基于截面恢复力模型在线更新的框架结构混合试验方法

Hybrid testing, which combines the advantages of numerical simulation and physical experimental, is an effective method to evaluate the dynamic response of structures. It is potential in assessing the seismic performance of large scale structures with the help of substructuring technique. However, limited by funding and equipment, only a few nonlinear members can be tested, and the remainder are simulated by numerical models with assumed parameters，which will bring in model errors. In this case, the advantage of substructure hybrid testing will be eclipsed if the whole structure is dominated by such numerical models. To reduce model errors and improve the accuracy of testing, hybrid testing method of frame structures with online updating of sectional restoring force model is proposed. First of all, sectional restoring force model of beam-column element considering multi-axial coupling and deteriorate effect is studied. Secondly, extended Kalman filter (EKF) is employed jointed with direct differentiation method (DDM) to identify the parameters of the sectional restoring force model. Thirdly, hybrid testing system based on finite element and considering online model updating is established. Finally, hybrid testing of frame structure considering model updating are carried out to verify the effectiveness of the new testing method and system. It is planned to make innovative research on the establishment of a sectional restoring model, formulation of a parameter identification method and development of a hybrid testing system, so as to lay a theoretical basis and offer a technical support for both analysis and testing of structures.

混合试验综合了数值模拟和物理试验的优点，是研究结构动力反应的有效方法之一。子结构技术的应用使混合试验在评估大型复杂结构抗震性能上具有很大的潜力。但限于资金和设备,只能对个别非线性构件开展试验，剩余非线性构件需用假定的数值模型模拟，这将带来模型误差，当此类构件在结构中所占比例较高时，子结构混合试验的优势就消失了。为了减少模型误差，提高试验精度，本项目提出了基于截面恢复力模型在线更新的框架结构混合试验方法。首先研究梁柱截面上考虑多轴耦合及退化效应的恢复力模型，接着把直接微分法与扩展的卡曼滤波器结合起来识别截面恢复力模型参数，然后建立基于有限元并考虑模型更新的混合试验系统平台，最后开展面向单元的框架结构模型更新混合试验，验证所提试验方法和平台的有效性。本项目拟在截面恢复力模型建立、参数识别方法推导和混合试验系统平台开发等方面做出新的研究，以期为结构分析和试验研究提供理论基础和技术支撑。

混合试验综合了数值模拟和物理试验的优点，是研究结构动力反应的有效方法之一。子结构技术的应用使混合试验在评估大型复杂结构抗震性能上具有很大的潜力。但限于资金和设备,只能对个别非线性构件开展试验，剩余非线性构件需用假定的数值模型代替，当此类构件在结构中所占比例较高时，子结构混合试验的优势就消失了。为了减少模型误差，提高试验精度，本项目提出了基于截面恢复力模型更新的框架结构混合试验方法。项目研究内容如下：（1）针对梁柱单元构建了截面上轴力与弯矩的屈服面，建立了考虑轴力与弯矩耦合的截面恢复力模型；阐述了截面状态确定和本构关系积分的方法；通过静力和动力弹塑性分析算例验证所提截面恢复力模型能兼顾计算精度与效率。（2）预设了框架柱的数值模型，推导了力法单元杆端力对截面恢复力模型参数的敏感性；提出了采用扩展的卡曼滤波器（EKF）结合直接微分法（DDM）估计梁柱单元截面恢复力模型参数的识别方法，通过数值仿真验证了该识别方法的精度、效率和鲁棒性。（3）提出了面向单元的杆系结构混合试验系统架构，嵌入参数识别与模型更新模块，形成基于有限元并考虑模型更新的混合试验系统，通过数值仿真，验证了试验系统的可靠性。（4）设计了钢框架原型结构并制作了足尺框架柱子结构试验模型，研发了平面框架柱的三个作动器几何非线性加载方案，开展了框架柱子结构混合试验和模型更新混合试验，验证了试验方法和平台是有效的。通过本项目的研究，建立了梁柱截面上考虑轴力弯矩耦合的恢复力模型，提出了基于直接微分法与扩展的卡曼滤波器的截面恢复力模型参数识别方法，建立基于有限元并考虑模型更新的混合试验系统平台，开展面向单元的框架结构模型更新混合试验，验证所提试验方法和平台的有效性。本项目的研究为结构分析和试验研究提供理论基础和技术支撑。