基于AR模型和因子分析的结构损伤预警研究

发布时间：2018-07-15 20:54

【摘要】：结构损伤会引起动力响应特征参数的变化,而环境因素(温度、湿度)的改变同样也会导致动力响应特征参数发生变化,有时甚至会淹没损伤引起的动力响应特征参数的变化,从而导致基于振动的结构损伤识别方法失效。为消除环境因素(温度)的影响,利用时间序列分析中的AR模型和多元统计分析中的因子分析并结合标准差控制图进行结构的损伤预警研究。首先利用AR模型对结构损伤前后的加速度响应数据进行动态建模,并提取模型系数,其次利用因子分析去除温度对AR模型系数的影响,最后利用标准差控制图进行损伤预警。数值模拟结果表明,在温度变化的条件下,该方法成功地对四层钢框架结构进行了损伤预警。
[Abstract]:The structural damage will cause the change of the dynamic response characteristic parameters, and the change of the environmental factors (temperature, humidity) will also lead to the change of the dynamic response characteristic parameters, and sometimes even submerge the changes of the dynamic response characteristic parameters caused by the damage. As a result, the method of structural damage identification based on vibration is invalid. In order to eliminate the influence of environmental factors (temperature), the AR model in time series analysis and factor analysis in multivariate statistical analysis were used to study the structural damage early warning combined with the control chart of standard deviation. Firstly, the acceleration response data before and after structural damage are modeled dynamically by AR model, and the model coefficients are extracted. Secondly, the influence of temperature on AR model coefficient is removed by factor analysis. Finally, damage warning is carried out by using standard deviation control chart. The numerical simulation results show that the method can successfully predict the damage of four-story steel frame structure under the condition of temperature change.
【作者单位】：青岛理工大学土木工程学院;
【基金】：国家自然科学基金(51179082) 山东省自然科学基金(ZR2010EL035) 山东省高等学校科技计划项目(J12LG09) 蓝色经济区工程建设与安全山东省协同创新中心
【分类号】：TU312.3

【参考文献】