基于GTO优化的VMD-CNN-GRU光伏发电功率预测

A PV power generation prediction based on the VMD-CNN-GRU model optimized by GTO

精确地预测光伏发电功率是保证电力系统稳定运行的关键。为改善光伏发电功率的预测的准确性,通过引入人工大猩猩部队优化(artificial gorilla troops optimizer,GTO)算法和变分模态分解(variational mode decomposition,VMD),提出了一种基于卷积神经网络(convolutional neural networks,CNN)和门控循环单元(gated recurrent unit,GRU)神经网络的组合预测模型(GTO-VMD-CNN-GRU)。研究基于皮尔逊相关系数的气象特征量提取方法,获取特征重要性并作为模型输入,针对VMD和模型参数手动设置的复杂性和不确定性,利用GTO对变分模态分解数量和惩罚因子进行寻优来确定最优组合,并对CNN-GRU模型主要超参数进行寻优。对光伏输出功率的预测进行分析,结果表明,GTO-VMD-CNN-GRU预测模型能有效提升光伏输出功率预测精度,再通过与其他4种方法的预测效果比较,发现所提方法各项误差指标表现最好,因此,优化后的模型可靠性更强。

Accurate prediction of photovoltaic(PV)power is the key to ensuring the stable operation of the power system.To improve the accuracy of PV power prediction,by introducing an artificial gorilla troops optimizer(GTO)algorithm and variational mode decomposition(VMD),a combined prediction model(GTO-VMD-CNN-GRU)based on convolutional neural networks(CNN)and gated recurrent unit(GRU)neural networks was proposed.In this study,the quantitative meteorological feature extraction method based on Pearson s correlation coefficient was used to obtain feature importances for use as model inputs.To address the complexity and uncertainty of manual settings of VMD and model parameters,GTO was used to optimize the number of VMD and penalty factors to determine the optimal combination,and the main hyperparameters of the CNN-GRU model was optimized.By analyzing the predictions of PV output power,the results show that the CTO-VMD-CNN-GRU prediction model effectively improves the accuracy of PV output power predictions.By comparing the prediction effects with those of the other four methods,it was found that the proposed method performed the best in every error index.Therefore,the optimized model is more reliable.