of time-series forecasting in complex energy systems.