|本期目录/Table of Contents|

[1]于晓辉,陈玉琛,代旷宇.基于机器学习的爆炸荷载下RC板最大位移响应分析[J].建筑科学与工程学报,2026,(01):85-94.
　YU Xiaohui,CHEN Yuchen,DAI Kuangyu.Maximum displacement response analysis of RC slabs under blast load based on machine learning method[J].Journal of Architecture and Civil Engineering,2026,(01):85-94.
点击复制

基于机器学习的爆炸荷载下RC板最大位移响应分析(PDF)
分享到：

《建筑科学与工程学报》[ISSN:1673-2049/CN:61-1442/TU]

卷:
期数:
2026年01期
页码:
85-94
栏目:
智能检测与建造技术专栏
出版日期:
2026-01-20

文章信息/Info

Title:
Maximum displacement response analysis of RC slabs under blast load based on machine learning method
作者:
于晓辉，陈玉琛，代旷宇
1. 桂林理工大学 广西绿色建材与建筑工业化重点实验室,广西 桂林 541004; 2. 郑州大学 土木工程学院,河南 郑州 450001
Author(s):
YU Xiaohui, CHEN Yuchen, DAI Kuangyu
1. Guangxi Key Laboratory of Green Building Materials and Construction Industrialization, Guilin University of Technology, Guilin 541004, Guangxi, China; 2. School of Civil Engineering, Zhengzhou University, Zhengzhou 450001, Henan, China
关键词:
RC板; 爆炸荷载; 最大位移响应; 可解释性机器学习; PSOXGBoost方法
Keywords:
RC slab;  blast load;  maximum displacement response;  interpretable machine learning;  PSO-XGBoost method
分类号:
-
DOI:
-
文献标志码:
A
摘要:
收集既有试验及数值模拟结果，建立了包含491种钢筋混凝土(RC)板在爆炸荷载作用下的位移响应数据库。采用板的长度、宽度、厚度、混凝土抗压强度、钢筋屈服强度、配筋率、边界条件、板的类型、爆炸距离和爆炸当量10个影响因素作为输入参数，采用3类共9种机器学习方法，分别建立了RC板在爆炸荷载下最大位移响应预测模型。采用可解释性机器学习方法，通过特征重要性分析、单因素部分依赖分析及交互性依赖分析，对所建立的机器学习模型进行解释，并对RC板在爆炸荷载下最大位移响应的影响因素的重要性进行了分析。结果表明：基于粒子群优化极端梯度增强方法(PSOXGBoost)的预测模型精度最高，且精度高于规范推荐的等效单自由度模型结果；在所考虑的影响因素中，爆炸当量、爆炸距离、板的厚度及配筋率对RC板在爆炸荷载作用下的最大位移响应影响最显著；RC板的抗爆设计应保证最小板厚达到150 mm，最小配筋率达到1.5%，且混凝土强度应达到50 MPa。
Abstract:
A displacement response database for 491 types of reinforced concrete (RC) slabs under blast loads was established based on the collection of existing experimental and numerical simulation results. Using 10 influencing factors including the length, width, thickness, compressive strength of concrete, yield strength of steel bars, reinforcement ratio, boundary conditions, type of slab, blast distance, and blast equivalent as input parameters, a total of 9 machine learning methods from 3 categories were used to establish prediction models for the maximum displacement response of RC slabs under blast loads. Using interpretable machine learning methods, the established machine learning model was explained through feature importance analysis, single factor partial dependency analysis, and interactive dependency analysis, and the importance of the influencing factors on the maximum displacement response of RC slabs under blast loads was analyzed. The results show that the prediction model based on particle swarm optimizationextreme gradient boosting (PSO XGBoost) has the highest accuracy, and the accuracy is higher than that of the equivalent single degree of freedom model recommended by the standard. Among the considered influencing factors, blast equivalent, blast distance, slab thickness and reinforcement ratio have the most significant impact on the maximum displacement response of RC slabs under blast loads. The blast resistant design of RC slabs should ensure a minimum thickness of 150 mm, a minimum reinforcement ratio of 1.5%, and a concrete strength of 50 MPa.

参考文献/References:

-

相似文献/References:

[1]曹雪叶,赵均海,李 艳.爆炸荷载作用下钢管混凝土柱非线性分析[J].建筑科学与工程学报,2015,32(01):58.
　CAO Xue-ye,ZHAO Jun-hai,LI Yan.Nonlinear Analysis of Concrete-filled Steel Tubular Columns Under Blast Load[J].Journal of Architecture and Civil Engineering,2015,32(01):58.
[2]姚宇飞,师燕超,李忠献.爆炸荷载下钢筋混凝土框架结构连续倒塌分析方法比较[J].建筑科学与工程学报,2015,32(01):64.
　YAO Yu-fei,SHI Yan-chao,LI Zhong-xian.Comparison of Progressive Collapse Analysis Methods for RC Frame Structures Under Blast Loads[J].Journal of Architecture and Civil Engineering,2015,32(01):64.
[3]田 力,范其华.多层框架结构在其地下室内部爆炸冲击下的连续倒塌机理研究[J].建筑科学与工程学报,2016,33(01):46.
　TIAN Li,FAN Qi-hua.Research on Progressive Collapse Mechanism of Multi-layered Frame Structure Under Blast Impact in Internal of Basement[J].Journal of Architecture and Civil Engineering,2016,33(01):46.

备注/Memo

备注/Memo:
-
更新日期/Last Update: 2026-01-20