预应力撑杆构件交互屈曲与承载力智能估计方法

Interactive Buckling in Prestressed Stayed Members and Load-Carrying Capacity Intelligent Evaluation Method

预应力撑杆构件造型优美、省材高效,但由于其柔度较大,易产生多个屈曲模态交互耦合即“交互屈曲”模式的破坏,且交互屈曲对其极限承载力、屈曲性能影响显著。研究了交互屈曲对预应力对称、不对称撑杆构件在轴压、压弯作用下屈曲性能、极限承载力的影响,并据此采用基于回归算法的人工神经网络(ANN)建立了其极限承载力估计模型。研究结果表明,对于预应力撑杆构件,模态交互易造成不稳定后屈曲,且对其极限承载力有显著的削弱效应;该削弱效应是否会产生影响由结构几何参数与参与交互的模态屈曲荷载决定,其削弱程度受结构撑杆不对称性、预应力水平影响;所建立的极限承载力ANN预测模型可考虑交互屈曲的影响,且准确、可靠、适用范围广泛。研究可为后续开发预应力撑杆构件智能设计软件平台提供基础,满足工程设计的需求。

Prestressed stayed members have beautiful shape and high material efficiency,but they are susceptible to multi-modal interaction failure so-called"interactive buckling"owing to their slender configuration,which can affect their loadcarrying capacity and buckling behaviour significantly.This paper investigates the interactive buckling in the symmetric and asymmetric prestressed stayed members as columns and beam-columns,such as its effects on buckling behaviour and load-carrying capacity,and develops the load-carrying capacity evaluation model via artificial neural network(ANN)based on regression algorithm.The results show that,for the prestressed stayed member,modal interaction can lead to unstable post-buckling behaviour and significant load-carrying capacity reduction.This reduction is dominated by structural geometric parameters and the buckling loads of interactive modes,and the reduction degree is affected by stay asymmetry and prestressing levels.The proposed ANN model for load-carrying capacity evaluation can take interactive buckling into accounts,and its accuracy,reliability and applicability areacceptable. This study can provide research foundation for developing the intelligent design software of theprestressed stayed member,which can satisfy the requirement for structural design in engineering practices.