To study the influence of subway wheel polygons on low-frequency vibration of the car body，the polygon wear of wheels of a subway line is investigated，on the basis of grasping the distribution characteristics of wheel polygons of subway lines. A vertical dynamic model of elastic car body considering wheel polygons is established，by the time-domain integral solution method. The relationship between wheel polygon excitation frequency and low frequency vibration of vehicle body is studied. By comparing the low-frequency vibration of wheel polygons of different orders，the effect of changes in the operating speed of metro vehicles under service conditions and changes in the radius of wheel wear on the polygonal action of the wheel body is discussed separately. It is shown that a wheel polygon of order 1—3 at common operating speeds generate a low-frequency excitation frequency of 0—20 Hz，and when the excitation frequency is close to the first-order droop frequency of 10.2 Hz resonance will occur. At the beginning of service，the influence of the second order wheel polygon becomes severe on the low frequency vibration of the car body. With the wear of the wheel radius during service，the influence of the first three order wheel polygon changes on the vibration law of the car body. This paper provides a good reference value for service subway operation and wheel maintenance.

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.

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.

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.

When early failures occur in planetary gearboxes，the weak fault features are difficult to extract and identify due to the interference of background noise in industrial environments and the attenuation of fault impacts in complex transmission paths. To address this issue，a sparse-guided improved empirical wavelet transform （IEWT） is proposed combined with multipoint optimal minimum entropy deconvolution adjusted （MOMEDA） method for weak fault feature extraction. Firstly，a new fault composite index （FCI） is introduced，and the original signal is adaptively decomposed into a set of IEWT components based on the amplitude envelope of the signal spectrum. Secondly，the sensitive components，selected through the sparse-guided method，are used as the sparse representation of the original weak fault signal. Finally，the MOMEDA technique is applied to the sensitive component signals to reduce signal noise and extract the weak fault feature frequencies for identification. The effectiveness of the proposed method is validated through simulations and experiments，successfully extracting and identifying the weak fault features of planetary gearboxes. This demonstrates that the method has good diagnostic performance for noisy，non-stationary，and non-linear fault signals in planetary gearboxes，providing a new approach for the diagnosis and identification of weak faults in engineering practice.

Wind tunnel pressure tests are conducted on a high-speed railway station roof to study the non-Gaussian characteristics and extreme wind pressure distribution on the long-span roof surface. First，the surface wind pressure is classified into Gaussian and non-Gaussian distributions. Then，the fitting effects of three different single probability distributions （Gumbel，Lognormal，and Weibull） and their corresponding combined distributions （double Gumbel，double Lognormal，and double Weibull） on the wind pressure time history of the roof surface are compared. The extreme wind pressures obtained from the combined probability distributions are compared with the estimates from the modified Hermite method. Finally，the extreme wind pressure distribution on the roof under all wind directions is presented. The results show that the combined probability distributions provide a better fit to the wind pressure time history than the single distributions. The extreme value estimates from each combined distribution at the same guarantee rate are more accurate than those from the single distribution. The combined distributions generally yield better estimates at the 99.90% guarantee rate compared to the modified Hermite method. The extreme wind pressure varies significantly with the wind direction，and under all wind directions，the minimum pressure coefficient reaches its lowest value at the middle of the roof edge side，reaching -5.9.

Aiming at the multi-source noise of inertial navigation in the attitude estimation of coal mine bolting jumbo，a noise reduction method of inertial navigation heterogeneous signal is proposed based on noise sensitive prior and improved variational mode decomposition（VMD），which avoids the over-decomposition and under-decomposition problems caused by the parameter fixation. Firstly，the noise sensitivity difference of the heterogeneous signals （acceleration and angular velocity） of coal mine bolting jumbo is investigated by using the variation of the signal characteristics in the time and frequency domains. Secondly，according to the noise-sensitive characteristics，the dual decomposition layer and energy fluctuation model are constructed，so that the decomposition parameters have the ability of adaptive optimization and the synchronous optimal decomposition of the inertial-guide heterogeneous signals is realized. Based on the Pearson correlation coefficient （PCC），the modal component screening parameter，correlation coefficient P，is designed to consider the noise sensitivity difference，to achieve screening practical modal components and simultaneous noise reduction of heterogeneous signals. Finally，the proposed method is compared with the noise reduction results of complementary ensemble empirical mode decomposition （CEEMD） and improved complete ensemble empirical mode decomposition with adaptive noise （ICEEMDAN）. The results show that the method proposed in this paper considers the noise sensitivity differences of heterogeneous signals，thereby improving the signal-to-noise ratio of inertial measurement and enhancing the attitude initialization accuracy of bolting jumbo. The pitch error is reduced by 81.818 %，and the yaw error is reduced by 87.958 %，which lays a good foundation for accurate roadway support.

To address the issue of metallic objects around a balise affecting its electromagnetic transmission performance，a balise antenna model is established in electromagnetic simulation software. The simulation experiments are carried out by varying three parameters：the metal surface area，the vertical distance between the metal surface and the balise，and the metal surface thickness. The uplink signal amplitude curves of the balise under different parameters are obtained，and the transmission performance metrics are calculated to analyze the impact. Results show that： a larger metal surface area leads to lower performance metrics，such as a reduced number of safety message frames received by the balise transmission module （BTM），with a more significant degradation and greater interference from the sidelobe region. When the metal area is greater than 320 mm×320 mm，the uplink field strength consistency requirement can no longer be met. A greater absolute distance between the metal surface and the balise results in less interference from the sidelobe region，and the distance must be greater than 123 mm to satisfy the field strength consistency requirement. Increased metal surface thickness causes greater interference from the sidelobe region，and the thickness should not exceed 1 mm.

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.

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.