At present，the big aviation countries already have mature gust wind tunnel test technology，but which is relatively backward in China，especially the gust wind tunnel tests equipment and technology of full aircraft model are lack. In this paper，a gust generator，a five-degree-of-freedom suspension system and a full elastic aircraft model are developed，and the wind tunnel tests of the whole model are carried out. The test results show that the gust field is stable，and the deviation of the gust velocity between the two ends and the center of the wind tunnel is less than 25%. The support stiffness of the model suspension system is small and the stability is good，which can meet the requirements of the gust wind tunnel test. The simulation results of the non-uniform gust field are close to those of the wind tunnel test，and the error of the moment of the wing root is less than 15%，and the error of wing tip overload is less than 0.2g.

In order to suppress the nonlinear vibration during the motion of a flexible manipulator，a model-free hybrid control strategy of trajectory tracking and vibration suppression based on a novel online observation of disturb forces is proposed. The Lagrange equation and singular perturbation method are employed to model and decouple the dynamics of the manipulator，which are decomposed into a slow subsystem representing rigid motion and a fast subsystem representing flexible vibration. Considering the complexity of modeling and uncertainty of model parameters，PD control method is adopted to realize trajectory tracking，and model-free adaptive control algorithm is proposed to realize nonlinear vibration control of flexible links. To solve the control divergence problem caused by unknown external disturb forces，a modified extended state observer is proposed to online estimate and real-time compensate the disturb force，which can improve the convergence performance of model-free vibration control algorithm effectively. The simulation results show that the proposed method can effectively suppress the vibration of the flexible manipulator in the presence of disturb force，and has good dynamic performance and robustness.

This paper investigates the dynamics and control problems of the long-term deorbiting of an electrodynamic tethered satellite system. The dynamics modeling of the system is carried out based on a dumbbell model assumption. To improve the accuracy of the system model，the orbital dynamics is described using a set of modified equinoctial elements，involving the effects of Lorentz force，atmospheric drag and J₂ perturbation force. Three current control strategies are proposed to regulate the electrodynamic forces for achieving a stable long-term deorbiting process，namely，the constant current input，the directionally variable current input，and the optimal control strategies. In the design of the optimal control strategy，the long-term deorbiting problem is formulated as an inverse problem of dynamics with nonlinear constraints，which is further solved via a nonlinear programming method to obtain the optimal reference trajectories. The deorbiting of the system is then achieved using the modified current control input obtained from a tracking feedback control law. Additionally，an energy-based current switch control strategy is adopted to ensure the stability of system and the efficient utilization of Lorentz force. Case studies of the system with designed physical parameters are conducted to analyze the deorbiting efficiency and to validate the effectiveness of the proposed control strategies.

Stochastic subspace identification （SSI） generates spurious modes in the process of identifying the dynamic characteristics of high-rise structures，which interferes with the automatic tracking of dynamic characteristics. This article has proved that the non-white noise excitation is one of the causes of spurious modes，and further proposed a signal reconstruction method based on multivariate variational mode decomposition （MVMD） for non-white noise excitation，which removes the influence of non-white noise excitation in signals and eliminates spurious modes. A Single-Pass clustering algorithm is proposed to eliminate discrete spurious poles. The above algorithm has been applied to on-site monitoring data of super high-rise structures，achieving long-term automatic identification and tracking of dynamic characteristics.

The key to damage pattern recognition lies in digging and classifying damage features from the response data of civil structures. To this end，a stack auto-encoder network with several auto-encoder hidden layers and a Softmax classification layer is built for analyzing frame structures. A hybrid learning mechanism is adopted to combining unsupervised and supervised learning strategies. Finite element analysis is used to generate the transmissibility function samples corresponding to different scenarios of a frame structure. The transmissibility samples are then divided into training，validation，and test sets. The parameters of the auto-encoder hidden layers，such as the weights and bias，are determined by a pre-training strategy in order to avoid the phenomenon of network over fitting. A fine-tuning step is employed to adjust the pre-trained network parameters，and the network hyper parameters are further adjusted based on the validation set. The measured transmissibility data are input into the network to evaluate the damage of the frame structure. The analysis results show that the proposed method can effectively extract and classify the damage features. Both the single and double damage scenarios at the frame joints were identified with higher accuracy and better anti-noise ability than the traditional shallow neural network.

Establishing human body dynamics model to obtain human natural vibration frequency is a common scientific challenge in various fields such as civil engineering，traffic engineering，aerospace，rehabilitation medicine and so on. The spring-mass-damper （SMD） model is most commonly used in previous studies，which actually is not consistent with the distribution characteristics of human mass and stiffness along the height. In this study，a distributed parameter dynamics model of the human body with a pair of biomechanical forces is proposed，and the analytical solution of human natural frequency is theoretically derived. Therefore，a frequency recognition method based on gait tests is proposed. 247 subjects are organized to conduct gait tests，and their stiffness and natural frequency are identified. The rationality and applicability of the proposed model are verified from multiple perspectives： by fitting the probability distribution of results，comparing the results with other researches，and analyzing the results across different age groups.

As a method for operational modal analysis （OMA），the Bayesian FFT algorithm has garnerd significant attention for its high accuracy and efficiency，as well as its ability of uncertainty quantification. However，different cases of OMA （e.g. well-separated mode，closely-spaced modes，and multi-setup OMA） require different optimization strategy，and it is tedious in computer coding. A new framework is proposed in this paper to unify the above-mentioned cases of OMA，and the implement is simplified as a consequence. Regarding the structural modal response as a latent variable，the single-setup and multi-setup Bayesian OMA is cast as latent variable models，which have been deeply investigated in statistics. An expectation-maximization （EM） algorithm is developed for both single-setup and multi-setup OMA. The introduction of latent variables decouples the parameter optimization in EM，and Louis identity is employed to calculate the Hessian matrix. Two field tests are applied to verify the performance of the proposed approach，with a comparison to the existing algorithm. Consistent results are obtained，and a great advantage in efficiency is observed in the case of closely-spaced modes. The proposed latent variable model unifies the cases of Bayesian OMA，with the advantage of simplified implementation and fast computation. It also paves a way for a further improvement of Bayesian OMA，e.g. with the approach of variational Bayes or Gibbs sampling.

In the realm of stochastic nonlinear response analysis for large and intricate structures，the Monte Carlo simulation method stands out as a pivotal approach. However，its widespread practicality is hampered by its exorbitant computational costs. To surmount this challenge，researchers have endeavored to develop the active learning-based Gaussian process surrogate model algorithm. Despite its promise in reducing computational expenses，the optimization strategy associated with active learning necessitates further refinement to meet the exacting demands of engineering applications. For this purpose，we introduce a search function endowed with ‘intelligent’ attention capabilities. This function is meticulously crafted to concentrate on exceedingly high-risk one-sided tail events in engineering scenarios. By incorporating this search function，we have engineered an algorithm that surpasses existing methodologies. Our algorithm finds successful application in the analysis of complex adhesive anchoring structures within subway tunnel rings and linings. Compared to conventional methodologies，our algorithm exhibits a remarkable 30% reduction in the estimation error of single-tailed probabilities. This advancement facilitates a more precise estimation of the one-tailed probability distribution governing the stochastic response of complex structures. Consequently，it enhances the precision of assessing the occurrence probability of extreme events. These findings yield invaluable insights for decision-making processes in pertinent engineering domains and insurance sectors.

There are two kinds of stochastic seismic ground motion simulation methods： frequency-domain methods and time-domain methods. Based on the time-domain model of single filtered white noise，this paper proposes the time-domain representation for simulating stationary and non-stationary seismic ground motion processes. In essence，the time-domain representation can be regarded as linear superposition of deterministic functions modulated by a series of standard orthogonal random variables，and the set of orthogonal random variables is defined as the form of random orthogonal functions to achieve efficient dimension-reduction. Therefore，by introducing three kinds of random orthogonal functions，i.e.，Legendre orthogonal polynomial of non-Gaussian type，Hartley orthogonal basis and Hartley orthogonal elementary of Gaussian type，the acceleration process of seismic ground motion can be accurately represented in the time-domain model with only one elementary random variable. Numerical examples of seismic stationary ground motion process show the effectiveness of the proposed method，which is superior to the Monte Carlo method. The analysis of fully nonstationary seismic ground motion shows the engineering applicability of the proposed method.

The traditional additional acoustic black hole （ABH） structure is mainly designed for vibration suppression of plate structures，but it is difficult to be applied to pipe structures that are widely available in engineering. In order to solve the vibration suppression problem of pipe structures，a new additional ABH device，‘Circular Spiral ABH Damper （CSABH）’ is proposed to be applied to pipe structures. By designing the ABH area in the form of a spiral，the modal density of the damper is increased，and a better coupling with the main structure is achieved. The results show that the CSABH has good wave aggregation characteristics and can achieve a vibration suppression of 20~5000 Hz for the pipe. Besides，when the pipe constraints and temperature conditions are changed，the CSABH with the same parameters can still play a good wide frequency damping effect，showing the robustness of the damping. The wide frequency，high efficiency and high robustness of CSABH in the vibration control of pipeline structures are verified experimentally.