基于组合神经网络的建筑冷负荷混合预测模型

删除或更新信息，请邮件至freekaoyan#163.com(#换成@)

本站小编 Free考研考试/2024-01-16

-->

于军琪,高之坤,赵安军,周敏,虎群.基于组合神经网络的建筑冷负荷混合预测模型[J].,2022,62(5):509-517

基于组合神经网络的建筑冷负荷混合预测模型

Hybrid forecasting model of building cooling load based on combined neural network

DOI：10.7511/dllgxb202205009

中文关键词:负荷预测特征提取组合神经网络混合预测模型预测精度

英文关键词:load forecasting feature extraction combined neural network hybrid forecasting model forecasting accuracy

基金项目:咸阳机场三期扩建工程绿色能源站系统智能管控咨询与顾问项目(20210103)；国家重点研发计划资助项目(2017YFC0704100).

作者	单位
于军琪,高之坤,赵安军,周敏,虎群

摘要点击次数:183

全文下载次数:254

中文摘要:

准确的建筑冷负荷预测是实现大型中央空调系统优化运行、节能降耗的关键．为此，提出一种天牛须搜索(BAS)算法优化的广义回归神经网络(GRNN)结合长短时记忆(LSTM)神经网络的混合预测模型(BAS-GRNN&LSTM)，用于建筑冷负荷预测．首先，采用递归特征消除法得到最佳特征数量并结合随机森林算法做特征选择；再将天牛须搜索算法优化后的广义回归神经网络与长短时记忆神经网络组合，构建冷负荷混合预测模型；最后利用某大型建筑的实测数据进行了仿真实验．结果表明：天牛须搜索算法有很好的稳定性和收敛性，适用于广义回归神经网络参数优化；利用随机森林算法结合递归特征消除法提取出的特征能够更好地建立预测模型，有效增加模型预测精度；相比其他预测模型，BAS-GRNN&LSTM的预测效果更为优越，并能对不同月份冷负荷进行有效预测，泛化能力强，适用于建筑冷负荷预测．

英文摘要:

Accurate building cooling load forecasting is the key to implement the optimal operation, energy saving and consumption reduction for large central air conditioning system. Therefore, a hybrid forecasting model based on generalized regression neural network (GRNN) optimized by beetle antennae search (BAS) algorithm and long short-term memory (LSTM) neural network (BAS-GRNN&LSTM) is proposed for building cooling load forecasting. Firstly, recursive feature elimination method is adopted to obtain the optimal number of features and combined with random forest algorithm to complete feature selection. Then, the hybrid forecasting model of building cooling load is constructed by combining the GRNN optimized by the BAS algorithm with the LSTM neural network. Finally, the simulation experiment is carried out by using the measured data of a large building. The results show that the BAS algorithm has good stability and convergence, and can be applied to the parameter optimization of GRNN. The features extracted by the combination of random forest algorithm and recursive feature elimination method can better establish the forecasting model and effectively increase the forecasting accuracy of the model. Compared with other forecasting models, BAS-GRNN&LSTM presents more superior forecasting performance, and further has strong generalization ability effectively forecasting the cooling load in different months, which can be applied to building cooling load forecasting.

查看全文查看/发表评论下载PDF阅读器

关闭

基于组合神经网络的建筑冷负荷混合预测模型

本站小编 Free考研考试/2024-01-16

相关话题/

领限时大额优惠券,享本站正版考研考试资料!