摘要
在分析影响光伏阵列发电功率的天气因素基础上,提出了一种基于天气预报信息和相似日分类选取的反馈型神经网络预测模型。运用天气预报和相似日评价函数,从天气的历史数据中选择出天气相似日,结合相似日的发电功率和相似日与预测日的天气数据建立Elman神经网络预测模型。实验结果表明,模型预测精度较好,具备一定有效性。
With the analysis of factors effecting the generated power of PV arrays,a feedback neural network prediction model was proposed based on the weather forecasting information and similar-day selection method in this paper.In the presented methodology,the similar-day was picked from the historical data using the evaluation function and the weather forecasting information.In addition,the applied Elman neural network prediction model was built based on the power generation data in the similar-day and the weather data of similar-day and the predicted date.The experiments show that the forecasting model has high prediction accuracy.
出处
《杭州电子科技大学学报(自然科学版)》
2015年第5期74-78,共5页
Journal of Hangzhou Dianzi University:Natural Sciences
基金
浙江省自然科学基金资助项目(LQ12E07001)
浙江省重大科技专项资助项目(2009C11020)