The permutation flow shop problem with two inter-machine buffer conditions (i.e.limited buffer,zero-wait) is studied arising from manufacturing industry,and a mathematical programming model is established with the objective of minimizing maximum completion time.A mixed artificial bee colony algorithm is then proposed combined with improved two-segment Tent chaotic mapping,adaptive Cauchy mutation and greedy algorithm. Firstly,the initial job sequence population is generated by improving two-segment Tent chaotic mapping. Then,in the employed bee phase,the neighborhood search based on adaptive Cauchy mutation is applied to generate new job sequences. In the onlooker bee stage,a selection fitness strategy and the inverse-order reversal operation based on adaptive Cauchy mutation are designed to optimize the job sequences.In the scout bee phase,unimproved job sequence is updated based on key/non-key factories by greedy algorithm. Finally,a large number of simulations and comparisons with various algorithms show that the proposed algorithm can obtain a good near-optimal solution in a reasonable computational time.

In the process of virtual debugging,there is a problem that the data of virtual or real controller and virtual model are not synchronized,which leads to the confusion of virtual process flow,seriously affects the quality of virtual debugging system,and makes the digital twin system has the problem of asynchronization between virtual and reality. Taking the bearing ring size detection line as an example,a virtual debugging system based on Mechatronics Concept Designer (MCD) was designed,the asynchronous problem of the virtual debugging system was analyzed,and the internal and external delay test of MCD was designed. Based on the analysis and explanation of the existing solutions,a solution considering the internal and external delay of MCD was proposed. The experiment shows that this solution can realize the synchronization of virtual debugging system and digital twin system.

In order to solve the wind turbine extrusion panel production scheduling problem with complex features such as packaging order flush and product changeover adjustment,a multi-objective collaborative optimization model was constructed to maximize the average utilization rate of the current starting equipment and maximize the order fulfillment rate;the wind turbine extrusion panel production scheduling problem was transformed into a Markov sequence decision-making problem,and 10 different scheduling strategies were designed as the action space,and the appropriate state features and reward functions were refined; a scheduling algorithm based on Dueling Double Deep Q Network (D3QN) was proposed. The effectiveness of the proposed algorithm was verified by comparing with Double DQN and Dueling DQN algorithms through simulation tests on actual data of an enterprise;and the objective values obtained by four different solution methods under 10 algorithms were compared,which verified that the proposed improved D3QN algorithm can get a high-quality solution to the problem,providing an intelligent method and reference for the production scheduling of wind turbine extrusion panel manufacturing enterprises.

The motion control of quadruped robots in unstructured terrains heavily relies on complex dynamics models and controller designs. Utilizing Deep Reinforcement Learning (DRL) methods to design controllers for quadruped robots has become a trend. To address issues such as slow convergence, susceptibility to local optima, and high computational resource consumption during the training process of DRL, an improved algorithm called Memory-integrated Twin Delayed Deep Deterministic policy gradient (M-TD3) was proposed. Firstly, models for both the quadruped robot and unstructured terrains were established. Then, the convergence state and learning efficiency of M-TD3 algorithm were analyzed. Finally, motion control simulation comparisons for various terrains were carried out, and production of prototype was tested to verify controller performance. Simulation results show that compared with traditional TD3 algorithm, M-TD3 algorithm has faster convergence, higher efficiency, and significantly improves motion control performance. Prototype test results show that the controller designed based on the improved TD3 algorithm can make quadruped robot effectively move over obstacles in unstructured terrain.

In view of the problem that the robot grinding trajectory deviates from the actual workpiece weld due to the manufacturing error of the workpiece and the positioning and clamping error in the current robot weld grinding process,resulting in poor consistency of grinding quality,an adaptive correction system of the robot weld grinding trajectory based on the feature matching of the point cloud model was developed. The system uses the surface structure light camera to scan the feature area of the workpiece to obtain the real morphology of the workpiece,matches with the standard point cloud template,and obtains the deviation conversion matrix of the actual workpiece coordinate system,which is applied to the robot master program to realize the rapid guidance of the initial point of welding grinding and the online adaptive correction of the grinding track. Finally,the experimental results show that the system has good performance of automatic correction and stability.

Aiming at the problem of robot path planning in workshop dynamic environment,a path planning method based on fuzzy DWA algorithm guided by prior knowledge was proposed. Considering the invariability of the static environment of the workshop,an adaptive A^* algorithm for global path planning was designed,and it was used as a prior information to guide the robot movement. The robot kinematics model was established,and a local path planning method based on fuzzy DWA was proposed under the guidance of prior information. This method can fuzzy adjust the weight of the evaluation function according to the complexity of the environment. The experimental results show that in the 30 m×30 m grid environment,the path length planned by the hybrid A^* algorithm is 51.32 m,and the path length planned by the adaptive A^* algorithm is 46.83 m,which is 8.75 % less than that of the hybrid A^* algorithm; in the local path planning,the security,flexibility and length of the fuzzy DWA algorithm are better than the standard DWA algorithm,which verifies the superiority of the fuzzy DWA algorithm.

A control method for an upper limb flexible exoskeleton system based on adaptive integral global sliding mode and extended state observer was proposed to address the problem of difficulty in high-precision trajectory tracking caused by external disturbance and parameter uncertainty in the upper limb flexible exoskeleton system. Firstly,the upper limb flexible exoskeleton system was modeled by Lagrangian function,and the model was decomposed into two subsystems for cascade control. Secondly,two extended state observers were used to compensate the parameter uncertainty and external disturbances. In the position trajectory control,an adaptive integral global sliding mode control was constructed,in which a nonlinear integral function was designed to avoid integral saturation. At the same time,an attenuation function was established to accelerate the convergence speed of the global sliding mode surface.The reaching law coefficient may be adaptive adjusted using adaptive control method to reduce chattering and to improve the dynamic performance of the system. Finally,the stability was proved by designed Lyapunov function. The experimental results show that compared with traditional linear active disturbance rejection control and integral sliding mode control,the proposed method has a more accurate trajectory tracking effect and stronger robustness, the control accuracy is improved,and high-precision trajectory tracking is still maintained under external disturbances and system parameter uncertainties,indicating the effectiveness of this control strategy.

The textured surface is a type of the structured surfaces,which has the effect of reducing friction.In order to obtain the regular textured surface of the inner hole,the tool structure of the super-hard abrasive reamer was designed firstly,and then the abrasive grains on the super-hard abrasive reamer were arranged in phyllotactic order,and the reaming simulation was carried out by using Matlab software. The effects of phyllotaxis coefficient and feed speed on the surface mesh morphology and parameters were studied.The simulation results show that the textured surface can be hinged by a one-time tool. The phyllotaxy coefficient affects the plateau support rate of the textured surface.The larger the phyllotaxis coefficient,the higher the plateau support rate of the textured surface. The feed speed affects the regularity,reticulation angle and plateau support rate of the textured surface. The higher the feed speed,the larger the reticulation angle and the higher the plateau support rate.

Based on the practical application of injection production integrated string technology in Bohai oil field and related literature,the performance analysis of the special shaped blade spring in the automatic steering tool was carried out. The high temperature compression and the one year high temperature and atmospheric pressure test of the special shaped blade spring were completed,and the changes of its mechanical properties under high temperature conditions were obtained. Based on the test results,it can be seen that the blade spring will reach a stable state after working at 250 ℃ for about half a year,and its maximum elasticity will be reduced by about 10 %~30 %. Meanwhile,the material parameters of the blade spring were modified. Based on the modified material parameters,a total of 35 sets of comprehensive dimensional simulation analysis were carried out on the thickness and width of the blade spring. The relationship curves between the thickness and width of the spring on the maximum elastic force and the maximum plastic deformation were obtained,and the mathematical model was established,which can provide reference for the practical application of the steering tool.

In order to analyze the influence of the modification amount of pinion on the meshing performance of face gears,the tooth surface equation of the face gear pair was obtained based on the rack model of pinion,the motion trajectory of the dressing grinding wheel was calculated,and the modification amount of pinion was solved through the derivation of the equidistant curve of the cross-section profile of the grinding wheel. The relationship between tooth surface modification amount and modification parameters was analyzed based on the calculation results of pinion modification amount under different modification parameters,and tooth contact analysis of the gear tooth surface under different modification amounts was completed. The correlation law between pinion modification amount and the meshing performance of the face gear was obtained,these results provided a theoretical basis for the precision machining of the face gear.

In order to enhance the sealing of car compartments, a commonly employed method in the past involved using a short-cycle bolt welding approach to secure bolts and reduce perforation count. However, the prevalent destructive testing method suffered from drawbacks such as low detection efficiency and an inability to identify welding joint issues during service. Addressing these issues, it proposed a bolt welding quality inspection scheme that combines dynamic welding process parameters with weld seam images. To overcome the insufficient number of unqualified samples in dynamic parameters, a distance-weighted data augmentation algorithm was introduced. Additionally, a deep learning model incorporating Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) was designed to achieve bolt welding quality inspection. Experimental results validated the effectiveness of the proposed scheme, demonstrating a 95.33 % accuracy in defect bolt identification. Hence, this short-cycle bolt welding quality inspection method, based on dynamic parameters and weld seam images, showcases high detection precision and provides an effective quality inspection approach for practical engineering applications.

Aerospace parts require high precision and a qualified rate of 100 %.Three-coordinate measurement is commonly adopted to improve the measurement precision and reduce the error caused by personnel measurement. High-precision and thin-wall shaft sleeve as a connecting part playing an axial positioning and radial fastening role in the work,plays an important role in mechanical transmission. In order to ensure the measurement precision and qualified rate of the parts,it is necessary to control the measurement process error within 1/3 of the part tolerance,including positioning error,clamping deformation error,etc. The experiment used SLA prototyping flexible fixture to clamp the shaft sleeve to reduce the deformation and damage caused by the metal clamp and eliminate the deformation error. Using the repeated measurement of the same shaft sleeve parts,the standard uncertainty caused by the positioning error combined with the process ability calculation was obtained to determine the feasibility of the measurement method.

In order to solve the problems of low efficiency and high energy consumption of cyclone separator in the actual separation process,a slit at the overflow pipe of cyclone separator and couple with a screen was proposed,according to which the structural models of single-layer,double-layer and triple-layer slits of the overflow pipe were designed,and 2 types of screens of 150 mesh and 200 mesh were coupled with the slit respectively,and 6 kinds of improvement schemes were obtained. The pressure drop error between the actual model and the porous step model was compared,and the effects of the number of layers of the overflow tube and the number of meshes on the separation efficiency under different velocities were analysed. The results showed that when the flow rate was greater than 10 m/s,the relative error in pressure drop between the actual model and the porous step model for the 2 types of 150 mesh and 200 mesh meshes was 0.70 % and 0.40 %,respectively,and this validation demonstrated the reliability of the porous step model. At the same inlet flow rate,the greater the number of layers of the overflow tube,the lower the tangential velocity of the cyclone. At the same inlet velocity,the more slotted layers,the lower the tangential velocity,the lower the pressure drop,and the slightly lower the cyclone separation efficiency. At the same number of slotted layers,the tangential and radial velocities under the 150 mesh screen coupling were lower than those under the 200 mesh screen coupling,and the maximum reduction in pressure drop of the improved cyclone could reach 18.76 %. The separation efficiency of the 6 improved cyclones under a 20 m/s inlet velocity is 97.25 % to 97.59 %, which is very close to the originel cyclone′s sepration efficiency of 97.63 %. The design of the cyclone separator overflow tube with open slot and coupling screen can effectively reduce the energy consumption under the premise of maintaining the efficiency.

The fault diagnosis method based on current signals has advantages in signal acquisition and anti-interference.It is difficult to extract gear fault features,due to the power frequency component. Therefore,removing the power frequency component from the current signal is a significant task,especially when the gearbox is in variable speed conditions. At present,there is still no ideal method for diagnosing planetary gearbox faults under variable speed conditions. A planetary gear fault diagnosis method based on reactive power,Gramion Angular Field,and average teacher semi-supervised model was proposed. Firstly,the power frequency component was removed from the time-varying current signal by calculating the reactive power. Then,Gramion Angular Field was used to transform 1D temporal data into 2D data with more complete feature information. Finally,the fault recognition of planetary gears was achieved based on an improved semi-supervised learning model. The experimental results showed that the recognition accuracy could reach 92.31 % when using 20 % labeled data.

In response to the issue of high maintenance cost and the failure to fully utilize the remaining service life of the system during the maintenance of CNC machine tools,based on the Log-Linear Process (LLP),a new method for non-equal cycle imperfect Preventive Maintenance (PM) of Computer Numerical Control (CNC) machine tools was proposed. The hybrid failure intensity model of imperfect PM of machine tools under the LLP was established,and the expressions of the age reduction factor and the failure intensity increasing factor were given according to the empirical value method. With reliability as the constraint,the dynamic maintenance intervals under different PM thresholds and replacement thresholds were derived. Furthermore,the optimal number of PM times was given to minimize the average maintenance cost rate. The proposed method was validated using the spindle system of CNC machine tools as an example. Comparison analysis was conducted between the maintenance model without considering replacement threshold and the maintenance model with mixed failure intensity factors following a uniform distribution,the results indicated that the proposed method can effectively reduce maintenance costs while maximizing the remaining life of the equipment. The research results provide a new method for formulating a reasonable PM strategy for CNC machine tools.

In order to suppress the longitudinal vibration of the mine hoisting system,a longitudinal vibration-damping suspension device based on magnetorheological damper was designed. Firstly,a magnetorheological damper mechanical model was established,and the structural design and magnetic circuit design of the vibration-damping suspension damper were carried out; secondly,a multi-physical field coupling model of the magnetorheological damper was established based on the multi-physical field simulation software COMSOL,and the internal magnetic field,the flow field and the pressure distribution conditions were analyzed in the flow-solid coupling environment of the electromagnetic field and the flow field to validate the reasonableness of the design of the mechanical structure and the distribution of the magnetic circuit; and then,through the dynamic simulation of coupled electromagnetic field,flow field and force field,the relationship between the damping force F and displacement x,frequency f,amplitude A,current I and damping adjustable coefficient β under sinusoidal excitation of the magnetorheological damper was obtained; the simulation results show that,in the range of operating current from 0~1.4 A,the maximum damping force of the designed magnetorheological damper was 4 993 N,and the maximum damping adjustable coefficient was 9.97. Finally,the trial prototype verifies that the designed magnetorheological damper can provide the required damping force for the vibration-damping suspension device to achieve the suppression of longitudinal vibration of the mine hoisting system.

Robust identification of the running state of transmission parts is of great significance to ensure the service performance of electric mountain wreckages.In the high-frequency variable load operation and high noise environment,the collected vibration signals are often affected by complex transmission paths and component coupling,resulting in the interference of irrelevant noise,which brings challenges to the accurate diagnosis of bearing faults of mountain wreckages. A fault diagnosis model of rolling bearing based on frequency domain Granian Angular Fidles and multi-scale Residual Networks (ResNet) is proposed. The frequency component of time series signal is reconstructed by Granian Angular Fidles,and the features are weighted and enhanced by multi-scale attention enhancement mechanism. This enables the model to adaptively focus on the features that are most important for fault diagnosis while suppressing the impact of noise. The residual connection is introduced to promote the training and information flow of the deep network,so as to promote the learning of complex features. The fault accelerated life test data set of rolling bearing in Xi’an Jiaotong University and the rolling bearing data set of mountain wrecking truck collected by using the rolling bearing test bed are used for verification. The fault recognition rate in the two data sets is more than 99 %,which verifies the effectiveness of the proposed fault diagnosis model. Compared with different variable load conditions,the fault recognition rate of the model is more than 98.5 %,and the recognition rate is still more than 90 % in the high-frequency variable load environment of -6 dB noise,which further verifies that the model can be used for bearing fault recognition of mountain wrecktrucks and has good robustness and generalization ability.

To tackle the challenge of rolling bearing fault diagnosis under conditions of weakly labeled data samples,a novel approach utilizing self-supervised learning within the time-frequency was developed to unearth latent fault features in fault-free labeled samples.Initially,feature representations were extracted from both the time and frequency domains by constructing respective time and frequency encoders. A model for the time-frequency self-supervised learning was designed to augment the mutual predictive capabilities between features across these domains. Furthermore,to refine the model learning process,a novel cross-correlation matrix loss function was devised,significantly enhancing the model′s proficiency in identifying complex failure modes. The efficacy of this method was confirmed by using both the Case Western Reserve University bearing fault open dataset and the Paderborn University bearing open dataset,where the experimental outcomes indicated superior diagnostic results with minimal fault-labeled data.

Aiming at the difficulties in fault vibration signal feature extraction and low recognition accuracy of reciprocating compressor plain bearing,an improved Mutiscale Entropy (MSE) bearing fault feature extraction method was proposed. Firstly,in order to solve the problem of redundant calculation and large amount of calculation in the calculation step of sample entropy,the symbolic idea was introduced to simplify the repeated calculation problem when counting the number of vectors whose distance between vectors was less than a threshold under different dimensions,and a fast sample entropy algorithm was obtained. Secondly,in view of the fact that the traditional cubic spline interpolation method cannot meet the complex and variable time series signal multi-scale processing,cubic triangular B-spline interpolation was proposed to replace the traditional cubic spline interpolation to multiscale the multi-scale sample entropy and improve the accuracy of MSE. The previous complex compressor sliding bearing clearance fault was the research object,and the improved MSE method was used to extract the fault signal feature. The research results showed that when the sample length was 24 056,the computational efficiency of the improved MSE algorithm feature extraction increases by nearly 9.12 times,and the percentage of increase was 816.62 %. At the same time,this method improved the accuracy of fault diagnosis and recognition of the original MSE method.