基于滑窗子空间算法的桥梁运营状态预警分析研究

以某典型斜拉桥振动台模型试验为研究背景,采用滑窗技术与数据驱动随机子空间识别算法对桥梁运营状态预警方法进行研究。为实现在时域上对桥梁模态参数进行实时跟踪识别,将滑窗技术引入子空间法,提出一种滑窗子空间模态参数识别方法。而后基于滑窗子空间算法识别结果,参考各国高速铁路桥梁相关规范的状态评价指标,提出基于频率值和频率变化率的铁路桥梁运营状态预警方法,实现铁路桥梁的实时状态评估。运用模型试验各工况下的测试数据,对该模型桥的状态进行实时评估和预警分析,并将状态预警结果与模型桥实际损伤情况进行对比分析,验证所提方法的正确性,该方法可直接应用于桥梁结构的运营状态预警分析中。

Study on Early Bridge Structure Operational Warning Based on Sliding Window Technique and Data-driven Stochastic Subspace Identification Algorithm

In order to promote the reliability of the safe operation of bridges,the sliding window technique and data-driven stochastic subspace identification(Data-Driven SSI) algorithm are employed to study the early operational warning of bridge structure based on the shaking table model test of a typical cable-stayed bridge.To achieve the real-time tracking and recognition of modal parameters of high-speed railway bridges in time domain,sliding window technique is introduced into data-driven stochastic subspace identification algorithm,and a sliding-window stochastic subspace identification method is proposed.Subsequently,based on the results of sliding window stochastic subspace identification algorithm and combining the state evaluation indices of high-speed railway bridges in various countries,the early operational warning of high-speed railway bridges based on frequency value and frequency changing rate is proposed,which is expected to realize the real-time state evaluation of railway bridges.By using the testing data of the model test under different testing conditions,the real-time evaluation and early warning and analysis of the state of the model bridge under different testing conditions are carried out.And then the real-time early warning results are compared with the actual cracking and damage conditions respectively to verify the correctness of the method.The method is proved applicable directly in the early operational warning and analysis of high-speed railway bridges.