预应力空间网格结构张拉全过程的反馈控制研究  

Research on tensioning process feedback control for prestressed space grid structures

作　　者：冯玉龙[1] 周臻[1] 孟少平[1] 王永泉[1] 吴京[1] 

机构地区：[1]东南大学混凝土及预应力混凝土教育部重点实验室,江苏南京210096

出　　处：《建筑结构学报》2013年第10期93-100,共8页Journal of Building Structures

基　　金：江苏省自然科学基金项目(BK2010428);中央高校基本科研业务费专项资金资助

摘　　要：预应力张拉是确保预应力空间网格结构实现预期设计状态的关键,但施工误差的随机变异影响了张拉过程的可控性。为此,提出一种预应力空间网格结构张拉全过程的反馈控制方法。在考虑施工材料参数与几何参数误差的基础上,首先对基准结构模型进行概率有限元分析,获取误差结构响应与预张力控制方案的样本对;然后利用BP神经网络的非线性映射功能,以概率有限元分析结果为训练样本,构建结构响应与预张力控制值的关系模型;在实际的多阶段张拉过程中,基于前一阶段的实测结构响应数据,利用构建的神经网络关系模型预测下一阶段的张拉控制力。利用ANSYS软件的概率设计模块与MATLAB软件的神经网络工具箱,编制了反馈控制的实现程序,并针对预应力弦支穹顶结构的数值算例进行模拟分析。结果表明:利用建议的方法能够实现对预张力控制方案的动态调整,确保结构张拉完成时实现预期设计目标。利用一3 m跨弦支穹顶模型,通过误差可调节点对结构施加节点安装误差,基于实测结构杆件内力与节点位移,对结构模型的径向索张拉过程进行了索力反馈控制试验。结果表明:施工误差影响下,结构张拉完成时径向索力达到目标预应力值,验证了反馈控制方法的有效性。Pretension is crucial for prestressed space grid structures to realize the expected design state. However, the random variation of construction error reduces the controllability of tensioning process. Therefore, a tensioning process feedback control method was developed for prestressed space grid structures. Considering the construction errors of material and geometrical parameters, probabilistic finite element analysis was first conducted for basic structural model. Taking the probabilistic finite element analysis results as training samples, the strong nonlinear mapping function of BP neural network was used to construct the relation model of structural responses and tensioning control forces. In the actual multi-stage tensioning process, the neural network model was used to predict the tensioning control force in the next stage, according to the measured structural responses in the previous stage. The feedback control method was programmed by using the probability design modulus of ANSYS and the neural network toolbox of MATLAB. A prestressed suspended dome structure was analyzed as the numerical example. Analysis results show that the proposed method can realize the dynamic adjustment for the tensioning control force scheme and then ensure the achievement of expected design state. A 3 m-span suspended dome structural model was constructed. The joint installation error was applied to the structural model through the error-adjustable joint. Based on the measured member forces and joint displacements, the cable-force feedback control test was conducted for the radial-cable tensioning process by using the feedback control method. Results show that the forces in radial cables reach the target prestress value when tensioning process is finished, which verifies the effectiveness of proposed feedback control method.

关 键 词：预应力空间网格结构 张拉过程 概率有限元分析 神经网络 反馈控制 模型试验 

分 类 号：TU394[建筑科学—结构工程]