Work

Clustering of distributed photovoltaic typical days, variable mode decomposition, and high-accuracy power forecasting

Year: 2026
Type: article
Abstract: Introduction Under the global transition of the energy structure, photovoltaic (PV) generation has developed rapidly due to its abundant resources and low-carbon characteristics. However, the intermittency and volatility of PV power output pose significant challenges to the stable operation of power grids, making high-accuracy short-term power forecasting essential. Existing forecasting approaches—including physical models, statistical methods, and hybrid techniques—still face limitations, such as the need to predefine the number of clusters and the difficulty of determining signal decomposition parameters. Methods To address these issues, this study proposes a distributed photovoltaic short-term power forecasting model that integrates weather clustering, signal decomposition, and a hybrid neural network. The method first preprocesses PV and meteorological data, employs a Gaussian Mixture Model to perform weather-scenario clustering, and applies Whale Optimization Algorithm–optimized Variational Mode Decomposition to process PV power sequences. Subsequently, a forecasting model combining a Temporal Convolutional Network and a Bidirectional Long Short-Term Memory network is constructed, and its hyperparameters are fine-tuned using the Hunger Games Search optimization algorithm. Results Experimental results demonstrate that the proposed model significantly enhances forecasting accuracy and stability, effectively overcoming the shortcomings of conventional methods. Discussion The proposed model provides valuable technical support for large-scale grid integration and consumption of distributed PV systems.
Language: English
Field: Energy
Open Access status: gold
Estimated APC paid: $3,077
OpenAlex ID: W7162767599
FWCI:
Cites: 19
Cited by:
Related to: 10