A Study On Transfer Entropy Theory Applied In Nonliear Dynamic Identification Of Engineering Structures

For any actual engineering structure,due to various environmental loads and service loads,the problem of damage accumulation is faced from the time the structure is in service.With the long-term effects of dynamic loads and static loads and long-term environmental erosion,the damage of the structure will become increasingly serious.In order to ensure the safety and durability of the structure during the service,it is very important to identify the damage of the structure.The research status of damage identification methods based on structural vibration was reviewed comprehensively in this paper,and the transfer entropy theory was used to carry out nonlinear dynamic identification of engineering structures.The main research work of this paper can be summarized as follows:1.The finite element software ANSYS was used to simulate vertical cantilever beams with cracks at different locations.The surrogate data algorithm and the improved surrogate data algorithm were used to reconstruct the time-history data of the vibration acceleration signal obtained by the numerical solution.Judging from the difference between the autocorrelation function of the two kinds of surrogate data and the autocorrelation function of the original data,the data generated by the improved surrogate data algorithm was more in line with the requirements of the subsequent calculation of the transfer entropy.The damage identification of the numerical model of cracked beams was studied by using the transfer entropy and the improved surrogate data algorithm.The results showed that for the group of points close to the crack location,the degree of deviation between the transfer entropy calculated from the original data and the transfer entropy calculated by the surrogate data increases with the increase of the crack size.However,there was no such variation in the transfer entropy of the group of points away from the crack location.This rule showed that the transfer entropy combined with the improved surrogate data algorithm had the ability to locate cracks(damage).The concept of relative reference state was proposed.On this basis,the damage index of the previous scholars was improved.It was found that the damage index value of the group of points near the crack on the cantilever beam is significantly positively correlated with the crack length.The crack length can be identified by the damage index,and the validity of the proposed index is verified.2.The Galerkin method combined with the fourth-order Runge-Kutta method was used to solve the governing equations of motion of a homogeneous Euler-Bernoulli simply supported beam that is rested on a nonlinear viscoelastic foundation and subjected to a time-dependent load.In order to simulate different nonlinearity levels of system,different ratios between nonlinear stiffness coefficient and linear stiffness coefficient of the foundation,as well as different Young’s modulus of the beam were used.Using the transfer entropy combined with the improved surrogate data algorithm,the transfer entropy was calculated respectively from the original data and the surrogate data of the transverse displacement of the beam under different nonlinearity conditions.The results showed that the values of the proposed index were positively correlated with the degree of nonlinearity.It was verified that the transfer entropy combined with the improved surrogate data algorithm can effectively identify the nonlinearity state of the foundation-structure system.By adding different intensity noise to the lateral displacement signal,the results verified that the nonlinearity state identification method based on the transfer entropy combined with the improved surrogate data algorithm had strong anti-noise ability3.Three different methods were used to analyze the measured response data of the wind turbine towers with flange damage.The results of the FFT spectral analysis method showed that this method is very insensitive and it was not possible to identify the flange damage by the natural frequency alone.The results of the identification method based on self-power spectral entropy showed that the method could identify flange damage but could not locate damage.Using the damage index proposed in this paper,the measured data of the wind turbine tower with the same model,the same construction conditions and good operating conditions was used as relative reference data.The transfer entropy was calculated from the measured vibration data at three sampling frequencies.The results showed that the transfer entropy combined with improved surrogate data algorithm had good damage identification effect.It was proved that the proposed method and the proposed index were not only applicable to the nonlinear dynamic identification of numerical models and semi-analytical numerical models,but also to the nonlinear damage identification of practical engineering structures.4.In order to produce obvious normal section failure and oblique section failure,two kinds of reinforced concrete simply supported beams were used.For each of the two failure modes,one complete beam and three broken beams were selected to test the crack identification effect of the method which was based on the transfer entropy and the improved surrogate data algorithm.By defining the appropriate discreteness index,the identification effect of the transfer entropy on microscopic cracks,macroscopic cracks and through cracks generated during the failure process was studied,and the variation law of the discreteness index in the process of two failure modes was analysed.The results showed that in addition to micro-cracks,the transfer entropy combined with the improved surrogate data algorithm could effectively identify and locate the other two types of cracks in reinforced concrete beams.The robustness of the method based on the transfer entropy combined with the improved surrogate data algorithm for damage identification of reinforced concrete structures was verified from the perspective of sampling frequency and excitation method.On the basis of the finite element numerical simulation,semi-analytical numerical method,engineering test and experimental research,the transfer entropy combined with improved surrogate data algorithm was used to study the dynamic identification of different nonlinear forms in engineering structures.The applicability and effectiveness of the proposed method were verified,and this method would provide a new and effective way for nonlinear dynamic identification of practical engineering structures.