一种非迭代式的分布式子结构混合试验方法

"A NON-ITERATIVE DISTRIBUTED SUBSTRUCTURE ONLINE HYBRID TEST METHOD "

  • 摘要: 该文针对混合试验中的子结构试验积分算法进行研究,提出了一种基于模型的预测-修正算法(M-PC)。M-PC在预测时采用初始刚度求解动力方程以得到预测边界位移,与差分得到的预测位移相比,具有较高的精度和稳定性。M-PC是一种非迭代式的算法,对于子结构边界自由度较多的混合试验,该算法具有效率高的特点。该文首先介绍了M-PC积分算法的基本公式和进行子结构混合试验的具体步骤,然后通过编程实现了与OpenSEES和ABAQUS两种有限元软件的通用接口、电液伺服式加载装置的控制程序以及网络通讯功能,最后通过门式框架的子结构数值试验和物理试验对M-PC算法进行验证。结果表明:M-PC算法有效地预测静力边界位移,结构整体响应合理,正确地模拟了各子结构的非线性行为。

     

    Abstract: Proposed in this paper is a model-based predicting and correcting algorithm (M-PC) which serves as the time integration algorithm for substructure hybrid tests. M-PC adopts the initial stiffness to solve the equation of motion to predict the boundary displacements. Compared with the difference-based predictor, the model-based predictor is characterized by a better accuracy and stability. M-PC is a non-iterative algorithm which is much efficient particularly for the cases with many degrees of freedom associated with substructure boundaries. The M-PC algorithm is first formulated and procedures to implement hybrid test are given. Then a hybrid test framework is developed to realize the interface to two finite element programs, i.e., OpenSEES and ABAQUS. The functions to control servo-hydraulic devices and exchange data via Internet are also integrated into the framework. Finally, a numerical hybrid simulation and physical hybrid test are conducted on a portal moment frame. Both results demonstrate that the M-PC algorithm is able to effectively predict the displacement associated with static boundaries. The responses of the entire structure are reasonable and correctly reproduce the nonlinearities of two substructures.

     

/

返回文章
返回