In order to analyze the reliablility of hypersonic flight vehicles，the longitudinal model of the vehicle is simplified as a cantilever beam structure，and a limit state function is formulated. To address the uncertainty of variable parameters within the limit state function，a hybrid reliability analysis method based on a two-stage Kriging model is proposed. For the first stage，initial sample points are selected to construct an initial Kriging model centered on potential failure points meeting specified accuracy requirements，ensuring the model satisfies this accuracy criterion. For the second stage，a hybrid reliability analysis of the flight vehicle is performed using the initial Kriging model and the first-order reliability method. The Kriging model is adaptively updated by incorporating learning functions，thereby enhancing the efficiency and accuracy of reliability calculations. Comparing the results with existing methods under different parameters of ultimate strength，cantilever beam height，and width，it is demonstrated that the proposed method can meet the requirements for real-time and accurate reliability analysis of the hypersonic vehicle.

Aiming at the chatter problem in the process of milling screw rotors，a chatter monitoring method based on RelifF algorithm to the least square support vector machine （RF-LSSVM） is proposed. Firstly，the vibration signals in the milling process of the screw rotor are decomposed，and feature extraction and selection are performed using the variational modal （VMD） and the RelifF algorithm. Secondly，the penalty factor，kernel parameter，the number of near neighbor samples of RelifF algorithm and the length of dimension reduction feature of LSSVM are iteratively optimized using the enhanced whale optimization algorithm （E-WOA）. Finally，a flutter identification model is established by inputting the reduced-dimensional flutter eigenvector matrix and outputting the flutter occurrence state. The experimental results show that the proposed VMD-RF-LSSVM model has a higher recognition accuracy than the unoptimized variational modal decomposition-support vector machine （VMD-SVM） model，reaching 99.5% accuracy. The proposed method can effectively monitor the chatter problem in the screw milling process，provides a thought for the optimization of the screw milling processing.

To investigate the fault mechanism of blade cracks and to analyze the effects of blade crack on the three-dimensional （3D） tip clearance of the rotor system，while comprehensively considering blade radial deformation，flap-wise bending，and chordwise bending，this paper develops a novel dynamic model of the rotor system based on continuum theory. With the blade breathing crack model under the three-dimensional stress state，a 3D tip clearance dynamic response model of rotor system with blade cracks is further established. The accuracy of the dynamic model is validated by comparing it with the finite element model and experiments. On this basis，the effect of blade crack depth and location on the 3D tip clearance in rotor system is further analyzed. The results show that the amplitudes of the high frequency doubling component of the 3D tip clearance increase with crack depth，while both the fundamental frequency and the high frequency doubling component of the 3D tip clearance show a non-monootonic trend as the relative crack location increases. The research results provide theoretical guidance for research on monitoring and diagnosis method of aero-engine blade crack based on 3D tip clearance.

In view of the current situation that the detection indices of maglev train boot rail system both domestically and internationally are relatively limited，and there is a lack of detection data within the 100—140 km/h speed range，a comprehensive detection framework for high-speed maglev train boot rail system is proposed. Firstly，in accordance with relevant standards and specifications of pantograph and conductor rail system testing，a detection method is developed，which includes conductor rail contact force，vibration acceleration，electric shoe current，arc combustion and transverse geometric parameters of the conductor rail. Secondly，a real-time side conductor rail monitoring system that combines video surveillance and data statistical analysis is proposed，and a supporting program for extraction，processing and analysis of original detection data is developed. Finally，a medium- and low-speed maglev line is taken as the test object，and the data measured at different speed levels of the maglev train are analyzed. The results show that the dynamic performance of the pantograph differs during the upward and downward runs of the maglev train； The contact force and vibration degree at the expansion joint are larger than those in the middle section，indicating poorer dynamic performance of the conductor rail. Toe vibration mainly comes from vertical vibration. Relevant studies reveal the characteristics and issues of the maglev train boot rail system under different working conditions，providing theoretical support and practical basis for the further optimization of the maglev train boot rail system.

Quasi-zero stiffness （QZS） vibration isolation，by introducing stiffness nonlinearity，effectively addresses the inherent contradiction between load-bearing capacity and isolation bandwidth in conventional linear isolators. As a result，it exhibits superior low-frequency isolation performance. The core challenge in realizing QZS isolation lies in designing mechanical structures whose force-displacement curves simultaneously demonstrate high static stiffness and low dynamic stiffness. Focusing on QZS isolation design methodologies，this paper first outlines the fundamental principles of QZS isolation and categorizes the traditional approaches according to the means of stiffness nonlinearization into four groups： geometric motion nonlinearity，geometric deformation nonlinearity，magnetic nonlinearity，and stress-strain nonlinearity. Subsequently，it introduces emerging design strategies based on nonlinear positive-stiffness structures，including hardening and softening types，and compares them with conventional approaches，with particular attention to their differences in static and dynamic behavior. Finally，the paper summarizes and discusses future directions from the perspectives of negative-stiffness structure design，QZS characteristic tuning，and potential applications，aiming to provide a comprehensive overview of the latest research progress and to offer insights into future development trends of QZS isolation systems.

To address the inadequacy of existing models for predicting the flow resistivity of kapok felt，the airflow resistivity of kapok felt with different bulk densities is tested. Initially，the airflow resistivity of kapok felt is calculated using empirical and theoretical models commonly used in fibrous materials. Subsequently，a new empirical model suitable for predicting the flow resistivity of kapok felt is developed by fitting the experimental data. Finally，considering the cross-sectional geometries and arrangement of kapok fibers at different bulk densities，theoretical models for the flow resistivity of kapok felt are derived by determining the average velocity and frictional force within micro-units，for both circular and flattened fiber cross-sections，based on the Tarnow model and the Hagen-Poiseuille flow assumption. The models are further modified using a flattening ratio parameter. Results demonstrate that compared to the measured values，the prediction accuracy of existing models for kapok felt flow resistivity is low. The modified model is applicable to transitional states between the two cross-sectional geometries. Within the bulk density range of 20 to 180 kg/m³，the modified model exhibits high prediction accuracy.