Self-pierce riveting technology, as a new connection technology, has excellent connection performance in the connection of thin plate materials and has a wide application scenario in the aerospace industry. 5A06 aluminum alloy and TA1 titanium alloy thin plate were selected for self-piercing riveting, and the forming process was subjected to finite element analysis. The effects of different materials and temperatures on the static mechanical properties and failure forms of the joint were investigated by conducting tensile-shear tests at room temperature and -30 ℃ environment. The results showed that the homogeneous joints were formed with good symmetry, and the cavities were easy to appear in the riveted harder plates; the average peak load of the joints decreased at low-temperature environment, but the average failure displacement and the average energy absorption increased significantly, which improved the cushioning and shock absorption ability of the joints; the average peak load of 5A06-5A06 joints (SAA joints) decreased by 13.6 % at -30 ℃, and the average peak load of TA1-TA1 joints (STT joints) showed a 6.6 % decrease at -30 ℃, which indicates that STT joints have better low-temperature resistance; under both room-temperature and low-temperature environment, SAA joints failed by pull-off and plastic deformation occurred at the failure point, STT joints showed an increase in the number of rivet cracks at low temperature, with brittle fracture at the beginning of crack expansion and ductile fracture at the edge of expansion.

Aiming at the low output flow of the valveless piezoelectric pump with triangular prism bluff body,a valveless piezoelectric pump with arrow-shaped bluff body was designed based on the low flow resistance characteristic of the arrow structure,and the structure and working principle of this pump were introduced.The velocity streamline diagram of the valveless piezoelectric pump with arrow-shaped bluff body was simulated by finite element software,and the resistance characteristic of this pump was compared with that of the valveless piezoelectric pump with triangular prism bluff body.The results show that the valveless piezoelectric pump with arrow-shaped bluff body can transport fluids in one direction,and its efficiency is 19.51 % higher than that of the valveless piezoelectric pump with triangular prism bluff body.A 3D printer was used to produce pump prototypes,and the flow rate and pressure difference tests were carried out.The tests results show that under the driving voltage of 300 V,the maximum flow rate and pressure difference of the valveless piezoelectric pump with arrow-shaped bluff body are 17.40 g/min and 470 Pa,which are 22.71 % and 37.03 % higher than those of the valveless piezoelectric pump with triangular prism bluff body,respectively.

Aiming at the Flexible Job shop Batch Scheduling Problem (FJBSP) with the goal of minimizing the makespan, an Improved Grey Wolf Optimization (IGWO) algorithm was proposed to optimize the job batching scheme. For the first time, the fluid model was applied to solve FJBSP through proposing a decoding method based on fluid model to get a better batch scheduling scheme. Secondly, the hierarchical system of the wolf pack was improved to avoid premature convergence of the algorithm. Thirdly, a new crossover method that adapts to the variable length code was designed to deeply exchange the batch information between two individuals, and it can enhance the search ability and stability of algorithm. Then, a self-adaptive grey wolf wandering strategy was proposed to dynamically update the individual′s wandering rate,and it can balance the search quality and convergence speed of algorithm. In addition, the self-adaptive neighborhood search was used on the leader wolves to dynamically adjust the selection probability of each neighborhoods for each kind of job and improve the local search ability of the algorithm. Finally, nine examples and three sets of experiments in total were designed to verify the effectiveness and superiority of the proposed IGWO algorithm.

For the disadvantages of low efficiency and difficulty in quality control of the main drive assembly process in shield machines, a main drive assembly production line integrating efficient assembly and data transmission was developed. Based on the overall design scheme of the main drive assembly production line of shield machine, a research was conducted on key technologies such as automated cleaning of main drive parts, material automated transportation, automated assembly of components, real-time detection and storage of tightening information for fasteners, and intelligent scheduling systems. Highly integrated automated equipment replaced traditional manual assembly processes, improving the quality and efficiency of main drive assembly while achieving the management and traceability of assembly information, forming the core technology system of the main drive assembly production line. The designed main drive assembly production line of shield machine was successfully built and the actual process verification was completed, achieving the design effect. It has now been successfully applied to the assembly production of various main drive specifications, achieving significant economic benefits.

Based on the characteristics of distributed flexible job-shop problems,a mathematical model was constructed with the objective functions of minimizing completion time,factory energy consumption,and equipment load,and the three objectives were normalized using linear weighted sum method.On the basis of the traditional Seagull Optimization Algorithm (SOA),an adaptive additional variable update strategy has been improved to improve the local optimization ability and convergence accuracy of the algorithm in the later stage. The flight mechanism in the sparrow algorithm was integrated to expand the local optimization range of individuals,and further improve the optimization accuracy.Introducing a variable neighborhood search algorithm for key factories has expanded the neighborhood search range and enhanced the local search ability of the Seagull algorithm. The effectiveness and feasibility of ISOA in solving multi-objective distributed and flexible job-shop problems were verified through standard and actual factory examples.

Aiming at the problem of aviation cable path planning,a joint point routing algorithm with non-rasterization processing was proposed. The concept of ″baseline″ was introduced. Firstly,the initial joint point set was obtained by baseline and routing environment information,and the path planning problem was transformed into the optimal selection problem of joint point set. Then the redundant joint points were screened and optimized according to the obstacle information. Then,the set of joint points with redundant points removed was iteratively optimized,and joint points were added and deleted under the convergence condition of optimal path. Finally,the joint point was used as the routing path node to obtain the cable path. The experimental results show that the joint point routing algorithm can effectively reduce the bending times and cable length of cable routing,and the routing time is better.

The emergence of mobile charging robots has solved the problem of charging electric vehicles in old parking lots, but the complex and random obstacle environment in parking lots puts forward higher requirements for their path planning and obstacle avoidance functions. The path planning problem of an electric vehicle mobile charging robot in a parking lot was simulated and analyzed by improving the traditional gray wolf optimization algorithm. The gray wolf optimization algorithm has fast iteration speed, but low optimization accuracy, and is prone to falling into local optima. After rasterizing the parking lot map, the traditional gray wolf optimization algorithm was improved from the perspective of fitness function, convergence factor and location update function. Simulation using MATLAB showed that the average iteration number of the improved gray wolf optimization algorithm was reduced by 39.4 % compared to the traditional gray wolf optimization algorithm, and the path length was reduced by 4.7 %, the path length was the shortest compared with other typical improved gray wolf optimization algorithm. At the same time, the path of mobile charging robots with different occupancy rates in the same parking lot was simulated using this algorithm, and the results showed that the algorithm can successfully run under random maps and different target locations, verifying the stability of the algorithm.

To solve the problems of inaccurate modeling of the end-effector pneumatic system and poor stability of grinding force caused by external environment interference during grinding of aluminium alloy wheel hub,a hybrid control strategy of MRAC+PID was proposed to achieve tracking control of grinding force.Firstly,the end-effector was modeled and analyzed.Then,the MRAC+PID control strategy was used to control the end-effector,in which the adaptive control law was designed by constructing the Lyapunov function.Finally,the effectiveness of the proposed method was verified by simulation and experiment.The results show that MRAC+PID has good control performance in improving the response speed and anti-interference of the system,which can track the desired grinding force quickly and accurately and improve the surface grinding quality of aluminum alloy wheel hub.

In order to improve the performance of the robot in the milling application,the screw theory was applied to the modeling of the milling robot.In order to solve the problem that the calculation of the inverse solution is too slow in the simulation process,an improved geometric method was proposed to optimize the inversion process.According to the geometric structure of the KUKA KR60 milling robot,the method considers that the milling position is relatively fixed,and ignores the inverse solution changes caused by the first three joints of the robot.The simulation results show that the improved geometric method can greatly improve the computing efficiency and reduce the simulation time compared with the inverse method of MATLAB.

To overcome the influence of uncertain factors such as mechanical friction,unknown loads,and model errors on the control accuracy of manipulator,a nonsingular fast integrating terminal sliding mode finite time robust control law was designed using exponential obstacle Lyapunov function.Firstly,a mathematical model of the manipulator with uncertain factors was established,and the interference value was accurately estimated by designing an observer. Then,the terminal sliding surface of nonsingular fast integration was designed by using the angle error of the manipulator,and the finite time robust control law was designed based on the exponential obstacle Lyapunov function. Finally,the stability analysis also verified that the designed finite time robust control law could converge within the effective time. The simulation results of a 3-DOF manipulator show that the proposed finite time robust control law can effectively overcome the influence of uncertainties factors on the control accuracy of the manipulator,the maximum estimation error of the observer is 0.1 (°)/s²,and the maximum tracking error of each joint angle of the manipulator is 0.1°,which verifies the effectiveness of the design method.In the 3-D space fixed coordinate positioning experiment,under the proposed finite time robust control law,the maximum positioning error of the manipulator is 0.22 cm,and the maximum running time is 2.2 s,which verifies that higher control accuracy and running efficiency can be achieved in practical engineering applications.

In recent years,with the rapid development of intelligent transportation system,including vehicle-vehicle communication,vehicle-road communication and other short-range real-time wireless communication information,as well as road traffic information and other long-distance traffic information,so that the vehicle is able to obtain real-time knowledge of the surrounding vehicle movement as well as the traffic environment in front of it,which is conducive to improving the vehicle′s perception of the surrounding traffic environment in order to achieve reasonable travel arrangements and driving control,thus improving the vehicle performance. In order to realize the energy-saving driving of fleet in multi-signal light scenarios,a co-optimization method based on reinforcement learning for vehicle queuq speed planning and energy management was proposed.Through the SUMO platform,a scenario including five vehicles for a fleet of vehicles passing through multiple signals was established.The results show that the proposed method outperforms the traditional driving model Intelligent Driver Model (IDM) in terms of comfort,economy and efficiency.

In order to improve the stability and safety of the vehicle during steering,an active steering controller considering the yaw angle velocity and the centroid side declination angle was designed on the basis of the feed-forward control of the ideal transmission ratio.According to the error between the actual value and the ideal value of the state parameter,the active steering controller calculates an additional additional turning angle independent of the driver through the adaptive sliding mode controller to compensate for the front wheel rotation angle,so that the actual response of the car follows the ideal value to improve the stability of the car.Finally,a co-simulation model of steer-by-wire was established in Matlab/Simulink software and Carsim software,and the proposed active steering control strategy was tested under the conditions of double line shift,sinusoidal input,high and low road adhesion coefficients,respectively.The results show that the active steering controller can significantly reduce the error between yaw angular velocity and centroid deflection angle and the ideal value,and improve the stability of the vehicle during steering.

Aiming at the common problems and key technologies in the forward development of torsion beams, using numerical simulation methods, the comparative studies were conducted on two types of torsion beams, namely closed tube beams and open plate beams, including key technologies such as torsion beam scheme design, DMU verification, product and process CAE verification, torsion beam deformation numerical simulation, and deformation compensation, providing a certain reference for the forward development of torsion beams by host manufacturers.The main research were as follows. 1) According to the design input, it was necessary to design the beam position and cross-section firstly, with the focus on determining the shear center of the beam. From the numerical simulation results, it can be seen that different beam cross-section designs have a significant impact on the Z-direction distance of the shear center. 2) In checking the kinematic clearance of the torsion beam suspension, compared to the parallel wheel jump condition, the reverse wheel jump condition is harsh, which can easily lead to the risk of interference and wear of the tires and wheel housings. 3) Using numerical simulation technology, the torsional stiffness of two types of torsion beam structures was studied. Compared with open plate beams, closed tube beams have better lightweight effects. Using numerical simulation technology, the external stresses of two types of torsion beam structures under torsion conditions and the internal stresses during the forming process of the two types of beams were studied. If the internal and external stresses at the same position of the beams were superimposed, there would inevitably be a greater risk of fatigue cracking. Solutions should be provided from the perspective of materials, structure, and technology. 4) The deformation of the torsion beam under the action of the vehicle′s offline spring load can have a significant impact on the toe and camber of the torsion beam, and the equivalent toe and camber of the torsion beam should be provided with reverse mechanical angle compensation.

Aiming at the problem of poor point cloud segmentation effect in the grinding process of aviation engine blades and the large amount of calculation caused by surface fitting in trajectory planning, a point cloud based method for aircraft engine blade segmentation and grinding trajectory planning was proposed, aiming to improve point cloud segmentation effect and reduce the amount of calculation. The point cloud of the blade to be polished was collected by the structured light camera, pre-processed such as filtering was carried out, and the Random Sample Consensus (RANSAC) plane segmentation algorithm was improved to improve the efficiency of the algorithm. A region growing segmentation algorithm combined with point cloud color information was designed to achieve stable segmentation of the blade area to be polished. A triple section method directly based on point cloud was proposed, combined with B-spline curve fitting algorithm to generate pre-grinding trajectory, and based on equal residual height algorithm to optimize line spacing, complete grinding trajectory planning. It has been verified by experiments that the efficiency of the plane segmentation algorithm is 40.25 % higher than that of the traditional method, and the data processing efficiency is greatly improved. In addition, through the improved region growing segmentation algorithm, the efficiency of blade grinding trajectory planning is significantly improved, and its realization effect is significantly enhanced.

The position of the turbine disc connecting blade of the steam turbine is usually a mortise groove structure. For small dovetail type mortise groove,a molding tool is often used for one-time processing,which leads to low precision of machined parts,poor surface roughness and high tool cost. In order to solve the limitation of the present manufacturing process,a design idea of combined cutting tools was proposed. Firstly,according to the structural characteristics of dovetail type mortise groove,the tool shape line and size parameters of four knives were designed.The model was built by UG modeling software,and the finite element analysis was carried out by ANSYS software. Then the manufacturing process of the combination tool was developed,and the path planning and programming were carried out by NUMROTO tool programming software.Finally,the qualified combined tool was grinded on the SAACKE five-axis tool grinder. Through simulation and practical verification,this scheme is feasible.Compared with traditional machining methods,it has the advantages of high machining accuracy, low surface roughness,low cutting heat and low cost.

Thin-disk parts are common parts on rotor mechanism,which usually need to be tested and calibrated for dynamic balance before assembly and commissioning.At present,the commonly used dynamic balancing calibration methods have problems such as the separation of processing and calibration,complex balancing process and low efficiency.Based on these problems,a method is proposed to correct the unevenness during the machining of thin-disk workpieces,and combine the spindle quasi-stop,tool positioning and de-weighting control functions of the composite CNC machine tool to perform online de-weighting correction of thin-disk workpieces.Meanwhile,in order to improve the correcting accuracy,an iterative control method with weight moving rate is proposed,and its stability is proved theoretically.The advantages of iterative control method with weight moving rate are discussed by simulation analysis,and the ideal weight moving rate selection method is discussed. The research content can realize online de-weighting correction of thin-disk workpieces unbalance without dynamic balancing machine,and the balance efficiency and accuracy can be improved,the balance cost can be reduced.

A new measurement method of sprocket end face geometry parameters based on vision principle was proposed aiming at the problems of low efficiency and measurement accuracy in the existing measurement methods of sprocket end face geometry parameters. Firstly,the inner and outer contours of the sprocket were extracted by sub-pixel edge detection of Zernike moments,and Hough transformation was used to obtain the center point coordinates and center hole radius of the inner contour. Secondly,an auxiliary circle was made on the outer contour,and the number of teeth was obtained by the pixel sequence counting algorithm. Then,according to the shape characteristics of the sprocket tooth profile,an improved local radius region algorithm was proposed to extract the tip and root points of each tooth profile,and the least square method was used to fit the pixel points to obtain the tooth apex and tooth root circle diameter. Finally,the least square method was used to fit the sprocket keyway edge line to obtain the keyway size.The repeated measurements were performed 10 times for each of the two sprockets with different specifications. The experimental results show that the measurement errors of sprocket end face geometric parameters are all less than 40 μm,and the relative errors are all within 0.4 %,which can meet the needs of high-precision automatic measurement.

Due to lightweight,miniaturization and other needs of head-mounted smart glasses,polymer materials are gradually introduced into the application,but due to factors such as mold accuracy and material shrinkage during the injection process,it is easy to cause the problem of mismatch between the 3D model of the spare parts and the real thing,which causes the size of the five-axis laser linkage cutting glasses parts to be out of line. In order to solve this problem, the dimensional error between the three-dimensional model of injection products and the real objects is analyzed with the help of ATOS optical scanner. Through reverse engineering modeling of injection product parts,it verifies the feasibility of extracting the tool path trajectory for five-axis laser cutting on the three-dimensional model generated by scanning, which provides a solution to improve the precision of 5-axis laser cutting for products that do not match the three-dimensional model and the injection molding size.

The flexspline is a vulnerable part of the harmonic reducer. Driven by the high rotation speed of the wave generator, its fatigue life has always been a research focus of concern. Takes flexible wheel of a cup-type harmonic reducer as the research object, establishes a finite element simulation model, the relationship between its maximum stress and the parameters such as the length of the cylinder, the thickness of the cup wall and the radius of different transition fillets were obtained. Based on the S-N curve of the flexspline, the fatigue life of the flexspline was calculated, and the prediction of fatigue life was achieved using BP neural network.

In order to solve the problems of low accuracy of tubing coupling cutting and high scrap rate in the production of tubing couplings, each equipment works isolately under a low automation situation, the precise feeding unit, locking cutting unit and automatic unloading unit were optimized on the basis of the traditional steel pipe cutting equipment. The hardware and software design of the tubing coupling automatic production control system was carried out based on S7-1500 PLC, and the control logic was designed by TIA Portal, and the working status of the production line was monitored in real time by Siemens KTP1200 human-machine interaction panel to realize the coordinated control of the machine. The experiments show that the designed tubing coupling automatic production line has a high degree of automation and operates stably, which significantly improves the production efficiency and automation of tubing coupling.

In order to improve the structural performance of H-type vertical axis wind turbine blades under the combined loads of aerodynamic force,centrifugal force and gravity,a multi-objective optimization method was proposed.A mathematical optimization model was established with the minimum mass and maximum stress as the objective function and the maximum deformation as the constraint.Through the Fluid-Structure-Interaction (FSI) method,the real-time and accurate extraction of the blade surface pressure was realized,and the finite element model of the blade under the combined load was established.Based on Optimal Space-Filling (OSF) method and Kriging model,the response surface model of each variable to stress,mass and deformation was established,and the sensitivity and change trend were analyzed.Finally,the Multi-Objective Genetic Algorithm (MOGA) was used to obtain the optimal solution of each variable,and the results were verified.The results showed that the optimized blade mass was reduced by 14.7 %,the maximum reduction of the maximum stress at each azimuth was 7.8 %,and the maximum reduction of the maximum deformation was 16.7 %.The research results can provide a reference for the structural optimization design of blades under combined loads.