基于扩展卡尔曼估计算法的地震模拟振动台模型识别  

Identification of shaking table model for seismic simulation based on an extended Kalman estimator

作　　者：赵博宇[1] 丁勇[1] 吴斌[1] 

机构地区：[1]哈尔滨工业大学土木工程学院,哈尔滨150090

出　　处：《振动与冲击》2014年第12期145-150,共6页Journal of Vibration and Shock

基　　金：国家自然科学基金(51161120360);哈尔滨工业大学创新基金(01319406)

摘　　要：由于地震模拟振动台的长期使用和多次维护,其实际工作中的模型往往难以建立,而地震模拟振动台模型的修正对于估计实验结构底部剪力有重要作用。为此利用扩展卡尔曼估计算法识别了哈尔滨工业大学地震模拟振动台模型。首先,对一单自由度结构进行仿真,通过测量位移响应,将结构参数准确的识别。随后,对抗震实验室振动台进行了3次空台加载,通过测量振动台台面位移,利用扩展的卡尔曼估计算法,识别地震模拟振动台的模型,更新了其质量,阻尼和刚度。而后分别进行两组不同载荷形成的空台加载实验,由识别参数重构计算得到的台面响应与实测台面响应吻合良好,由此表明振动台模型参数估计准确。通过此次研究既可以有效的认识不同构建形式的振动台模型,同样也可用于确定某些难以测得参数的复杂结构的模型。The dymanic model of a long used shaking table for seismic simulation is difficult to establish.An exact shaking table model for seismic simulation is significant for estimating a tested structure＇s base shear forces.The shaking table model for seismic simulation at Harbin Institute of Technology was successfully identified based on an extended Kalman estimator here.Firstly,the displacement of a SDOF structure subjected to earthquake was measured.Using the proposed method and the measured displacement,its parameters were identified correctly.Then,the shaking table of our university was loaded for three times,the proposed method and the measured table surface displacements were used to identify the shaking table dymanic model,its mass,damping and stiffness were updated.To evaluate the accuracy of identified parameters,the shaking table was loades with other two types.The table surface responses calculated with the updated parameters of the model agrees well with that measured with sensors.The result showed that the parametric estimation of the shaking table dynamic model is correct with the proposed method.Furthermore,the proposed method could be used to identify shaking table models with different forms and the models of complex structures.

关 键 词：模型更新 结构参数识别 地震模拟振动台 扩展卡尔曼滤波估计 

分 类 号：TU311.4[建筑科学—结构工程]