Scholarlink

<p>In response to increasing demand for more efficient and safe energy solutions, improving the performance of the electric power system through the advanced data-driven approaches is significant more than ever before. The knowledge for this research is important as it aims at improving the performance and stability of power systems in light of the growing challenges in power systems. It is widely known that current systems have a problem of inadequate forecasting, unpredictable resource management, and lack of data availability, which in return impacts the overall operational cost. In particular, the integration of Feed forward Neural Networks (FNN) and Generative Adversarial Networks (GAN) increases the accuracy of the forecast by a hybrid approach. For example, when it reaches an MAPE of 4.5% and RMSE of 10·2 MW in load forecasting the management of the grid is able to improve on the use of the available resources to an extent that it achieves operational cost reductions of approximately 12%. If compared with conventional approaches describing the model by an MAPE of 7% and the RMSE of 15 MW, the present work shows enhanced performance increase especially for the scenarios, which were constructed using Python frameworks for dynamic testing and validation. It also eliminates current obstacles and creates conditions for further development of the improvement of the state of power systems based on the efficiency of data processing and prediction.</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

Intelligent Data Driven Approaches to Optimizing Power System Performance

  • 期刊论文

  • 发表日期：

    2025年12月18日

查看全文
  • Zixu Zhu,
  • Xinjia Li,
  • Yupeng Zhang,
  • Yina Du,
  • Mengjia Liu,
  • Liming Wang

摘要

In response to increasing demand for more efficient and safe energy solutions, improving the performance of the electric power system through the advanced data-driven approaches is significant more than ever before. The knowledge for this research is important as it aims at improving the performance and stability of power systems in light of the growing challenges in power systems. It is widely known that current systems have a problem of inadequate forecasting, unpredictable resource management, and lack of data availability, which in return impacts the overall operational cost. In particular, the integration of Feed forward Neural Networks (FNN) and Generative Adversarial Networks (GAN) increases the accuracy of the forecast by a hybrid approach. For example, when it reaches an MAPE of 4.5% and RMSE of 10·2 MW in load forecasting the management of the grid is able to improve on the use of the available resources to an extent that it achieves operational cost reductions of approximately 12%. If compared with conventional approaches describing the model by an MAPE of 7% and the RMSE of 15 MW, the present work shows enhanced performance increase especially for the scenarios, which were constructed using Python frameworks for dynamic testing and validation. It also eliminates current obstacles and creates conditions for further development of the improvement of the state of power systems based on the efficiency of data processing and prediction.