摘要
目前,多数在冷负荷的预测模型中,由于预测模型的输入量和输出量的相关性较差、输入变量相互影响,信息冗余度较高,导致模型在预测精度和运算速度等方面均不理想。针对上述问题,首采用主成分分析法筛选出与空调冷负荷相关性高的变量将其作为模型的输入变量,随后使用遗传算法对网络参数进行优化,进而预测冰蓄冷空调系统的动态冷负荷。研究中分别建立三种不同的预测模型进行对比实验,相对于传统神经网络预测模型,改进模型提高了预测精度,加快了预测速度,可以较好地应用于实际项目中。
At present,most of the prediction models of cooling load,due to the poor correlation between the input and output of the prediction model,the mutual influence of input variables,and the high degree of information redundancy,the model is not ideal in terms of prediction accuracy and computing speed.In response to the above problems,this paper first uses principal component analysis to screen out variables that are highly correlated with air conditioning cooling load and use them as input variables of the model,and then uses genetic algorithms to optimize network parameters to predict the dynamic cooling load of the ice storage air conditioning system.In this paper,three different prediction models are established for comparison experiments.Compared with the traditional neural network prediction model,the improved model improved the prediction accuracy and speeded up the prediction,which can be better applied to actual projects.
出处
《工业控制计算机》
2021年第8期63-66,共4页
Industrial Control Computer
基金
安徽建筑大学智能建筑与建筑节能安徽省重点实验室开放课题:基于群智能的高层建筑集中供暖系统控制与优化研究(IBBE2018KX10)。
关键词
主成分分析
灰色神经网络预测
冰蓄冷空调
遗传算法
PCA
grey neural network prediction
ice storage air conditioning
genetic algorithm
