Prediction of power generation and maintenance using AOC‐ResNet50 network
Abstract With the continuous expansion of the photovoltaic industry, the application of solar photovoltaic power generation systems is becoming increasingly widespread. Due to the obvious intermittency and volatility of photovoltaic power generation, integration of large‐scale photovoltaic power gen...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2024-10-01
|
| Series: | IET Renewable Power Generation |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/rpg2.13081 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841546527534743552 |
|---|---|
| author | Yueqiang Chu Wanpeng Cao Cheng Xiao Yubin Song |
| author_facet | Yueqiang Chu Wanpeng Cao Cheng Xiao Yubin Song |
| author_sort | Yueqiang Chu |
| collection | DOAJ |
| description | Abstract With the continuous expansion of the photovoltaic industry, the application of solar photovoltaic power generation systems is becoming increasingly widespread. Due to the obvious intermittency and volatility of photovoltaic power generation, integration of large‐scale photovoltaic power generation into the power grid can cause certain impacts on the security and stability of the grid. Photovoltaic power prediction is essential to solve this problem, as it can improve the quality of photovoltaic grid connection, optimize grid scheduling, and ensure the safe operation of the grid. In this article, the deep learning method is selected for photovoltaic power prediction. Based on the analysis of the OctConv (Octave Convolution) network structure, the AOctConv (Attention Octave Convolution) convolutional neural network structure is proposed, which is combined with the ResNet50 backbone network to obtain AOC‐ResNet50. It is then applied to the prediction of the generation of photovoltaic power. The prediction performance is compared with the ResNet50 network and the Oct‐ResNet50 network, and it is found that the AOC‐ResNet50 network has the best prediction performance, with an MAE of only 0.176888. Based on the exemplar work, a framework is proposed to illustrate this method. Its general application is discussed. |
| format | Article |
| id | doaj-art-a7089c88447345d48e2516d8de6ef21b |
| institution | Kabale University |
| issn | 1752-1416 1752-1424 |
| language | English |
| publishDate | 2024-10-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Renewable Power Generation |
| spelling | doaj-art-a7089c88447345d48e2516d8de6ef21b2025-01-10T17:41:03ZengWileyIET Renewable Power Generation1752-14161752-14242024-10-0118142381239310.1049/rpg2.13081Prediction of power generation and maintenance using AOC‐ResNet50 networkYueqiang Chu0Wanpeng Cao1Cheng Xiao2Yubin Song3Department of Automation North China Institute of Aerospace Engineering Langfang ChinaElectrical and Electronic Teaching Center China Suntien Green Energy Corp. Ltd. Huaian ChinaDepartment of Automation North China Institute of Aerospace Engineering Langfang ChinaDepartment of Automation North China Institute of Aerospace Engineering Langfang ChinaAbstract With the continuous expansion of the photovoltaic industry, the application of solar photovoltaic power generation systems is becoming increasingly widespread. Due to the obvious intermittency and volatility of photovoltaic power generation, integration of large‐scale photovoltaic power generation into the power grid can cause certain impacts on the security and stability of the grid. Photovoltaic power prediction is essential to solve this problem, as it can improve the quality of photovoltaic grid connection, optimize grid scheduling, and ensure the safe operation of the grid. In this article, the deep learning method is selected for photovoltaic power prediction. Based on the analysis of the OctConv (Octave Convolution) network structure, the AOctConv (Attention Octave Convolution) convolutional neural network structure is proposed, which is combined with the ResNet50 backbone network to obtain AOC‐ResNet50. It is then applied to the prediction of the generation of photovoltaic power. The prediction performance is compared with the ResNet50 network and the Oct‐ResNet50 network, and it is found that the AOC‐ResNet50 network has the best prediction performance, with an MAE of only 0.176888. Based on the exemplar work, a framework is proposed to illustrate this method. Its general application is discussed.https://doi.org/10.1049/rpg2.13081photovoltaic power systemspower control |
| spellingShingle | Yueqiang Chu Wanpeng Cao Cheng Xiao Yubin Song Prediction of power generation and maintenance using AOC‐ResNet50 network IET Renewable Power Generation photovoltaic power systems power control |
| title | Prediction of power generation and maintenance using AOC‐ResNet50 network |
| title_full | Prediction of power generation and maintenance using AOC‐ResNet50 network |
| title_fullStr | Prediction of power generation and maintenance using AOC‐ResNet50 network |
| title_full_unstemmed | Prediction of power generation and maintenance using AOC‐ResNet50 network |
| title_short | Prediction of power generation and maintenance using AOC‐ResNet50 network |
| title_sort | prediction of power generation and maintenance using aoc resnet50 network |
| topic | photovoltaic power systems power control |
| url | https://doi.org/10.1049/rpg2.13081 |
| work_keys_str_mv | AT yueqiangchu predictionofpowergenerationandmaintenanceusingaocresnet50network AT wanpengcao predictionofpowergenerationandmaintenanceusingaocresnet50network AT chengxiao predictionofpowergenerationandmaintenanceusingaocresnet50network AT yubinsong predictionofpowergenerationandmaintenanceusingaocresnet50network |