基于WOA-VMD结合BiTCN-BiLSTM模型的光伏功率预测

打开文本图片集

中图分类号：TM615 文献标志码：A

Abstract：Photovoltaic （PV）power generation is subject to significant volatility and intermittency due to multiple influencing factors，posing challenges to power grid stability. Thus，accurate PV power prediction is essential for effective grid management andoptimizing PV power utilization.This study introduces a PV power prediction model that combines the whale optimization algorithm（WOA）-optimized variational mode decomposition（VMD）withamixedneural network architecture，incorporating bidirectional temporal convolutional network（BiTCN）and bidirectional long short-term memory network（BiLSTM）. The WOA enhanced VMD hyperparameter optimization， improving the accuracy of PV power data decomposition. The resulting power componentsand features were then processed bythe BiTCN-BiLSTM model，leveraging the strengths of both architectures for precise power output predictions. The model was evaluated using data from a solarpower station in Xinjiang，China.It demonstrated superior performanceover alternative methods regarding mean absolute error，root mean squared error，and coefficient of determination，highlighting its efficiency and stability in PV power prediction.

Key words： photovoltaic power generation；power grid dispatching；neural network；photovoltaic power prediction；variational mode decomposition

在全球变暖问题不断升级和能源危机加剧的情况下，光伏能源的开发已成为许多国家调整能源结构的主要方向[1]。（剩余8941字）