基于tGSSA-DELM的短期光伏发电功率预测

Short-term Photovoltaic Power Generation Prediction Based on tGSSA-DELM

针对目前光伏发电预测的预测耗时和预测精度不足等问题,提出了一种基于皮尔逊相关性分析、改进的麻雀算法(tGSSA)和深度极限学习机(DELM)的组合预测方法。该方法首先通过皮尔逊相关性分析方法对影响光伏出力的主要因素进行筛选,然后采用黄金正弦搜索策略、自适应t分布和动态选择策略来增强麻雀算法的全局搜索能力和局部寻优能力,最后利用tGSSA群智能优化算法对DELM中的输入权重和偏置进行寻优,在得到最优输入权重和偏置的情况下对光伏发电功率进行预测。以澳大利亚某光伏站一年数据按季节划分后进行预测研究,将本文模型与DELM,SSA-DELM,GA-DELM,ABC-DELM,WOA-DELM进行预测对比,结果表明,相比于其他算法改进模型和传统模型,tGSSA-DELM在预测精度、预测稳定性和工作效率中具有较大优势,具有更强的适用性。

Targeting the problems of prediction time consuming and insufficient prediction accuracy of photovoltaic power generation prediction,this paper proposes a combined prediction method based on Pearson correlation analysis,improved sparrow algorithm(tGSSA)and depth limit learning machine(DELM).This method firstly screens the main factors influencing photovoltaic output based on Pearson correlation analysis,then uses golden sine search strategy,adaptive t distribution and dynamic selection strategy to enhance the global search ability and local optimization ability of sparrow algorithm,and finally uses tGSSA swarm intelligent optimization algorithm to optimize the input weight and bias in DELM.With the optimal input weight and bias,the photovoltaic power generation is predicted.Based on the annual data divided by seasons from a photovoltaic station in Australia,the prediction research is carried out,and the proposed model is compared with DELM,SSA-DELM,GA-DELM,ABC-DELM and WOA-DELM.The results show that compared with other improved models and traditional models,tGSSA-DELM has greater advantages in prediction accuracy,prediction stability and work efficiency,and has stronger applicability.