Modal Analysis Identification Under Ambient Excitations Using Stochastic Subspace Method

Due to environmental and geological conditions,irregular behavior during construction,construction materials or equipment that do not meet regulations,and other factors,the structure gradually becomes damaged during use.When structural damage accumulates to a certain extent,it is easy to cause sudden collapse or destruction.In order to ensure structural safety,it is meaningful to establish a structural health monitoring system that can achieve monitoring fast and reliably.Structural damage often leads to modal changes.If we identify structural modalities accurately and compare the identification values with the modalities of healthy structures,the structural damage will be effectively monitored.Most system identification methods are based on system incentives and system responses,but large buildings are difficult to apply incentives.Therefore,it is meaningful to study those system identification methods only based on environmental incentives.The stochastic subspace method is a modal identification method born in the 1990 s.It can extract structural modal parameters only from the structural response.Compared with other recognition methods,the processing idea of the stochastic subspace method is easier to understand,and it is still as simple as the low-order system when dealing with high-order multivariable systems.Therefore,the stochastic subspace method has a good practical value in the health monitoring of large civil structures.This paper firstly introduces the theoretical basis and identification process of the stochastic subspace method.Secondly,it sets up five models of five-degree-of-freedom massstiffness-damping system and frame structure and white noise excitation signal is applied on them.The displacement information of each node is collected in this paper,and the stochastic subspace method based on covariance or data are applied respectively.The stochastic subspace method analyzes the displacement information of this two models,compares the identified frequency and damping information with the calculated theoretical modes.The recognition accuracy of the stochastic subspace method is verified.Then,the ANSYS model of Nanjing South Railway Station is established and white noise excitation is applied on it.Collecting the displacement time history information of some nodes on the model randomly,and the collected displacement information is analyzed by stochastic subspace method to get the frequency of Nanjing South Railway Station.Compared with the frequency calculated in ANSYS,the results show that the random subspace method can accurately identify the low-order frequencies of large civil structures.Finally,in order to study the recognition effect of random subspace method under other noise conditions,this paper establishes a filtered white noise signal model and a non-white noise model,which are loaded on a five-degree-of-freedom structure and frame structure to acquire displacement information.The stochastic subspace method is used to identify the system modality.The results show that the random subspace method can also identify some modal frequencies on ordinary excitation.