PDF全文阅读

王    凯 ，范路生 ，孙    畅 ，刘    琼

 

（ 上海理工大学 环境与建筑学院 ，上海 200093)

摘   要 ：基于决策树和随机森林两种机器学习算法分别建立再生混凝土梁抗剪承载力预测模型 ，将大量随  机配合比输入再生混凝土梁抗剪承载力预测模型中 ，确定获得最大抗剪承载力时 ，各配合比因素的最佳范围。 结果表明： 随机森林算法对于再生混凝土梁抗剪承载力的预测精度优于决策树算法 ；随机森林算法模型的均方  根误差为 0. 338 、平均绝对误差为 0. 253 、平均绝对百分比误差为 11 . 57 、拟合优度为 0. 913 ;  为获得最大抗剪承  载力 ，再生混凝土梁的最佳再生粗骨料取代率为 50% ,  最佳横向钢筋配筋率为 0. 2% ,  最佳纵向钢筋配筋率为  2. 5% ,  最佳剪跨比范围为 0. 8 ~ 1 . 1 ,   最佳混凝土抗压强度为 55 MPa。

关键词：再生混凝土梁 ；抗剪承载力 ；决策树 ；随机森林 ；预测

中图分类号：TU528 . 59            

文献标志码：A           

文章编号： 1005- 8249   (2024)  04- 0006- 09

DOI:10. 19860/j.cnki.issn1005 - 8249.2024.04.002

 

WANG Kai ,  FAN Lusheng ,  SUN Chang ,  LIU Qiong

(School of Environment and Architecture ,  University of Shanghai for Science and

Technology ,  Shanghai 200093 ,  China)

Abstract：The prediction models of recycled concrete beam shear capacity are established based on decision tree and random forest machine learning algorithms. By inputting a large number of random mix proportions into the shear bearing capacity prediction model of recycled concrete beams, the optimal range of mix proportion factors is determined to obtain the maximum shear bearing capacity. The results show that the prediction accuracy of the random forest algorithm for the shear capacity of recycled concrete beams is better than that of the decision tree algorithm. The root mean square error of the random forest algorithm model is 0.338, the mean absolute error is 0.253, the mean absolute percentage error is 11.57, and the coefficient of determination is 0.913; In order to obtain the maximum shear bearing capacity, the optimal recycled coarse aggregate replacement ratio of recycled concrete beams is 50%, the optimal transverse reinforcement ratio is 0.2%, the optimal longitudinal reinforcement ratio is 2.5%, the optimal shear span ratio is 0.8 ~ 1.1, and the optimal concrete compressive strength is 55 MPa.

Keywords: recycled concrete beam; shear capacity; decision tree; random forest; prediction

*基金项目：上海市科技创新行动计划启明星项目（22QC1400900）。

作者简介：王  凯（1997—），男，硕士研究生，研究方向为混凝土材料。

通信作者：孙  畅（1987—），女，博士，副教授，研究方向：再生混凝土结构研究。

收稿日期：2023-10-07