To evaluate the performance of various equivalent stress intensity factor models in predicting mixed-mode fatigue crack growth and to address the challenge of parameter estimation under limited sample conditions. A crack growth parameter estimation method based on the Bootstrap method resampling technique was proposed firstly. Mode Ⅰ fatigue crack growth tests were conducted on CT specimens to obtain the material parameters, and the proposed method was employed to expand the sample set and mitigate the issue of data scarcity. Then, using the statistically augmented material parameters,mixed-mode Ⅰ+Ⅱ fatigue crack growth experiments were performed on 6005A-T6 aluminum alloy CTS specimens under loading angles of 0°, 30°, 45° and 60°, employing a Richard-type loading fixture, to validate the accuracy of various equivalent stress intensity factor models. The results indicate that the Irwin model achieved the highest goodness-of-fit, with a value of 0.942 1, demonstrating the best crack growth prediction performance. Increasing the loading angle was found to reduce the initial crack growth rate, highlighting the need for angle-specific experiments to obtain appropriate Paris law parameters. This study confirms the applicability of multiple ΔK_eq models and provides theoretical support for fatigue life prediction in mixed-mode crack growth scenarios.

To address the issue of damage caused by low-speed impacts on composite material laminates coated with polyurethane coating, a numerical analysis method based on three-dimensional progressive cumulative damage in composite laminates and a yield damage criterion for polyurethane coatings was proposed. Firstly, a damage numerical model of polyurethane coating-carbon fiber reinforced composite laminates under erosion was established, and a Vumat subroutine was written. Subsequently, referring to the ASTM D7136 test standard, impact tests with various energy levels were conducted on samples coated with 1 mm and 2 mm polyurethane coatings and uncoated samples. Simultaneously, the proposed damage model was employed to study the formation reasons and propagation patterns of primary damages such as fiber damage, matrix damage, and delamination, thereby revealing the mechanism of polyurethane coating in absorbing impact energy. The results indicated that the mechanical response results calculated by the proposed damage model showed a high degree of agreement with the test results, validating the correctness of the proposed model. Additionally, comparative tests demonstrated the enhancement effect of polyurethane coating on the impact damage resistance of carbon fiber composite laminates. The findings of this study can provide a reference for the design of protective coatings for aircraft.

To enhance the computational efficiency of structural lightweight design for complex structures, a structural lightweight design method based on the Kriging surrogate model is proposed. The proposed method incorporates a hybrid addition strategy and a sample deletion strategy considering a distance threshold，aiming to rapidly improve the fitting accuracy of the Kriging surrogate model. This model was then applied to a multi-objective lightweight design model of the truck frame, with the optimization objectives of minimizing frame mass and maximum stress. Subsequently, the multi-objective lightweight model was solved using the non-dominated sorting genetic algorithm-II (NSGA-II). The results demonstrate that the proposed hybrid addition strategy and sample deletion strategy considering the distance threshold effectively enhance the update process of the Kriging surrogate model. The structural lightweight design method based on the Kriging surrogate model exhibits significant advantages in both computational efficiency and lightweight performance.

In view of the problem that the formulary design load based on the design standard can't truly reflect the actual service conditions of the welded frames of subway vehicles, and a large number of online measured data information has not been fully explored in the structural design of the frame, an anti-fatigue design spectrum compilation method based on the stress-time history sample information measured at the weak position of the subway vehicle frame was proposed.The small stress threshold value was determined by clustering ordered samples using the rain-flow counting method to compile measured stress spectrum reflecting structural damage, determining the stress-frequency-damage relation based on the fatigue damage theory, and using the methods of Bayesian parameter estimation and kernel density estimation to obtain stress extrapolation results. Considering the stress concentration caused by structural geometry changes in the cross-section of the welding site, the hot spot stress method and stress linearization method were used to obtain the stress concentration coefficient, and the measured stress spectrum was corrected to achieve the compilation of the design load spectrum. The research results show that the small stress threshold value is 3.18 MPa determined by extrapolation of stress extremes and the relation analysis between normalized stress-frequency-damage, which is 8.19% higher than the stress threshold value determined by traditional methods, the effect of discarding the number of small stress cycles is significant. Considering the load dispersion and stress concentration factor at the weld seam, while ensuring that the structure meets the current service conditions, the relation between the design mileage and equivalent stress is determined, the necessity of compiling the design spectrum is emphasized further. The above research builds the construction method of load spectrum compilation and the equal strength design of structures.

Aiming at the problem of the low fault diagnosis accuracy caused by the lack of fault samples for the rolling bearings of doubly fed wind turbines under normal conditions for a long time, an improved generative adversarial network fault diagnosis method based on expanding high-quality fault samples and using dual feature extraction was proposed. Firstly,a finite number of rolling bearing fault samples were expanded through a Wasserstein type generative adversarial network with maximum mean discrepancy and penalty constraints. Secondly, based on the dual feature extraction model, the time-frequency converted temporal features and local features were extracted separately. Finally, the fault diagnosis of the rolling bearing balance data was completed through a classifier. The standard dataset and test results show that the proposed method improves the fault diagnosis performance while lacking fault samples.

The aircraft’s inlet structure is connected to the engine sleeve using countersunk rivets. During maintenance,fatigue fractures were discovered in some rivets. It suggests that the inadequate perpendicularity of rivet holes during manufacturing causes the rivet misalignment, reducing the load-bearing capacity, and leading to fatigue fractures under aircraft vibrations. The finite element simulation was used to study the effect of inclined rivet holes on the load-bearing capacity,simulation results show that inclined holes cause uneven stress distribution across the rivet section. The higher the tilt angle,the higher the maximum stress and the more uneven stress distribution on the rivet head section. Fatigue tests under axial loads at different tilt angles demonstrated a reduction in the rivet’s fatigue life due to the hole inclination. The study concludes that non-compliance with perpendicularity standards during hole fabrication results in uneven stress distribution, decreasing load-bearing capacity. Therefore, the strict control over rivet hole perpendicularity during the aircraft manufacturing is crucial to ensure structural reliability.

To enhance the energy absorption efficiency of conventional sandwich panels, a biomimetic tree-like fractal core (BTLFC) inspired by the dendritic fractal structure of the royal lotus leaf vein was designed. Firstly, quasi-static compression tests revealed that the 2-order BTLFC exhibited a specific energy absorption 5.69% higher and an average load 4.46% greater than traditional honeycomb cores. Secondly, a finite element numerical model of the BTLFC was established;combined with quasi-static compression test data, the finite element model error was within 2.2%, demonstrating high accuracy of the model. Finally, Latin hypercube test design, Kriging surrogate model, and the non-dominated sorting genetic algorithm-II (NSGA-II) were employed to perform multi-objective optimization on the structural parameter combinations of the BTLFC (size ratio r, bifurcation angle θ, fractal order D). The optimized BTLFC structure exhibited superior comprehensive performance, with specific energy absorption increased by 10.19%, peak crushing force reduced by 12.27%,and mass decreased by 11.79% compared to traditional honeycomb cores. The findings provide novel biomimetic design insights for developing high-performance energy absorption structures.

A fault diagnosis method based on improved dung beetle optimizer (IDBO)-time varying filtered empirical mode decomposition (TVFEMD) with improved wavelet threshold functions was proposed aiming at that the vibration signal of rolling bearing fault tends to be disturbed and overwhelmed by strong noise background. IDBO was primarily developed to iteratively optimize B-spline order and bandwidth threshold ξ in TVFEMD，and the optimal parameter combination was obtained. Applying TVFEMD on the original signal, the decomposition for intrinsic mode function (IMF) component series were achieved, among which the irrelevant components were removed by correlation coefficient criterion, and target signals were reconstructed. Then the improved wavelet threshold function was employed on the new signal for further denoising.Finally, the envelope spectrum of the processed signal was calculated, from which the typical fault characteristic frequency was extracted. Through simulation signal and fault simulation test analysis, the fault diagnosis method combined with IDBO-TVFEMD and improved wavelet threshold function was compared with empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD) and complete EEMD with adaptive noise (CEEMDAN) denoising methods. The research results show that the algorithm model proposed in this paper has higher efficiency.

A set of anti-impact protection device was designed for the test fracture accident of 30 MN tension sensor calibration device. Firstly, based on the kinematic theory, the kinematic model of each component in the fracture process of test fixture was established. Then, different protective device structures at three impact locations were designed. Finally, the finite element models of three buffer structures were established and verified, calculated and optimized. The results show that the egg-box structure protection device at the top plate of the upper reaction rack and the upper ball joints can effectively solve the problems of small protection space and large impact value. The whole device dissipates 59.5% impact kinetic energy of the upper reaction rack, 60.7% impact kinetic energy of the lower reaction rack and 100% impact kinetic energy of the lower ball joints. After the improvement, the initial peak load of the protective device at the lower ball joints is reduced by 62.7%.

The gap between the rotor and the ball bearing in a wide temperature range increases with the increase in temperature. At the same time, the internal clearance of the bearing changes with the temperature. The inner ring of the bearing is subject to increased friction torque, which reduces its speed. This results in the deviation of the characteristic frequency of the outer ring defect, which is not conducive to the fault diagnosis of the ball bearing and the stable operation of the equipment. Considering the temperature-variable gap between the rotor and the bearing, a dynamic model of the bearing with an outer ring defect in a wide temperature range was established. The time-domain waveform and frequency of the model were analyzed. The simulation and test results show that the bearing vibration increases with the temperature, and the characteristic frequency of the outer ring defect decreases with the increasing temperature. Properly increasing the interference between the rotor and the bearing according to process requirements is beneficial for reducing the bearing system vibration in a wide temperature range and improving the accuracy of bearing defect frequency identification. The results provide reference for the use and health monitoring of ball bearings in a wide temperature range.