Imagen de Portada

A GWO-LSTM based approach for photovoltaic power generation prediction under extreme climate conditions

Guardado en:
Publicado en:Journal of Physics: Conference Series vol. 3163, no. 1 (Dec 2025), p. 012009
Autor principal: Daoyuan, Wang
Otros Autores: Haomin, Zhang, Xi, Lin, Yu, Yu, Jie, Zheng, Jizhong, Zhu
Publicado:
IOP Publishing
Materias:
Electric power generation
Dust storms
Weather
Cluster analysis
Time series
Prediction models
Clustering
Vector quantization
Back propagation networks
Optimization
Acceso en línea:Citation/Abstract
Full Text - PDF
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
Resumen:As the participation rate of photovoltaic power generation in the entire power system continues to increase, accurate photovoltaic power prediction technology is crucial for optimizing power system scheduling. However, in recent years, the frequent occurrence of extreme weather events has significantly increased the difficulty of accurately predicting photovoltaic power generation. Regarding photovoltaic power prediction methods under extreme weather conditions, this study first reviews existing research on photovoltaic power prediction methods from three aspects: dust storms, heavy rain, and snowfall. Based on this review, a GWO-LSTM grey wolf optimization time series prediction model based on K-Means clustering is proposed. First, the K-Means clustering algorithm is used to classify weather types. Then, based on the weather classification results, the predictive performance of the GWO-LSTM grey wolf optimization time series model, random forest prediction model, BP neural network model, and LSTM model is compared. The prediction results show that the GWO-LSTM model achieves the highest prediction accuracy, with an accuracy of approximately 95% under four weather conditions. This provides effective data support for the safe and stable operation of new power systems with a high proportion of photovoltaic grid connection.
ISSN:1742-6588
1742-6596
DOI:10.1088/1742-6596/3163/1/012009
Fuente:Advanced Technologies & Aerospace Database