基于数值天气预报因子扩充和改进集成学习的高寒地区短期光伏功率预测

Short term photovoltaic power prediction in high-altitude and cold regions based on numerrical weather prediction factor expansion and improved ensemble learning

高寒地区光伏系统因气象条件影响,其光伏功率表现出更显著的波动性。本文以黑龙江某光伏电站为例,提出基于数值天气预报(NWP)因子扩充和改进常规Stacking集成学习的高寒地区短期光伏功率预测方法。针对高寒地区光伏功率波动大的特点,引入NWP差分因子作为交叉特征,提升模型对天气变化的敏感性。随后,以极致梯度提升(XGBoost)和长短期记忆(LSTM)网络为基学习器,时间卷积网络(TCN)为元学习器,构建集成学习模型,并利用前向验证优化模型结构。最后,进行对比实验分析,结果表明本文所提方法具有更高的预测准确度和稳定性。

Due to meteorological conditions,photovoltaic systems in high-altitude and cold regions exhibit more significant fluctuations in their photovoltaic power.This article takes a photovoltaic power station in Heilongjiang as an example and proposes a short-term photovoltaic power prediction method for high-altitude and cold regions based on numerical weather prediction(NWP)factor expansion and improved conventional Stacking ensemble learning.In response to the large fluctuations in photovoltaic power in high-altitude and cold regions,the NWP differential factor is introduced as a cross feature to enhance the sensitivity of the model to weather changes.Subsequently,an ensemble learning model is constructed using extreme gradient boosting(XGBoost)and long short term memory(LSTM)network as base learners,and temporal convolutional network(TCN)as meta learners,and the model structure is optimized using forward validation.Finally,comparative experimental analysis is conducted,and the results show that the proposed method has higher prediction accuracy and stability.