钢纤维混凝土梁抗剪强度的集成学习预测方法

An Ensemble Learning Prediction Method for Shear Strength of Steel Fiber Reinforced Concrete Beams

传统经验方法预测钢纤维混凝土梁的抗剪强度多基于有限的数据且难处理强非线性问题,因此难以有效快速地预测抗剪强度。为此,开发了一种基于集成学习的钢纤维混凝土梁的抗剪强度预测方法,数据包含来源于文献的330组梁。输入参数包含梁的有效尺寸、纵向配筋率、混凝土抗压强度、骨料尺寸、剪跨比、钢纤维系数及抗拉强度。首先对数据集进行热力图与相关系数分析,得到各输入参数之间几乎无冗余,均可作为有效输入参数进行建模,并得到了输入与输出参数之间的线性关系程度;然后将数据集分割为测试集和训练集,分别用于不同的集成学习模型进行计算并记录运行过程;最后将结果与传统回归方法进行比较。结果表明,集成学习中GradientBoost算法预测抗剪性能的准确度最高,达到了0.950,比传统回归方法的平均准确度要高,证明该预测方法可用于抗剪性能预测。

Traditional empirical methods to predict the shear strength of steel fiber reinforced concrete(SFRC)beams are based on limited data and difficult to deal with strong nonlinear problems,so it is difficult to predict the shear strength effectively and quickly.In this paper,A shear strength prediction method of SFRC beams based on ensemble learning was developed.The data included 330 groups of beams from literature.The input parameters included the effective size of the beam,longitudinal reinforcement ratio,concrete compressive strength,aggregate size,shear-span to effective depth ratio,steel fiber factor and tensile strength of fibers.Firstly,the thermodynamic diagram and correlation coefficient of the data set were analyzed,and it was found that there was almost no redundancy between the input parameters,which could be used as effective input parameters for modeling,and the linear relations hip between the input and output parameters was obtained;then the data set was divided into test set and training set for different ensemble learning models to calculate and record the running process;finally,the results were compared with traditional regression methods.The results showed that the accuracy of the GradientBoost algorithm in predicting the shear strength in the ensemble learning was the highest,reaching 0.950,which was higher than the average accuracy of the traditional regression method.It proved that the method could be used to predict the shear strength of SFRC.