The wave generator, as the driving component, has its profile curve directly affecting the deformation and stress distribution of the flexible wheel. To reduce the stress on the flexible wheel and improve its fatigue performance, a design method for a three-term cosine cam wave generator was proposed.

The profile curve of the three-term cosine wave generator was composed of a constant term, represented by the base circle radius, and three cosine terms superimposed. By using the curvature variation coefficient at the major axis, the correction factor at the minor axis, and the clearance of the flexible bearing as variable parameters, the influence of these parameters on the performance of the wave generator was analyzed, and the optimization of a specific harmonic reducer model was validated. Finite element analysis was performed on the flexible wheel structure after incorporating the new cam structure under expansion, no-load, and loading conditions, and comparisons were made with the pre-optimized structure to analyze the stress distribution on the flexible wheel.

The finite element results indicate that, compared to the traditional cosine wave generator, the flexible wheel under the three-term cosine wave generator exhibits superior stress performance under various working conditions and achieves longer fatigue life.

A health state assessment method combining deep residual shrinkage network (DRSN) and adversarial domain adaptation (ADA) was proposed to address the problems of vibration signal noise interference and inconsistent data distribution under different working conditions in the remaining useful life (RUL) prediction of rolling bearings, so as to improve the accuracy and generalization ability of RUL prediction.

Firstly, a health state assessment model combining deep residual shrinkage network and adversarial domain adaptation was constructed. The performance of DRSN in avoiding noise in vibration signals and adaptively extracting bearing degradation features was utilized to build the health indicator curve. Then, ADA was used to align the distribution of health indicators between the test set and the training set, so as to eliminate the difference in data distribution under different working conditions. Finally, the health indicators output by the DRSN-ADA model were input into the convolutional long short-term memory (ConvLSTM) network model, and the accurate RUL prediction of rolling bearings was realized.

In the XJTU-SY dataset and engineering tests, the health indicators constructed by DRSN-ADA are superior to the comparison methods in monotonicity, robustness and correlation, with their mean values reaching 0.61, 0.97 and 0.98 respectively. The mean values of mean squared error (MSE) and mean absolute error (MAE) of the RUL prediction results are 2.52% and 2.19% respectively, and the average score is 0.86, which is significantly better than the DRN, principal component analysis and root mean square (RMS) methods. These results verify the effectiveness of the proposed method in noise suppression and cross-working condition prediction.

To address the problem that pure rolling bevel gears are prone to angular velocity mutation, vibration, and noise excitation under the influence of mounting errors, a tooth surface design method with low sensitivity to mounting errors was proposed.

Firstly, based on the influence of installation errors on angular velocity, the transmission error curve was preset as a parabolic type. Secondly, the theoretical tooth surface of the pinion was modified using the preset transmission error curve to establish a mathematical model of the target tooth surface. Finally, a 3D solid model of the modified bevel gear was constructed, and tooth contact analysis was performed.

The results show that the derived transmission error curve is consistent with the preset curve, validating the effectiveness of the modification. The actual meshing points of the optimized tooth surface deviate from the edges, and the overall contact area centers on the tooth surface, effectively avoiding edge contact and stress concentration.

Contact fatigue failure is the primary failure mode for gears operating under extreme conditions such as high speed, heavy duty, and high temperature. It has become a significant technical bottleneck limiting the development of aviation equipment toward higher reliability, longer life, and higher power density. Predicting contact fatigue life is now a critical focus of both engineering and academic research. The quantitative prediction research of gear contact fatigue life was carried out.

Firstly, over 400 rolling contact fatigue specimens made from 9310 aviation gear steel were prepared and subjected to more than 5 000 hours of fatigue testing. The investigation examined the effects of service conditions, such as contact stress and slip rate, as well as surface integrity states created by over ten processes including carburizing grinding, shot peening, fine particle peening, and rolling finishing, on rolling contact fatigue life. Secondly, the Lundberg-Palmgren （LP） theory was modified using multiple linear regression methods, resulting in an LP rolling contact fatigue life prediction formula based on service conditions and surface integrity parameters.

It is found that increasing the slip rate from 10% to 30% reduces the fatigue life at a contact stress of 3 000 MPa by 52.6%, from 2.28×10⁶ r to 1.08×10⁶ r. Additionally, when the contact stress is 3 000 MPa and the slip rate is 20%, dual shot peening significantly improves fatigue life, increasing it by 113.8% from 1.52×10⁶ r in the carburizing grinding state to 3.25×10⁶ r. The derived LP contact fatigue life prediction formula, considering service conditions and surface integrity, demonstrates an error margin within twice dispersion band, thus meeting engineering application requirements.

To solve the problem of no direct mechanical or hydraulic connection between the brake actuator and the brake pedal in an electronic mechanical braking system, leading to no feedback of road feel, a brake pedal feeling simulator was proposed based on magnetorheological dampers.

Ansys/Maxwell and Matlab/Simulink were used as the platform. The structural design of the sinking brake pedal feeling simulator, the structural design of magnetorheological damper, magnetic circuit analysis and electromagnetic simulation were carried out respectively, and the simulation analysis of traditional proportional-integral-derivative (PID) and fuzzy adaptive PID control under different working conditions of the whole system was compared.

The simulation results show that the brake pedal feeling simulator can track the characteristic curve of the traditional pedal well under different working conditions and has wide applicability. Compared with the traditional PID control effect, the fuzzy adaptive PID control has higher control precision and smaller control error, and has good application prospects.

To address the core problems of traditional magnetic tile-type magnetic wheels in the application of boiler wall-climbing robots, such as unclear magnetic field distribution, redundant safety factors of empirical formulas, low magnet utilization rate, and excessive weight, a magnetic tile arrangement scheme was identified with better magnetic attraction in the N-S alternating arrangement under the same volume and mass constraints. The differences in magnetic field distribution and magnetic attraction force magnitude when the magnetic wheel pressed against three versus two boiler water-cooled wall tubes were also clarified.

Taking a water-cooled wall inspection robot for thermal power plant boilers as an example, four magnetic wheel models with the same external dimensions but different magnet arrangement modes were established. Using Maxwell software, simulation analysis of the magnetic attraction force of these four magnetic wheels with different structures was conducted at a distance of 2 mm from the water-cooled wall tubes.

The results show that the adsorption force of the magnetic wheel on the water-cooled wall tube is 5%-20% greater when adsorbed at the joint of the magnetic tiles than at the middle of the magnetic tiles. When adsorbed at the middle of the magnetic tiles, regardless of whether the magnetic wheel presses against three or two wall tubes, the magnetic wheel with the 4-magnetic-tile arrangement exhibits the maximum attraction. Additionally, the adsorption force when pressing against three water-cooled wall tubes is approximately 10%-20% greater than that when pressing against two tubes.

To accurately and efficiently analyze the magnetic and dynamic characteristics of multishaft double ring-plate magnetic gears (MDRMGs), improve the working conditions of CPMG arm bearings, and extend their service life, a multishaft double-ring magnetic gear transmission structure was designed by combining magnetic gears with mechanical ring-plate gears.

A magnetic field unit classification method was proposed, and then a mathematical model for air-gap magnetic field and static torque was established, which was efficient and took the end leakage magnetic effect into account. Meanwhile, based on the Riccati transfer matrix method, a dynamic model of the eccentric shaft rotor system of MDRMG was constructed.

A comparison between the magnetic field unit classification method and the finite element method shows that the results of magnetic flux density and magnetic force obtained by the two methods are highly consistent, but the calculation time of the magnetic field unit classification method is shorter. The analysis also reveals that changes in the ring-plate spacing affect the classification calculation of magnetic field units and the lumped parameters in the dynamic model, the static magnetic torque of MDRMG increases with the increase of ring-plate spacing, while the critical speed of the eccentric shaft decreases as the ring-plate spacing increases. The magnetic field unit classification method can efficiently and accurately analyze the air-gap magnetic field and torque characteristics of MDRMG. In addition, the ring-plate spacing has a certain impact on the magnetic field and dynamic performance of MDRMGs.

To address the longitudinal vibration of long-distance belt conveyors during starting, the dynamic equations of the belt conveyor were established based on the analysis of the viscoelastic properties of the conveyor belt. Taking a practical long-distance belt conveyor as an example, a dynamic simulation model was constructed using AMESim software to study the starting acceleration curve and loading time.

The effects of common starting curves on the dynamic tension of the conveyor belt were analyzed, and a new combined starting curve (sine+parabolic) was proposed. The curve was optimized by introducing a creep phase and adjusting the pre-starting speed. The tension variations under different loading times were then analyzed.

The results show that the optimized sine and parabolic starting curve reduces the maximum tension of the conveyor belt by 5.8% compared with the commonly used sine acceleration curve. Furthermore, loading materials after stable operation effectively reduces the tension impact and extends the service life of the conveyor belt.

To address the issues of crawling and poor motion stability in low-speed and heavy-duty conveyor chain transmissions, a dynamic model of the above conveyor chain transmissions was established based on multi-body dynamic theory, and the research on dynamic characteristics was carried out, aiming to provide guidance for enhancing the stability of such conveyor chain transmission systems.

By considering the structural characteristics of low-speed and heavy-duty conveyor chain transmission systems and the effects of conveying loads, the mechanical properties of the conveyor chain nodes were simulated using Hooke’s law, and a dynamic model with 107 degrees of freedom of the conveyor chain was constructed. The modal superposition method was adopted to solve the dynamic model, the dynamic characteristics of the low-speed and heavy-load conveyor chain were obtained, and the influence laws of the slack side sag and the tight side friction force of the conveyor chain on the transmission speed variation were studied.

The results show that reducing the slack side sag and decreasing the friction coefficient between the conveyor and the guide rail can effectively reduce the occurrence of the crawling phenomenon in conveyor chain transmission and improve the motion stability of the conveyor chain transmission systems.

Aiming at the problems of unbalanced magnetic pull (UMP) and low structural strength of high-speed rotor in the operation of permanent magnet assisted magnetic gear, relevant research was conducted.

Firstly, the phase tuning method was used to study the influence of different transmission ratios on the UMP. Then, according to the structural characteristics for the high-speed permanent magnet rotor of the magnetic gear, the analytical solution of the rotor strength was obtained by using the equivalent mass ring method. Finally, taking reducing the maximum stress and ensuring a certain output torque as the optimization objectives, the multi-objective optimization was performed on the relevant parameters of the magnetic isolation bridge and magnetic barrier.

The analysis results indicate that a transmission ratio where the maximum common divisor of the number of magnetic blocks and the number of poles of the low-speed rotor G_{\mathrm{C}\mathrm{D}}=\mathrm{2} can effectively decrease UMP. The relative error between the analytical solution of the maximum stress obtained from the equivalent mass ring method and the result of finite element simulation is less than or equal to 1%, validating the accuracy of the analytical method. Through the optimized design, the maximum stress on the rotor is significantly reduced.