基于WMRA的隔震结构时变非线性参数识别

Time-Varying Nonlinear Parametric Identification of Isolated Structure Based on Wavelet Multiresolution Anslysis

基于小波多分辨率分析(wavelet multi-resolution analysis,简称WMRA)理论,建立结构时变非线性参数识别方法,研究隔震结构在地震作用下表现出来的时变非线性力学行为。首先,以描述隔震支座的迟滞非线性Boucwen模型为研究对象,引入3阶有效数值差分技术,将运动方程中非线性模型的恢复力增量进行线性化,形成递推观测方程;其次,基于WMRA理论,将观测方程中每一待识别参数采用小波多分辨率展开,将模型参数识别问题转化成多元线性回归模型中小波重构系数的估计问题;最后,分别采用数值模拟、串联隔震体系(serial seismic isolation system,简称SIS)振动台试验测试数据,识别结构及隔震层参数随时间的演化规律,验证该识别方法的正确性和适用性,为隔震结构在施工、运营期各阶段状态评估提供一定参考。

In order to study the time-varying nonlinear mechanical behavior of isolated structures under earthquake action,a novel method of identifying time-varying nonlinear parameters of isolated structures is proposed based on wavelet multi-resolution analysis(WMRA)theory.The classic hysteresis nonlinear Bouc-wen model used to describe isolation bearings is studied.In the motion equation established,the three-order effective numerical difference technique is cited to linearize increment expression of restoring force in nonlinear model and form a recursive observation equation.The time-varying parameters to be identified in observation equation are expressed approximately using WMRA,and their identification can be transformed into time-invariant wavelet reconstruction coefficient estimation in multivariable linear regression model.Numerical simulation and shaking table test data of serial seismic isolation system(SIS)are employed to identify evolution law with time of structural parameters and isolation layer parameters,and correctness and applicability of the method are verified.It can be used to evaluate mechanical states of the isolated structures during both construction period and operation period.