In response to unclear mechanisms in compressor fault diagnosis research，this article aims to reveal the variation law of component characteristic changes caused by blade wear，numerical simulation analysis method is used for studying the changes in component characteristics of blades under different degrees of wear. The results show that the degree of compressor performance degradation caused by blade surface wear is more significant than that caused by blade tip wear. With the increased speed，the degree of performance degradation becomes greater. The efficiency decay value caused by the increase of blade tip clearance shows a pattern of first decreasing and then increasing with the increase of pressure ratio. Near the working line of the compressor，the efficiency decay value is relatively stable with the change of pressure ratio，while the flow decay value increases with the increase of pressure ratio. After the increase of blade surface roughness，the decay values of efficiency and flow rate both increase with the increase of pressure ratio. Moreover，there is a decreasing trend in the decay values near stall pressure ratio. Near the working line，varying degrees of blade wear have a greater impact on flow rate than on efficiency. The results can provide a theoretical reference for the research of fault diagnosis methods for aviation engine gas paths.

To study the wind effect characteristics of super high-rise buildings in different incoming flows under the fluid-structure interaction effect，a full-scale numerical wind tunnel simulation of Pingan Financial Building in Shenzhen is carried out using the detached eddy simulation （DES），established the aeroelastic model. A new turbulent fluctuating flow field generation method named the discretizing and synthesizing random flow generator （DSRFG） is used to simulate the turbulent flows of the atmospheric boundary layer and the uniform flows. The wind pressure and wind-induced response results of the model are obtained. The calculated results are compared with the corresponding data from wind tunnel tests and field measurements to verify the accuracy of the numerical simulation. The analysis shows that the wind pressure of the building under turbulent incoming flow obtained from the DES is consistent with the distribution trend of wind tunnel test and field measurement results. The distribution of the mean wind pressure coefficient of the building under both conditions is similar，and the fluctuating wind pressure coefficient on the windward side under turbulent incoming flow is larger than that under uniform incoming flow. In addition，the cross-wind acceleration response is larger under turbulent incoming flow than that under uniform incoming flow，and the acceleration response power spectrum shows three peaks and the displacement response power spectrum shows two peaks under uniform incoming flow，while the turbulent incoming flow only shows a single peak. The acceleration response power spectrum shows three peaks and displacement response power spectrum shows double peaks under uniform incoming flow with only single peak under turbulent incoming flow. In the flow field，the wind velocity is more uniform in the uniform incoming flow than in the turbulent incoming flow，the wake vortex is flatter and narrower，and the overall vorticity magnitude is smaller.

In order to monitor the operation state of each wheel motor in distributed drive electric vehicle and ensure the safety of the vehicle，a fault diagnosis method of in-wheel motor based on improved multi-class support vector data description （MCSVDD） is proposed. The method incorporates two major improvements. First，a classification judgment rule based on the minimum distance to the cluster center within the class is proposed using the affinity propagation （AP） clustering algorithm to enhance MCSVDD. Second，a Weibull kernel function is constructed from the Weibull distribution to optimize data description model. Meanwhile，a dimensionality reduction method based on minimum-distance propagation discriminant projection （MPDP） is proposed for the multi-dimensional feature set of in-wheel motor operating state，which improves the separability of in-wheel motor fault states under different working conditions. Finally，in-wheel motors with typical bearing faults are customized respectively to collect vibration signals under 7 rotating speeds for verifying the effectiveness of the proposed method. The results show that the reduced dimension data's separability of observed samples of in-wheel motor operating state based on MPDP is better than that of linear discriminant analysis （LDA），minimum-distance discriminant projection （MDP） and locality preserving projection （LPP），and the recognition accuracy of MCSVDD's state recognition system based on Weibull kernel function is higher than that of polynomial and Gaussian kernel function.

Nonlinear structures exhibit multiple responses under steady-state excitation，making it challenging to directly obtain their dynamic characteristics using traditional vibration test methods. To address this issue，a constant-force dynamic characteristics testing method based on acceleration-response-controlled step-sine frequency sweep test technology is proposed. First，the acceleration response at the excitation point is selected as the control signal，and a step-sine frequency sweep experiment is performed by maintaining a constant response amplitude. Secondly，the resulting simple harmonic excitation spectrum and acceleration response spectrum are obtained through experiments under various acceleration amplitude conditions，and a smooth simple harmonic force surface is constructed using linear interpolation. Then，by extracting the contour lines corresponding to constant force amplitude from this surface，the frequency response curve of the nonlinear structure under constant-force conditions is derived. Finally，the dynamic characteristics of a typical bolted nonlinear structure are investigated. The results show that the proposed approach can accurately capture the frequency response characteristics of bolted nonlinear structures，revealing the pronounced nonlinear dependence on force amplitude. Furthermore，bolt preload and structural reassembly are found to significantly influence the dynamic characteristics of the connection structure.

The friction block of high-speed train brake pads exist in two configurations： holed and non-holed. To investigate the influence of the holed structure on the braking performance，drag braking tests using both types of friction blocks are carried out on a self-developed scaled brake dynamometer for high-speed trains. In addition，finite element simulations are carried out to analyze vibration noise and interface thermal distribution during braking. Experimental and numerical results indicate that the non-holed friction blocks produces continuous self-excited vibration and excites high-intensity squeal noise，whereas the perforated block effectively suppress system instability and reduce squeal noise to some extent. Moreover，the holed structure improves the interfacial thermal distribution，leading to more uniform temperatures on both the friction block surface and the matching brake disc compared to the non-holed blocks.

For tip clearance measurement，a multi-section tip clearance measurement and verification simulation rotor test bench is established，and multi-channel tip clearance online measurement software is developed. First，tip clearance is measured using a customized dual-fiber sensor based on laser triangulation. Second，a high-precision displacement platform is utilized to change the tip clearance value，and accurate tip clearance values are measured at different rotational speeds. Last，to address the trigger threshold voltage issue of the acquisition card for measuring tip clearance with dual-fiber sensors，a multi-channel integrated voltage comparator has been designed，which eliminates the time difference on the order of 10^-5 s caused by simultaneous measurement of the two channels of the dual-fiber sensor. The results show that，based on the tip timing principle，the tip clearance can be calculated by measuring only the arrival time of the tip. The mean error of each set of clearances is within 5%，and the standard deviation of the tip clearance measurements in each group is less than 22 µm，demonstrating the accuracy and repeatability of the tip clearance measurement system.

Aiming to address the operation stability affected by the parameter time-variation of the ultrasonic motor and the harmonic effect of the drive voltage，this paper focuses on the design and optimization of LLCC resonant network topology as well as total harmonic distortion （THD） of the output voltage for driving linear ultrasonic motors （LUMs）. Such method can effectively overcome the issue on the variations of the driving voltages caused by the parameter time-variation to improve their operation stability. Firstly，the calculation method for the LLCC matching parameters is derived by using a contact-based equivalent circuit of LUMs considering the stator/mover contact boundary conditions，and a compensation capacitor is added to improve the elasticity margin and the stability of impedance matching. Furthermore，the filtering characteristics of the LLCC resonant network near the resonant frequency is discussed in depth，and the mathematical relationships are derived between the THD of the output voltage and the parameters of the LLCC resonant network are derived. Furthermore，the influence of the parasitic parameters of the transformer on the LLCC resonant network is also analyzed. On this basis，the design optimization methodology for the LLCC resonant network is proposed while acting the THD of the sinusoidal output voltage as the main target. Finally，a LLCC resonant driver is designed for a V-shape LUM，and the corresponding experiments are conducted. The results indicated that the gain and the THD of the output voltage as well as the peak amplitude of the series capacitor voltage itself are determined by the ratio of the series capacitor and the parallel capacitor in the resonant network. The series inductor is the dominating factor for the soft switching characteristics. The THD of the output voltage is controlled below 3%，which is improved more than 70% compared to the unoptimized LLCC resonant driver.

To investigate the data-driven evolution of rail corrugation in subways，a spatio-temporally dense measurement method is proposed for rapid detection. First，high-precision sensors are used to measure car-body vibration，collecting triaxial acceleration data. Second，vehicle speed and mileage position are estimated by fusing triaxial vibration acceleration from different car bodies. Then，the vertical vibration acceleration of the car body is decomposed via wavelet packet analysis，and a vibration energy ratio index is defined as the ratio of the energy in the characteristic frequency band excited by rail corrugation wavelength to the total vibration energy. The vibration energy ratio threshold is set to automatically identify rail corrugation and output mileage information. Finally，the corrugation wavelength is derived from the ratio of vehicle speed to characteristic frequency，and the relationship between vibration energy ratio and corrugation amplitude is analyzed. Results show that using a vibration energy ratio threshold of 0.2 yields corrugation mileage distribution consistent with that from noise-based identification，and the calculated corrugation wavelength matches the measured value of 175 mm from a corrugation analyzer. Statistical clustering reveals that the relationship between rail corrugation amplitude and vibration energy ratio is not purely linear. Line-wide rapid detection shows that non-corrugated sections account for 76.56% of the track，while corrugation sections account for 23.44%. Among the corrugated sections，those with wavelengths below 60 mm dominate （77.88%），whereas those above 60 mm account for 22.12%.

The theoretical framework of Hertzian contact model for soft sphere collisions is well-established； however limited research has been done on the collision between rigid particles and elastoplastic materials. In this study，a novel damping form is introduced by incorporating the elastoplastic half-space contact constitutive relationship，and a viscoelastoplastic collision model between hard spherical particles and polyurethane surfaces is established. The nonlinear dynamic equations for particle and elastoplastic surface collision are derived. Furthermore，by conducting experiments to measure the coefficient of restitution for coal pellets colliding with polyurethane，the Meyer's index of the polyurethane material and the damping coefficient in the dynamic equations are determined. Additionally，the correctness of the damping model proposed in this study is validated by considering different damping forms in the equations. The changes in contact force at different stages are analyzed under various damping coefficients. The variations of displacement，velocity，and contact force during the collision process between particles and elastoplastic polyurethane are investigated. The results reveal that with an increase in the initial collision velocity，the irreversible plastic deformation of polyurethane increases from 1.636×10^-4 m to 5.657×10^-4 m，the coefficient of restitution decreases from 0.583 2 to 0.501 2，the collision time decreases from 6.963×10^-4 s to 4.737×10^-4 s，and the proportion of plastic compression stage decreases from 59.81% to 59.04%.The model established in this paper can be used in scenarios where particles collide with softer planes，such as the separation of metal ores and particle dampers. This paper provides theoretical support for the transportation，collision，impact，and separation of particle systems.

In response to the challenge of quantitatively diagnosing corrosion damage thickness within pipelines，a quantitative imaging method for pipeline corrosion damage using ultrasonic guided waves is proposed. Firstly，based on the frequency domain finite difference method，a numerical model for multi-path helical propagation of guided waves in pipes is established，enabling rapid calculation of guided wave reception signals when thickness map is known. Secondly，by calculating the received signals in the presence of randomly distributed damage，a database comprising 3 500 samples of damage signals is generated through iteratively running the numerical model. Subsequently，a one-dimensional convolutional neural network imaging model is constructed. The model is trained using the generated database to establish a mapping relationship between thickness maps and reception signals，and inputting the reception signals into the imaging model yields corresponding thickness maps. Finally，the feasibility of the proposed method is experimentally validated. The mean square error between experimental imaging results and actual values is 8.6048×10^-4，the correlation coefficient is 0.711 6，and the imaging model runtime is 0.538 5 seconds. The results indicate that the proposed method can achieve quantitative imaging of corrosion damage thickness within pipelines with high imaging efficiency.