For slotted disk magnetic couplers, the slotted conductor disk complicates the magnetic circuit during operation, increasing the difficulty of calculating electromagnetic torque. A reliable electromagnetic torque prediction model was established, and its torque characteristics were studied.

A 18-pole 16-slot disk magnetic coupler was studied. Firstly, via the equivalent magnetic circuit method, induced eddy currents from adjacent and self-magnetic circuits were introduced as a branch into a new model. Considering 3D end effects and combining Ampere's and Kirchhoff's laws, air gap flux and output torque expressions were derived. Secondly, the coupler was simulated and analyzed by the finite element simulation software, obtaining distributions of air gap magnetic field and eddy currents, and torque variations with air gap thickness and speed difference under adjacent eddy current influence. Finally, a test platform was established to verify theoretical and simulation results.

The results show that the results of theoretical calculation, finite element simulation and test are basically consistent. The proposed theoretical model has high accuracy, providing a reliable prediction model for studying the torque performance of such couplers.

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.

In order to improve the torque to volume ratio of magnetorheological （MR） brakes, a MR brake with internal and external fluid flow channels was developed.

Firstly, the structure and working principle of the MR brake with internal and external fluid flow channels were introduced, and the mathematical model of the braking torque was established based on Bingham constitutive model. Secondly, in order to obtain the optimal structure size of the MR brake, structural optimization design was carried out based on the non-dominated sorting genetic algorithm (NSGA-Ⅱ). Finally, the prototype of MR brake was fabricated, the torque performance test system was built, and the braking performance test was conducted for the MR brake.

There are both internal and external fluid flow channels in the MR brake. Six effective damping gaps are obtained by reasonable setting of magnetic conductive and magnetic isolating materials in the MR brake, so that it can produce excellent torque performance under the premise of the same size. The theoretical calculation results show that the braking torque and the adjustable range after optimization are increased by 30.23% and 16.58% respectively compared with those before optimization. Test results show that at the applied current of 2.0 A, the maximum braking torque is 44.28 N·m, and the dynamic adjustable range is 17.8. The relative errors of the braking torque and the dynamic adjustable range between the test values and the theoretical calculation values are 6.5% and 16.1% respectively, which verifies the rationality of the design.

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 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.

Alternating magnetic fields induce vibrations in mechanical components, thereby generating noise. Fluctuations in electromagnetic excitation forces and electromagnetic torque are the primary causes of electromagnetic vibration noise. To analyze the generation mechanisms and functional patterns of these fluctuations, an electromagnetic vibration noise analysis was conducted on a slotted disk-type asynchronous magnetic coupler with 9 pole pairs and 16 slots.

Firstly, theoretical formulas for air-gap magnetic flux density and electromagnetic excitation force were derived using the magnetic scalar potential permeance method and Maxwell stress tensor method. Combined with finite element simulation, the harmonic order amplitudes of the Fourier decomposition of air-gap magnetic flux density and electromagnetic excitation force were obtained. Secondly, based on the energy method, an expression for cogging torque was derived. Finite element simulation was employed to determine the cogging torque and electromagnetic torque fluctuations generated during the operation of the magnetic coupler. Thirdly, an electromagnetics-structural-acoustic multi-physics coupling model was established. Using the modal superposition method, vibration acceleration and displacement produced during stable operation of the magnetic coupler were obtained, and the characteristics of its electromagnetic noise were analyzed. Finally, a test platform for the magnetic coupler was constructed to measure electromagnetic noise during stable operation. Test results were compared with simulation outcomes to validate the theoretical analysis.

The results indicate that low-order electromagnetic excitation forces are the main causes of vibrations in the magnetic coupler, and significant vibrations occur when the frequency of the electromagnetic excitation force approaches the natural frequency of the magnetic coupler. Comparison with simulation results shows that the test data obtained from the magnetic coupler test platform confirm the accuracy of the theoretical analysis.

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 solve the problems of fixed damping force and poor pseudo-humanity of traditional lower limb prosthetic knee joints, a magnetorheological damper was designed to meet the vibration reduction requirements of the lower limb prosthetic knee joint.

Through theoretical calculation, the maximum damping forces required for the knee joint swing phase during flat walking and flat running were obtained, which were 179.6 N and 1 377 N respectively. In order to adapt to the motion state of the lower limb prosthesis, a vibration absorber was designed to meet the damping force required by the knee joint swing phase. Through numerical simulation and test research, the influence of external disturbed magnetic field and temperature rise effect on the dynamics characteristics of magnetorheological damper were analyzed. The test of influence of external disturbance magnetic field and temperature rise effect on the dynamics characteristic of the lower limb prosthetic knee joint was conducted by using the lower limb prosthetic knee joint simulator.

The results show that the output damping force of the magnetorheological damper increases with the increasing magnetic flux density of the external disturbed magnetic field. Under the same conditions, the output damping force of the magnetorheological damper decreases with the rising temperature of the magnetorheological fluid. In the early stage of knee joint swing and the first half of its middle stage, with the increase of the magnetic flux density of external disturbed magnetic field, the hysteresis of knee joint movement increases, and the angle error increases. When the magnetic flux density of external disturbed magnetic field is 10, 20 and 30 mT respectively, the maximum bending angle of the lower limb prosthetic knee joint is 59.0°, 57.8° and 55.7° respectively, and the maximum angle error reaches 3.0°, 6.8° and 11.9° respectively. As the rise of the temperature of the magnetorheological fluid, the hysteresis of knee joint movement increases, and the angle error increases. When the temperature of the damper rises to 30, 35 and 40 ℃ respectively, the maximum bending angle of the lower limb prosthetic knee joint is 57.1°, 54.0° and 49.8° respectively, and the maximum angle error reaches 1.9°, 5.1° and 9.8° respectively. These conclusions provide a basis for the design and optimization of the lower limb prosthetic knee joint based on the magnetorheological damper.

The risks of losing magnetism upon power-off and high energy consumption are suffered by traditional electromagnetic grippers. And the problem of lacking the active magnetic force regulation function is still faced by permanent magnegrippers. A lightweight magnetic pole rotation gripper with optimized methodology was developed.

Based on the magnetic flux continuity principle and magnetic field superposition effect, the optimal design strategy integrating theoretical analysis, numerical simulation, and test verification was established through coordinated regulation of three key parameters: magnetic pole rotation angle, geometric dimensions, and air gap distance. The global optimal solution of the magnetic pole structural parameters was finally obtained.

Both simulation and test results demonstrate that the optimized gripper achieves superior magnetic adhesion performance per unit mass compared to existing models. The proposed device exhibits distinctive advantages including compact structure, simplified control mechanism, and quasi-linear control characteristics, showing broad application potential in material handling operations.

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 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.

The pipe belt conveyors are crucial equipment for bulk material transportation with significant environmental advantages. This study is aimed to quantify the lateral bending stiffness of steel cord conveyor belts.

Based on the analysis of standard ISO 703:2017, the measurement and analysis method for the lateral bending stiffness was determined. Numerical model and 3D simulation model of the steel cord conveyor belt were established. Deformation data under different schemes was obtained using numerical analysis method and finite element method. Error analysis was conducted to demonstrate the validity of the models. Furthermore, a generalized deflection formula for the lateral bending stiffness was derived based on the functional dependence of the belt's troughability on elastic modulus, linear mass, and cross-sectional geometric parameters.

The results provide a new perspective for quantifying the lateral bending stiffness of pipe conveyor belts and offer a basis for their design and engineering practice.

Existing research on flexible hinges corresponds to complex expressions for flexibility and rotational accuracy calculations. To address this issue, a new type of catenary flexure hinge was designed, and a method for establishing the hinge’s compliance and rotational accuracy model by approximating arc segments with straight-line segments was proposed.

Firstly, by defining the flexure hinge as a series combination of tapered and expanded sections, the curve in the tapered segment was divided into several arc segments, and the curve segments were approximated with straight segments. Based on the Castigliano’s second theorem, a method by calculating the flexibility of the tapered section and then establishing the hinge flexibility and rotational accuracy model through matrix operations was established. Secondly, using specific examples, the derived formula, literature formulas, and the finite element method were employed for calculations. When the curve segment was finely divided, the calculation results align well, thereby verifying the formula’s correctness. Thirdly, the influence of structural parameters on the flexibility, rotation accuracy, and flexibility-accuracy ratio of catenary flexure hinges was analyzed. Finally, the bending flexibility and flexibility-accuracy ratio of the catenary, conic, and their hybrid hinges were analyzed with the same structural parameters.

The results show that a single parameter has a negative correlation with flexibility and rotation accuracy of the catenary hinge, and reducing the minimum thickness is the best way to improve flexibility. Under the same structural parameters, the flexibility and flexibility-accuracy ratio of the catenary hinge is between parabolic and circular shapes. Choosing a hybrid hinge with a section of high flexibility for the tapered section and a section of low flexibility for the expanded section allows for a balance between flexibility and motion accuracy.The greater the difference in flexibility, the better the flexibility-accuracy ratio.

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.

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 novel two-stage compound amplification mechanism design scheme was proposed to address the friction and clearance issues inherent in traditional revolute pairs within the large stroke design of micro-displacement platforms. The aim was to achieve high-precision and significant stroke displacement amplification through structural innovation.

Utilizing the theory of material mechanics, a static model was established. A two-stage compound amplification structure that integrated a flexible hinge lever amplification mechanism with a bridge amplification mechanism was employed. Piezoelectric ceramics served as the driving source, and a parameter optimization model was developed using Matlab software. The impact of key structural parameters on both the amplification ratio and input stiffness was systematically analyzed to identify the optimal parameter combination. The optimized structure underwent validation through multi-physical field simulation via finite element analysis.

Following optimization, the mechanism attains an impressive displacement amplification ratio of 13.1 times, with its natural frequency reaching 92.2 Hz. The maximum discrepancies between theoretical calculations and simulation results of the amplification ratio and natural frequency are recorded at 2.4% and 3.5%, respectively, thereby demonstrating the feasibility of this structural design.

The industrial robot industry has put forward higher requirements for RV reducers, and the precision life reflects the ability of the reducer to maintain transmission accuracy, which is one of the most important design criteria and usage indicators. To improve the precision performance of precision reducers, it is crucial to evaluate their reliability. Therefore, the degradation characteristics of precision reducers were analyzed.

Taking the RV80E reducer as an example, a random degradation model based on Gamma process was proposed. Combined with the performance degradation data of the reducer transmission accuracy, the model parameters were estimated based on the matrix method and the maximum likelihood estimation method. A Gaussian process regression model optimized by genetic algorithm was established using vibration characteristic data to optimize the prediction of transmission accuracy.

The results show that the prediction accuracy based on Gaussian process regression model is significantly better than that of the traditional regression model. The posterior distribution parameters of the random degradation model are updated by using the algorithm to predict the results, which can effectively evaluate the reliability of the accuracy life of RV reducer and lay the foundation for further reliability optimization design of accuracy life.

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 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.

In order to increase the depth of planting seedlings and reduce the width of transplanting holes, so as to improve the uprightness of rice pot seedlings after transplanting, the comprehensive design and test research of the transplanting mechanism based on the belt-shaped transplanting trajectory were carried out.

Firstly, a belt-shaped transplanting trajectory that can reduce the width of the hole was proposed, and the posture analysis of 10 key posture points on the transplanting trajectory was conducted according to the parameters of the transplanting mechanism and the requirements of motion design. Secondly, the seven-tooth non-circular gear planetary gear train transplanting mechanism was simplified into a planar 2R open chain mechanism, and a mechanism mathematical model based on the belt-shaped transplanting trajectory was constructed using the multi-position motion synthesis theory. Then, the objective function of the minimum change of the length of the planet carrier was established, and the App Design module of Matlab software was used to develop the multi-position motion comprehensive solution software for the rice pot seedling transplanting mechanism based on the genetic algorithm, and the optimal solution set of the key parameters of the transplanting mechanism was obtained. Finally, a simulation on the kinematic characteristics of the transplanting mechanism was carried out using Adams software, and a test prototype was fabricated to conduct the test research on field transplanting.

The simulation results show that the transplanting mechanism has the characteristic of quick return, and all the parameters and kinematic characteristics of the mechanism meet the design requirements, where the planting depth and transplanting hole width are 22.19 mm and 9.78 mm, respectively, which represent increases of 41.9% and decreases of 27.9% respectively compared to the corresponding value in relevant literature. The results of field test show that the average planting depth of rice pot seedlings is 22.84 mm, the average hole width is 10.37 mm, the excellent rate of verticality is 93.3%, and the transplanting effect is satisfactory.

In response to the visualization requirements of the models and analysis results necessary for the development of simulation analysis software for spiral bevel gear transmission systems, based on the structural characteristics of the bevel gear transmission system, the topological structure of the transmission system and the complete expression of the interrelationships among components were achieved by applying graph theory and object-oriented data structures.

The generation method of regularized point sets on the geometric surfaces of key heterogeneous components in the transmission system was investigated, and the precise construction and rapid assembly of the geometric models of bevel gears, transmission shafts, and bearings were accomplished using the open-source 3D computer graphics tool VTK. On this basis, the mapping between the geometric model of the transmission system and the mechanical model of loading contact analysis was established, and the visualization methods for the analysis results such as system deformation under loading and tooth surface meshing state were studied. Finally, the visualization effects of modeling and simulation analysis results of the bevel gear transmission system were verified through examples.

Research has shown that the comprehensive application of graph theory and object-oriented data structures can achieve a complete expression of the topological configuration and geometric correlation properties of bevel gear transmission systems. By applying the parameter expression modeling, sweep modeling, and triangulation modeling methods of the VTK library, accurate modeling of heterogeneous components can be achieved. Based on this, a mapping between the system geometry model and the mechanical analysis model can be established to complete the loading contact analysis of the bevel gear transmission system, and visualize the analysis results. The above research results provide technical support for the development of modeling and simulation analysis software for bevel gear transmission systems.

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.