Scholarlink

<p>This paper addresses the identification and classification of distributed generation (DG) connected to the secondary distribution network based on the non-intrusive load monitoring framework. We built a new public dataset with real-world data comprising samples of electrical variables aggregating loads and distributed generation data. Traditionally, NILM methods are concerned with disaggregating, identifying, and classifying electrical loads. On the other hand, behind the meter (BTM) estimation methods separate the consumption of electrical loads from the DG power generated by prosumers. Our work expands the traditional NILM and BTM analysis, presenting an ablation study of DG’s impact on the identification of electrical loads and the impact that aggregate loads represent for the identification of DG. We use state-of-the-art deep learning-based methods for disaggregation and classification on our new dataset and achieved up to 100% F1-Score for DG identification and up to 98% F1-Score for load disaggregation with the presence of DG. Data and codes are fully available at <a href="https://github.com/evertoneie/DG-NILM">https://github.com/evertoneie/DG-NILM</a>.</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

Load Identification with Photovoltaic Distributed Generation and a Novel Public High-Frequency Dataset

  • 期刊论文

  • 发表日期：

    2025年2月17日

查看全文
  • Everton L. de Aguiar,
  • Ricardo Bernardi,
  • André E. Lazzaretti,
  • Daniel R. Pipa,
  • Emerson G. Carati,
  • Rafael Cardoso

摘要

This paper addresses the identification and classification of distributed generation (DG) connected to the secondary distribution network based on the non-intrusive load monitoring framework. We built a new public dataset with real-world data comprising samples of electrical variables aggregating loads and distributed generation data. Traditionally, NILM methods are concerned with disaggregating, identifying, and classifying electrical loads. On the other hand, behind the meter (BTM) estimation methods separate the consumption of electrical loads from the DG power generated by prosumers. Our work expands the traditional NILM and BTM analysis, presenting an ablation study of DG’s impact on the identification of electrical loads and the impact that aggregate loads represent for the identification of DG. We use state-of-the-art deep learning-based methods for disaggregation and classification on our new dataset and achieved up to 100% F1-Score for DG identification and up to 98% F1-Score for load disaggregation with the presence of DG. Data and codes are fully available at https://github.com/evertoneie/DG-NILM.