摘要
混凝土碳化深度的影响因素众多又复杂且对时间有较强的依赖性,现有的混凝土碳化深度预测方法不能把所有影响因素充分考虑进去,导致其预测精度不够准确。提出一种BP-AR融合算法,该算法利用BP神经网络预测碳化深度,再通过时间序列方法对预测值进一步修正。通过试验分析,时间序列方法能够通过BP神经网络预测的碳化深度值发现碳化反应随时间变化的规律,BP-AR算法比BP神经网络预测碳化深度精度更高,弥补了因数据量有限而造成较大的预测误差。
The influence factors of concrete carbonation depth are numerous and complex,and carbonation reaction has strong dependence on time.The accuracy of existing concrete carbonation depth prediction methods is not accurate enough,because all factors can not be taken into account.The BP-AR fusion algorithm is proposed.In this algorithm,the BP neural network is used to predict the carbonation depth,and then the prediction value is further corrected by time series method.It has been experimentally verified that by using the time series method,the regularity of carbonation reaction with time can be found through the carbonation depth value predicted by BP neural network.The BP-AR algorithm predicts the carbonation depth more accurately than BP neural network,and makes up for the large prediction errors caused by limited data volume.
作者
杨雪华
董振平
于军琪
赵安军
魏廷剑
YANG Xuehua;DONG Zhenping;YU Junqi;ZHAO Anjun;WEI Tingjian(School of Information and Control Engineering,Xi′an University of Architecture and Technology,Xi′an 710055,China;School of Civil Engineering,Xi′an University of Architecture and Technology,Xi′an 710055,China)
出处
《混凝土》
CAS
北大核心
2019年第11期6-9,共4页
Concrete
基金
国家重点研发计划项目（2016YFC0701309-02）
关键词
碳化深度
BP神经网络
时间序列
预测
carbonation depth
BP neural network
time series
prediction
