大跨度空间钢结构震后损伤类型快速判别研究

Seismic damage analysis shows that after earthquakes there are element damage, joint damage and support damage for space steel structures and there are different security risks corresponding to different damage types. Damage type recognition is a fundamental problem in damage identification research field. At present, the damage location and damage degree detection are studied supposing the damage type is known, and the study on the damage type rapid identification is relatively less. In this project, according to characteristics of space steel structures, the dynamic property of space steel structure under different type damage is studied firstly and the sensitive and anti-noise damage index is proposed using the structural acceleration response during two key stages before-and-after earthquakes. The mathematical model between the damage index and damage types is built by pattern recognition method and neural network, and the fast detection problem of damage type is solved through numerical simulation and dynamic experiment. The research results can be used as the theoretical basis and technological supports for structural safety evaluation and the solution of this problem would be important to improve health monitoring system of space structures.

大量震害分析表明，空间钢结构震后常出现杆件损伤、节点损伤和支座损伤三种损伤类型，不同的损伤类型对应不同的结构安全隐患，损伤类型的判别是开展结构健康监测研究需要先行解决的基础问题。目前结构健康监测研究多为已知损伤类型来判断损伤位置和损伤程度，而对损伤类型的快速判别的研究相对较少。本项目针对空间钢结构节点和杆件众多的特点，通过多传感器获得的结构震前和震后两个关键阶段的动力响应监测数据，从空间钢结构各种损伤类型的动力特性分析入手，建立适用于判别空间钢结构各种损伤类型的灵敏性高、抗噪性能好的损伤指标。通过数值模拟和动力试验，基于模式识别和神经网络技术，建立损伤指标和各种损伤类型之间的数学模型，解决空间钢结构震后损伤类型快速判别的问题。该问题的解决将为震后空间结构的快速评估提供理论指导和技术支持，完善空间结构健康监测研究体系。

空间钢结构震后常出现杆件损伤、节点损伤和支座损伤三种损伤类型，不同的损伤类型对应不同的结构安全隐患，损伤类型的判别是开展结构健康监测研究需要先行解决的基础问题本项目针对空间钢结构节点和杆件众多的特点，通过获得的结构震前和震后两个关键阶段的加速度响应监测数据，从空间钢结构各种损伤类型的动力特性分析入手，已完成构建适用于判别空间钢结构各种损伤类型的灵敏性高、抗噪性能好的损伤指标。并通过数值模拟和动力试验，基于模式识别和神经网络技术，建立了损伤指标和各种损伤类型之间的数学模型，完成构建损伤类型判别的MATLAB软件操作平台，已建立空间钢结构震后损伤类型快速判别方法。通过单层网壳振动试验和振动台试验数据，对提出方法的有效性和工程应用性完成试验验证。该问题的解决为震后空间结构的快速评估提供理论指导和技术支持，完善空间结构健康监测研究体系。