| Autor: |
Xidong Zheng, Feifei Bai, Tao Jin |
| Jazyk: |
angličtina |
| Rok vydání: |
2025 |
| Předmět: |
|
| Zdroj: |
International Journal of Electrical Power & Energy Systems, Vol 164, Iss , Pp 110368- (2025) |
| Druh dokumentu: |
article |
| ISSN: |
0142-0615 |
| DOI: |
10.1016/j.ijepes.2024.110368 |
| Popis: |
The demand response (DR)-considered microgrid (MG) provides a large amount of electricity consumption information, and the value of these data has attracted increasing attention because accurately identifying customers’ electricity consumption behaviour patterns helps public utilities’ dispatch planning and precise services. This paper investigates how to achieve MGs’ optimal scheduling for analysing customer-side DR identification. To maintain the economics of the MG itself from the optimal scheduling of multiple MGs and the upper-level power system (ULP), a new master–slave management (MSM) is proposed. Then, by integrating the machine learning (ML)-based classifiers, the customer-side DR identification issues caused by abnormal data, such as data missing and label errors in MGs, are solved. A case study using the China State Grid data set proves the effectiveness of the proposed MSM and DR identification strategies. The assessment reveals that the integrated classification and identification centre (ICIC) helps ensure 4.615 average electricity purchase cost assessment (EPCA) and 4.835 for electricity power assessment (EPA), which is higher than abnormal situations without machine learning-based identification. The proposed method maximises customer satisfaction while reducing MG costs. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
Full Text from ScienceDirect