摘要
针对目前桥梁结构安全状态评估分级方法主观性强、费用较高的问题,本文采用基于人工神经网络改进响应面法进行结构极限承载力可靠度分析,提出了基于可靠度指标的有效极限承载能力比作为结构安全状况分级指标,以实现桥梁安全状况的定量分析。以一座钢管混凝土拱-连续梁组合体系桥梁为例,计算了钢管混凝土拱、预应力混凝土主梁、吊杆三种构件的极限承载能力可靠度指标,得到的以有效极限承载能力比为指标的结构状态评估结果与规范评定结果误差在5%以内,说明该方法能够有效实现结构安全状况分级评定。
Current methods of evaluating bridge safety are either subjective or expensive. In this study, we applied an improved response surface method based on the neural network to analyze the structural ultimate bearing capacity reliability. We defined an effective ultimate bearing capacity ratio based on the reliability index to evaluate the safety conditions of damaged structures. A quantitative analysis was achieved. Considering a concrete-filled steel tubular （CFST） arch/continuous beam bridge as an example, we calculated the reliabilities of the ultimate bearing capacity of a CFST arch and its pre-stressed concrete beam and suspenders. The results of the bridge safety evaluation based on the effective ratio of the ultimate bearing capacity coincide well with the specifications, with errors of less than 5%. This method can effectively evaluate safety conditions with respect to structural safety.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2016年第4期550-555,共6页
Journal of Harbin Engineering University
基金
国家自然科学基金项目(11202080)
关键词
神经网络
响应面法
可靠度
有效极限承载能力比
安全状况评级
neural network
response surface method
condition evaluation reliability
effective ultimate bearing capacity ratio
safety
