Parameter Identification Of Local Connection Nonlinerity Based On Instantaneous Characteristics Of Structural Response

For nonlinear structures,model building and related assumptions based on linear theory may no longer be applicable.Therefore,it is of great practical significance to study the model establishment of nonlinear structure and the parameter identification of nonlinear model.Space truss structure is widely used in civil engineering and aerospace fields due to its advantages of light weight,high stiffness and rapid construction.However,due to the influence of factors such as clearance and friction collision in the connecting joints of the structure,the truss structure as a whole shows strong nonlinear characteristics.For the assembled space truss structure,dynamic modeling is needed to understand the uncontrollable influence of its structural characteristics and working state on the system.It is of great significance to study the nonlinear characteristics of space truss structure and identify the nonlinear parameters of space truss structure.The vibration response of nonlinear structures generally contains multiple principal components with time-varying frequency and time-varying amplitude.The principal component of each vibration signal is obtained by VMD decomposition technique,and then the analytical signal of each principal component can be constructed by Hilbert transform.The instantaneous characteristics extracted by the analytical signal maintain the instantaneous amplitude and instantaneous frequency information of the component.As an information processing system that can simulate the human brain nervous system to process,memorize and process information,the neural network has excellent generalization ability and can well fit the relationship between the parameters of the nonlinear structural model and the instantaneous parameters of the vibration response.At present,neural network,as a calculation method of colleges and universities,has been applied in the parameter identification of nonlinear structural model,which greatly improves the accuracy and efficiency of calculation.The main contents of this paper include the following aspects :Through an example of Duffing nonlinear free vibration system,the principal component of the signal is decomposed by VMD,the instantaneous frequency and instantaneous amplitude of the principal component of the vibration response of the system are obtained by Hilbert transform,and then the parameters of the vibration are obtained by least square fitting.The parameters obtained by fitting are compared with the mathematical theoretical solution of the system.It is proved that VMD decomposition and Hilbert transform can effectively extract the instantaneous characteristics of the vibration system.The assembled space truss structure is taken as the research object.The finite element model of the assembled space truss structure is established by ANSYS finite element software.Considering the nonlinearity caused by contact collision and friction in the joint,the nonlinear spring-damper element is used to simulate the truss joint.The joint is simplified as a six-degreeof-freedom spring-damper element.Taking the bilinear hysteretic model and the piecewise stiffness model as examples,the nonlinear dynamic response of the assembled space truss structure under simple harmonic excitation and sinusoidal sweep excitation is calculated respectively.The influence of different nonlinear model parameters on the dynamic response of the truss structure is analyzed.Variational mode decomposition(VMD)and radial basis function(RBF)neural network are combined to identify the nonlinear parameters of the assembled space truss structure model.The vibration signal of the truss structure is decomposed by VMD decomposition method to obtain the principal component of the vibration signal,and then the instantaneous characteristics of the principal component of the signal are obtained by Hilbert Transform(HT).The complex nonlinear mapping relationship between the nonlinear joint parameters and the instantaneous characteristics of the principal component is fitted by the powerful fitting and generalization ability of the radial basis neural network to achieve the purpose of identifying the nonlinear joint parameters.The effectiveness of the method is verified by numerical simulation of space truss structure.