Aiming at the optimization of internal hole layout in hydraulic manifold block,an optimization model was established to minimize both the length of hole pathways and pressure loss.Leveraging the structural characteristics of hydraulic manifold block,a multi-endpoint hole layout optimization algorithm based on the Rectilinear Steiner Minimal Tree (RSMT) was proposed. This algorithm transforms the three-dimensional problem of optimizing the hole pathway into a two-dimensional problem of optimizing connectivity among endpoint sets.Initially,the hole layers were assigned based on the spatial coordinates of each endpoint,and then the hole layers were merged onto the same plane.The optimization objective of the hole pathway was achieved by solving the minimum Steiner tree within the plane.The RSMT algorithm,built upon Kruskal algorithm,was employed to construct the minimum Steiner tree with right angles.Subsequently,the edges of the resulting RSMT were matched to the corresponding hole layers of endpoints,yielding the optimized overall connectivity of the hole network.The validation results from practical examples demonstrate that the layer assignment and connectivity algorithm can efficiently and effectively perform the optimization design of hole connectivity for manifold block with multi-endpoints.

To solve the problem of assembly line balance problem with a determined production cycle and improve the balance degree of assembly line, a single-objective assembly line balance optimization model was established. According to the different priority weighting rate of optimization goals, the linear weight sum method was used to establish the model. The optimization goals of the model are the assembly line smoothing index and assembly line balance rate, considering the assembly job allocation, the number of workstations and other factors. The hybrid algorithm of Genetic Algorithm (GA) and Ant Colony Optimization (ACO) algorithm for solving the model was improved. The new fitness function and distance information matrix were constructed. Finally, numerical experiments are carried out on the classical examples. The experimental results show that the improved model and algorithm can improve the balance of the assembly line to a greater extent and verify the effectiveness of the model and algorithm.

With the progress of society,production safety and workshop management issues are receiving increasing attention.The management of traditional workshops mostly relies on manual work,and managers often fail to timely detect and identify disturbances in the workshop.This is not conducive to quickly resolving disturbance events and ensuring the safety of personnel and equipment.In order to improve management efficiency and ensure security,a disturbance determination method based on BA-PNN algorithm and digital twin technology was proposed.First,data were collected through sensors,the data were analyzed and preprocessed.Subsequently,the traditional PNN model and BA-PNN disturbance judgment model were built,and the latter aims at the optimization of algorithm recognition rate.BA-PNN model was integrated into twin platform by digital twin technology.Finally,through simulation and result analysis,compared with the model effect before optimization,the recognition effect of the model is significantly improved,while reflecting the characteristics of digital twins,proving the effectiveness of the method.

Aiming at the lack of existing research on mobile robot monitoring system and the defects of poor visualization,low transparency and low flexibility of existing monitoring system,a three-dimensional visual monitoring system of mobile robot based on digital twin was proposed.The framework of monitoring system was introduced,based on this,key technologies such as virtual model construction,data acquisition and mapping in the system were elaborated. In order to solve the problem of real-time state mapping,a virtual real-time mapping technology was proposed.Through online acquisition and twin mapping,the real-time information interaction was realized.A prototype three-dimensional visual monitoring system for mobile robot was developed,which proved the effectivity of the proposed system.

In order to solve the problems of poor real-time monitoring,insufficient accuracy and low visualization degree in the production process of spinning machine sheet metal workshop,a method of constructing visual monitoring system of spinning machine sheet metal workshop based on digital twin was proposed.Firstly,the twin geometric model was constructed based on the production factors of the sheet metal workshop of the spinning machine,and the twin geometric model was lightweight to meet the real-time response requirements.Then,the behavior logic model was designed based on time-assigned Petri nets,and the trigger response in the twin shop was realized by collision detection algorithm.On this basis,the information modeling and event-driven methods of OPC UA were combined to realize the faithful mapping of product processing and product in process flow of various process routes.Finally,a prototype system was designed and developed in a spinning machinery sheet metal workshop,and the feasibility and effectiveness of the proposed method were verified.

In order to address the drawbacks of traditional ant colony algorithm in mobile robot path planning,such as blind search,slow convergence speed,multiple path turning points,it proposes a mobile robot path planning algorithm based on improved ant colony algorithm.Firstly,the Jump Point Search (JPS) algorithm is utilized to unevenly distribute initial pheromones,reducing the likelihood of blind search during the early stages of the ant colony.Then,a Chebyshev distance weighting factor and turning cost are introduced to improve the heuristic function,enhancing the algorithm′s convergence speed,global path optimization capability,and smoothness of the search path.Finally,a novel pheromone update strategy is proposed that introduces an adaptive reward-punishment factor to adaptively adjust the pheromone reward-punishment factor during pre-and post-iteration phases,ensuring the algorithm′s global optimal convergence.Experimental simulation results demonstrate that,in various map environments and compared to existing literature results,the proposed algorithm effectively reduces the number of iterations and optimal path length required for path search while increasing path smoothness.

In order to study the influence of rail cant coupling wheel diameter difference on subway vehicle dynamic performance and wheel-rail contact characteristics,three types of rail cant(1/20,1/30,1/40) and five types of wheel diameter difference conditions were set up.By establishing a subway vehicle dynamics simulation model and a three-dimensional elastic finite element model of wheel-rail contact,the influence of coupling wheel diameter difference of rail cant on linear stability and curve passing performance of subway vehicles was analyzed.The results show,under the straight section,the influence of wheel diameter difference on vehicle smoothness index is greater than that of rail cant and the wheel-rail equivalent stress and contact stress increase with the increase of wheel diameter difference.Each dynamic index of the curve section will increase with the increase of wheel diameter difference.Considering the dynamic index and wheel-rail contact characteristics,the use of 1/20 rail cant under the straight section is minimized.Under the curve section,1/30 rail cant can be selected to reduce the impact of wheel diameter difference on the vehicle.

Compared to passive suspension,active suspension has an additional device that can generate active force,effectively adjusting stiffness and damping in uneven road conditions,improving the shock absorption performance and riding comfort of the vehicle chassis. The wheelbase preview information is considered and the principle of Linear Matrix Inequality (LMI) is used for reference to establish the model of the half car active suspension under the action of pitching torque,and the traditional H₂/H_∞ state feedback controller and the H₂/H_∞ state feedback controller with preview information are designed. The multi-objective optimization control of the active suspension is carried out,and the performance of the two controllers is simulated in the time domain. The results indicate that both controllers can effectively optimize the performance parameters of the active suspension,but the H₂/H_∞ state feedback controller considering preview information has a more prominent optimization ability.

In order to improve the vibration isolation performance of exhaust system of a certain model,the finite element model of exhaust system was constructed with HyperMesh softwore and analyzed for vibration isolation performance. After testing and analysis, the transfer force is too large, the sum of the transfer force and the standard deviation of each ear were reset as the optimization target 1; and the sum of static displacement of each ear and the standard deviation of preload force were taken as the optimization objective 2. In HyperStudy softwore,the Z-direction stiffness of the six lifting lugs was finally determined by Latin hypercube sampling and DOE variable screening,and the response surface model was constructed,and the optimal lifting lug stiffness value was obtained by Multi-Objective Genetic Algorithm (MOGA). The best lifting lug stiffness value was brought into the finite element model of the exhaust system for further analysis,and the results show that the optimization target was improved. Finally,the actual production of lifting lugs was tested on the actual road after loading,and the test results show that the vibration isolation rate is greater than 20 dB, which is the same as the finite element result,which meets the production design requirements.

An accurate understanding of the driving environment is one of the prerequisites for autonomous driving.In order to improve the scene understanding ability of autonomous driving vehicles,an end-to-end autonomous driving model based on semantic segmentation,horizontal disparity,and angle coding multi-modal intermediate representations was proposed.The end-to-end autonomous driving model used deep learning technology to build perception-planning network.The perception network generated multi-modal intermediate representations with RGB and depth images as inputs to realize the spatial distribution description of road environment and surrounding obstacles.The planning network used multi-modal intermediate representations to extract road environment features and predict waypoints. Model training and performance testing were conducte based on the CARLA simulation platform.The results showed that the end-to-end autonomous driving model can realize the scene understanding of urban road environment and effectively reduce collisions.Compared with the baseline model based on the single modal intermediate representation,its driving performance index is 31.47 % better than the baseline model.

A direct yaw moment differential braking control strategy based on fuzzy Proportion、Integral、Differential (PID) control is proposed for the dangerous situation of double-trailer train under high-speed overtaking condition, such as lateral instability, trailer folding and tail-throwing. Two kinds of lateral stability control strategies for differential braking were designed, namely single-control mode controlled only by tractor and multi-control mode controlled by both tractor and trailer, aiming at the lateral deflection angle and yaw velocity deviation and deviation change rate of the TruckSim nonlinear model and the four-degree-of-freedom linear model respectively. The effectiveness of the control strategy was verified through the co-simulation of MATLAB/Simulink software and TruckSim software and the Hardware-in-the-Loop (HIL) platform. The results show that, in high-speed overtaking conditions, the lateral stability control system of multi-control vehicles is better able to reduce the lateral declination angle, yaw speed, articulated angle and Rearward Amplification (RWA) coefficient of the vehicle than that of uncontrolled vehicles and single-control vehicles, and is significantly better in improving the lateral stability of double-truck trains.

The problems of section warpage,burr and accelerated tool wear of aluminum foil during precision cutting cannot be solved qualitatively.Lemaitre shear damage model was established to accurately analyze the elastoplastic deformation and fracture evolution process of aluminum foil in shear.By analyzing the influence of different shear gaps on the quality of the shear section and the change of the stress state of the blade,the height of the shear band in the shear section and the stress of the blade were put forward as the optimization evaluation indexes of the aluminum foil shear.Finally,according to the analysis of simulation data and test results,the optimal shearing clearance is 30 %t,which meets the shearing quality.At the same time,the life of the cutting edge of an aluminum foil shearing machine is increased by 15.9 %.

There is machining errors in the leading and trailing edges of aero-engine blade after Electrochemical Machining (ECM).Electric Discharge Machining (EDM) technology was proposed to trim the leading and trailing edges of blade to reduce machining errors,and then the machining errors of leading and trialing edges were analyzed.The self-developed six-axis EDM machine was used to trim the trailing edges of blade by an electrode with curved surface.The point cloud data of blade surface was collected using a set of Coordinate Measuring Machine (CMM).The Genetic Algorithm (GA) and Iterative Closest Point (ICP) algorithm were used to register the blade point cloud data after envelope processing with the theoretical model of blade,and the least square ellipse and NURBS curve fitting were used to process the point cloud data.Aiming at the fitted blade contour curve after machining,the machining errors at the trailing edge of the blade and the overall trend of machining errors were reflected by taking equidistant points on the curve and calculating the distance between these points and the theoretical contour of the blade.The results show that the profile error of the trailing edge of the blade is between-16.4~88.5 μm,and the average error is 53.7 μm.The profile error at the edge head of the trailing edge of the blade is the smallest,which verifies the feasibility of the EDM technology to cover the leading and trailing edges of the blade.

In order to improve the processing efficiency of cylindrical gears and the bearing capacity of gear pairs,a new type of rapid prototyping method for plane helical cylindrical gears is proposed,that is,the gears are milled quickly by using a large rotating cutter head,multiple tool bars and blades are arranged according to the plane helical line to continuously cut the tooth groove of the gear,the pitch of the plane helix is equal to the pitch of the gear to be processed,the cutter head rotates one circle and the gear to be processed rotates one tooth,realizing continuous and rapid cutting of the gear processing,the feasibility of this method is verified through VERICUT processing simulation; a special processing machine tool is developed to implement the processing of new gears; the main parameters of the processed plane helical cylindrical gear are clarified,the tooth line radius of the concave tooth surface are slightly larger than the radius of the tooth line of the convex tooth surface,the teeth are slightly drum-shaped,which can reduce the meshing load of the gear pair. Through the three-dimensional meshing analysis of the plane helical cylindrical gear pair,it is concluded that the contact interval of the plane helical cylindrical gear pair is mainly distributed in the middle part of the contact tooth surface,and the force on both ends of the gear teeth gradually decreases,effectively reducing the bending stress at the tooth root at both ends of the gear teeth and improving the bearing capacity of the gear pair.

Regular inspection of drainage pipes can find serious defects in time,which is of great significance to ensure the healthy operation of the drainage system and the safety of the urban environment.Aiming at the difficulty of detecting low illumination and low resolution of lower drainage pipes,an improved drainage pipeline defect detection and geometric characterization method of YOLOv7 algorithm is proposed.Firstly,the Contrast-Limited Adaptive Histogram Equalization (CLAHE) image enhancement technique is used to improve the contrast and detail of the image,so as to improve the detection network′s ability to capture drainage pipe defects.Secondly,based on the design of Drop-CA and MC modules,the YOLOv7 algorithm is improved,so that the network can obtain the semantic information of shallow defects,reduce the false detection rate,and improve the classification and localization capabilities of the model.Finally,for the two serious defects of crack and fracture,a method is designed to quantitatively describe the geometric characteristics of the defect to evaluate the size of the defect.Experimental results show that the final average accuracy of the improved network model reaches 93.3 %,and the detection speed reaches 42.9 f/s. This method effectively improves the accuracy of defect detection and classification of drainage pipelines,and can effectively characterize the geometric characteristics of defects.

Aiming at the problem that the existing disassembly depth decision method cannot objectively and accurately obtain the optimal disassembly depth of waste smartphones,a disassembly depth optimization method for waste smartphones based on improved Teaching-Learning-Based Optimization (TLBO) algorithm was proposed. The disassembly profit model,disassembly time model and disassembly energy consumption model were established and the disassembly depth optimization target was determined,an optimal disassembly depth solution set was obtained by the improved TLBO algorithm. The disassembly process of Xiaomi 5 was selected as an example to verify the effectiveness of the proposed method. The result show that the optimal solution search ability and the convergence speed of the proposed method are enhanced.

A novel synchronous drive Permanent Magnet Claw Coupling(PMCC) based on magnet repulsion was proposed. Firstly,the structure of PMCC was introduced.Secondly,the structural design problems of PMCC were analyzed.Thirdly,the magnet circuits of PMCC were calculated by finite element simulation software. The influences of some important parameters such as the number of pole pairs,the thickness angle of permanent magnet and the angular gap on torque transfer characteristics of PMCC were analyzed respectively. Finally,the torque curves of PMCC with strong and weak magnetic combination were calculated. The results showed that the transmission torque of PMCC can be increased by increasing number of pole pairs or thickness angle of permanent magnet. When the initial angular gap was increased,the elastic buffering characteristics of PMCC were improved. In addition,when the design scheme of strong and weak magnetic combination is adopted,it can increase the transmission torque and the initial angular gap of PMCC, which provides an important reference for the development and design of PMCC and its application in mechanical transmission system.

In order to solve the problem of gearbox rolling bearing failure caused by frequent start-stop of hybrid electric vehicles in urban operation,the Maximum Correlation Kurtosis Deconvolution (MCKD) method was applied to the early fault diagnosis of gearbox bearing.In order to solve the problem that the filter order and shift number need to be selected manually in MCKD,a fault diagnosis method of maximum correlation kurtosis deconvolution with adaptive parameters was proposed.In this method,the maximum correlation kurtosis in the signal envelope spectrum was taken as the objective function,and the filter coefficient L and shift number M were optimized by the Improved Particle Swarm Optimization (IPSO) algorithm.Finally,the bearing fault characteristics were extracted by the envelope spectrum. Simulation and experimental results show that this method can effectively reduce environmental interference,accurately extract fault features from strong noise,and realize fault diagnosis.

Aiming at the problem of early respiratory crack location of beam structures,based on the difference of bilinear degree of different measuring points of beams with respiratory cracks,a bilinear frequency separation method for respiratory crack location of beam structures was proposed. By using the techniques of zero search and signal segmentation,the structural responses of different spatial positions were decomposed into positive and negative response components corresponding to bilinear effects,which reflect the dynamic behavior of the structure in the crack opening state and closed state,respectively. The difference of frequency components of split components at different measuring points is related to the location and depth of cracks,so the distance damage index of bilinear frequency components was constructed to realize the location of cracks. The numerical study of the influencing factors such as crack location,crack depth,excitation frequency and amplitude,natural structure shape and noise shows the effectiveness of the method and can provide a theoretical basis for practical engineering applications.

Predicting the Residual Useful Life(RUL) of bearings based on traditional methods has many steps,is expensive,and the model is not generalizable.And the existing deep learning-based prediction methods often lead to overfitting of the model due to the excessive amount of data,which results in poor model accuracy.In order to overcome the above drawbacks,a rolling bearing RUL prediction method based on Manifold Regularization Stack Denoising Auto Encoder (MRSDAE) combined with Self-Organizing Mapping (SOM) combined with HGRU was proposed. Firstly,a bearing Health Indicator (HI) was constructed using an unsupervised network MRSDAE-SOM. Then,a prediction model was built through a Hierarchical Gated Recurrent Unit (HGRU) network. The HGRU network has the ability to capture temporal features and integrate attention information from different time scales by adding multi-scale and dense layers. Finally,experimental validation shows that compared with other data-driven methods,the proposed method constructs health indicator in an unsupervised way,which is efficient and easy to apply. The model generalizes well and effectively prevents overfitting,and achieves higher prediction accuracy.

The friction characteristics of silver (Ag),polytetrafluoroethylene (PTFE) and molybdenum disulphide (MoS₂) coatings were tested under different contact stresses and different sliding speeds using a high-speed rotary friction and wear tester in an ultra-low temperature liquid nitrogen (LN₂) environment.The shape of the wear marks and the main components of the steel discs were analysed using an XM-200 surface morphometer,a JSM-7800F scanning electron microscope (SEM) and an energy spectrometer (EDS),and the friction factors and wear mechanisms of the three coatings were investigated under ultra-low temperature environment. The results show that the friction factor of the coated steel discs does not vary much under two speeds (high sliding speed of 2 m/s and low sliding speed of 0.4 m/s),while the friction factor of the uncoated steel discs is influenced by the sliding speed,with MoS₂ and PTFE having the best friction reduction effect.The wear of the three coatings increases with increasing load for the same operating conditions.The wear resistance of PTFE in the ultra-low temperature environment is poor and the wear amount is higher,while the wear amount of steel/MoS₂ coating is the smallest and steel/Ag coating is the second,indicating that both Ag and MoS₂ show good wear resistance in the ultra-low temperature environment and the wear amount is smaller.