建筑结构学报2025，Vol.46Issue(12)：83-98,16.DOI:10.14006/j.jzjgxb.2025.0068
侧向循环荷载作用下耗能钢管混凝土分体柱损伤分析及智能损伤识别系统研究
Research on damage analysis and intelligent damage detection system of energy consuming concrete filled steel tube slotted columns under lateral cyclic loading
摘要
Abstract
For large-section concrete-filled steel tubular(CFST)columns,a novel energy consuming concrete filled steel tube slotted column(E-CFSST)was proposed to address issues such as weak cooperative working performance between steel and concrete and low material utilization.To clarify the mechanical mechanism and damage characterization of this new composite column,its seismic performance and damage process were thoroughly analyzed,and an intelligent damage detection system was established based on image recognition and neural networks.Low-cycle reversed loading tests were conducted on E-CFSST specimens,and surface damage and actual damage characterization methods were proposed based on the test results.A theoretical damage model for E-CFSST was developed using large deformation plate and shell theory to optimize the extraction method of surface damage characterization parameters.An extraction model for surface damage characterization parameters was established using image recognition technology.Multiple neural network models were employed to clarify the mapping relationship between surface damage and actual damage,comprehensively constructing an intelligent damage detection system.The research results demonstrate that the proposed E-CFSST exhibits excellent seismic performance,with energy dissipation capacity increased by approximately 33％compared to structures without dampers.By combining theoretical derivation and machine learning,the established damage detection system can accurately determine the actual damage level of specimens through images,with the coefficient of determination R2 for all neural network prediction models exceeding 0.85.This intelligent damage detection system is also applicable to rapid intelligent inspection of CFST structures,providing innovative methods and theoretical support for damage detection in CFST structures.关键词
耗能钢管混凝土分体柱/图像识别/神经网络/智能损伤识别系统/抗震性能/理论损伤模型Key words
E-CFSST/image recognition/neural network/intelligent damage detection system/seismic performance/theoretical damage model分类
建筑与水利引用本文复制引用
戎翀,田文凯,史庆轩,王朋..侧向循环荷载作用下耗能钢管混凝土分体柱损伤分析及智能损伤识别系统研究[J].建筑结构学报,2025,46(12):83-98,16.基金项目
国家自然科学基金项目(52108171,52178159,52178505). （52108171,52178159,52178505）
