Aiming at the problem of motion planning for the efficiency and safety of multi joint manipulator, the motion planning based on the open source robot operation system (ROS) is studied. The robot arm model is built by SolidWorks, and the robot URDF file described in XML format is generated by SW2URDF. The setting assistant Moveit!Setup Assistant in ROS is used to complete the configuration of motion planning related files. In the aspect of trajectory planning, by using the C++ interface under the framework of moveit!, the interpolation of cartesian space straight line and arc is realized under the 3D visualization platform Rviz. In the aspect of path planning, the algorithm of RRT (rapidly expanding random tree) and RRTConnect is used to complete the collision free path planning of high-dimensional space and complex constraints. The simulation results show that RRTConnect algorithm has fast convergence speed and strong search ability, which can effectively solve the path planning problem in complex environment and ensure the motion safety.

In view of the large assembly error and high repair rate in the production process of the aircraft wing assembly workshop, the application of digital twin technology in the assembly process was studied, and the quality control method of the assembly process based on digital twin was proposed. Various digital assembly and inspection equipment collect and analyze assembly data in real time. Based on the research and analysis of historical assembly quality data, the correlation model between quality control points is constructed through Copula theory, and then the assembly quality entropy and mutual information are used. Theoretically build a product assembly stability measurement model, optimize the control threshold of assembly quality control points, predict and give the next optimal assembly plan, and transfer the best assembly plan to the physical workshop for actual assembly. Finally, the prototype system is verified by an example of wing box assembly in a workshop, which proves the feasibility of the whole method.

In view of the disadvantages of traditional PMSM closed-loop control system, such as large volume, high cost and low reliability, a sensorless closed-loop control system and on-line identification method are proposed. Based on the two-phase static coordinate system model of PMSM, a sliding mode current observer is designed to estimate the position and angular velocity of PMSM. Considering that the change of internal environment will lead to the change of motor parameters in the process of motor operation, an on-line identification of motor parameters based on genetic algorithm is designed. The simulation results show that the system can quickly and accurately predict the position angle and speed of the motor, and has high identification accuracy for the stator resistance and inductance of the motor. It can meet the needs of sensorless control of PMSM.

During the operation of the robot, in order to make the servo motor achieve the target speed under the optimal performance, and have stronger anti-disturbance ability in the working process, and to avoid the problem of vibration and resonance, resulting in a serious reduction in the dynamic stability of the robot. The control model of servo motor is analyzed, and a particle swarm optimization algorithm based on nonlinear dynamic learning factor is proposed. The algorithm takes the speed controller in the servo system control model as the core, and can identify the load's moment of inertia in real time, so that the internal control parameters of the servo system can be adjusted according to the actual conditions. By using the identification value, the PI parameter value of the speed control is obtained through calculation, and the PI parameter value of the speed controller is corrected in real time. The results of MATLAB/SIMULINK simulation show that compared with the traditional particle swarm optimization algorithm, this method has faster response speed, higher control accuracy and stronger anti-interference ability, whether in the motor starting process or under the load disturbance.

To improve the working performance of the robot and meet the needs of actual complex working conditions, the motion state and trajectory planning of the SNR3-C30 6R industrial robot were studied. First, the coordinate system and kinematics equations of the robot arm link are established according to the D-H method. Secondly, based on the posture matrix characteristics of its end effector, a geometric inversion optimization algorithm combined with posture separation is proposed. This algorithm uses the geometric positional relationship of each joint of the robot relative to the wrist to solve the first three joint angles θ₁, θ₂, θ₃, and then on the basis of this, the remaining joint angles θ₄, θ₅, θ₆ are solved by rotating the sub-matrix. Finally, the improved algorithm is verified in the MATLAB Robotics simulation module, and the polynomial interpolation method is used to plan the joint space trajectory. The research results show that the improved algorithm has superiority and effectiveness, and when the fifth-order polynomial interpolation method is used for trajectory planning, the robot runs more smoothly.

Rolling bearing fault diagnosis is an important technology in the development of modern industry. Aiming at the problem of feature extraction and intelligent diagnosis of rolling bearing signals, this paper proposes a fault diagnosis method for rolling bearing based on WPD-CNN two-dimensional time-frequency images. First, the signal is converted into a two-dimensional time-frequency image by wavelet packet decomposition (WPD). Second, the time-frequency image is input into the VGG19 Convolutional Neural Network (CNN) model to automatically extract effective features, and input into the Softmax classifier for training. Finally, use the trained classifier to complete the fault diagnosis task of rolling bearings. The experimental results show that the recognition accuracy of the 10 types of fault data is about 98.3%, which is higher than other deep learning and traditional methods. Therefore, the proposed fault diagnosis model can effectively perform the feature extraction and classification tasks of complex signals of rolling bearings.

The known data fitting based on moving least squares (MLS) has a large lifting space, this paper proposes an improved moving least squares (IMLS). Through the analysis and comparison of the normal weighting function and two typical weighting functions, the superior fitting performance of the normal weighting is proved. The moving least square method is improved by taking the normal weighting as the weighting function, and the influence of its influence radius and shape parameter on the fitting performance is deeply studied, numerical examples of curve fitting and surface fitting are given, numerical examples and measurement data processing show that the improved moving least square method is superior to the moving least square method in curve fitting and surface construction.

In order to solve the problem that the instantaneous frequency is difficult to estimate in the order analysis method based on the instantaneous frequency estimation, a order analysis method based on the synchronous extraction transform (SET) is proposed to analyze the variable speed vibration signal of the bearing. In this method, the instantaneous frequency of bearing vibration signal in strong noise environment is extracted by using the good energy concentration characteristic of set. The time domain signal is sampled equiangular by phase discrimination time scale of calculating instantaneous frequency, so as to obtain the steady angular domain signal. By analyzing the variable speed vibration signal of the bearing experimental platform, it shows that this method can effectively suppress the noise interference, accurately extract the instantaneous frequency of the variable speed signal, and successfully realize the fault diagnosis of the bearing in the variable speed state.

The paper aims to investigate the influence of ultrasonic vibration-assisted on the cutting performance of single-crystal silicon, to reveal the mechanism of substrate removal in ultrasonic vibration-assisted cutting.Based on Tersoff and Morse potential, the ultrasonic vibration-assisted cutting behavior with different vibration frequencies and amplitudes was studied by molecular dynamics.According to the change of atom motion track, tool force, substrate potential energy, subsurface damage, von Mises stress and number of substrate atoms removed, analyzed the cutting performance and removal mechanism of ultrasonic vibration-assisted cutting on single-crystal silicon.The results show that traditional cutting removes atoms by friction pushing, ultrasonic vibration-assisted cutting removes atoms by friction stripping.The increase of vibration frequency and amplitude is beneficial to reduce the cutting force and normal force of tool, increase the work of tool, reduce the thickness of the noncrystal layer, and increase the number of substrate atoms removed, ultrasonic vibration improves the cutting performance and efficiency of single-crystal silicon.The results of this study can provide theoretical guidance for high-efficient and high-quality processing of single-crystal silicon.

A hybrid force/position control method based on neural network is proposed to solve the problems of trajectory tracking and constant force control of the end-effector of the head grinding robot and vibration suppression of the head grinding robot. The force/bit hybrid control method is adopted to complete the requirements of grinding machine manpower and bit simultaneous control. A neural network robust controller is designed for the vibration interference of the polishing robot platform during its walking process, and the system parameter error and vibration interference are suppressed by the neural network. The simulation results show that the hybrid controller can not only ensure the precise control of the position of the end-effector and the effective constant force control, but also has a good suppression effect on the external vibration interference.

Dynamic model based control is an effective method to improve the performance of robot motion control.In order to obtain an accurate dynamic model, it is necessary to obtain accurate dynamic parameters of the robot.The identification algorithm combining weighted least squares method and genetic particle swarm optimization algorithm is used to obtain accurate dynamic parameters.Joint friction affects the accuracy of dynamic model.The accuracy of traditional Coulomb viscous friction model is not high.A new continuous friction model is introduced.The Newton Euler method is used to establish the robot dynamic model.Collect data of robot under excitation track motion.First, the weighted least square method is used to identify the initial solution.On the basis of the initial solution, the boundary of the solution is set.The genetic algorithm Particle swarm optimization algorithm and genetic particle swarm optimization hybrid algorithm are used to identify dynamic parameters, and compared with existing methods, the results show that genetic particle swarm optimization hybrid algorithm has higher accuracy in identifying dynamic parameters.Finally, the accuracy of the dynamic parameters is verified by selecting the verification trajectory.The results show that the identified dynamic parameters can establish an accurate dynamic model.

Unmanned aerial vehicle (UAV) inspection has become an important means for daily operation and routine maintenance of overhead transmission lines. When the inspection system is in contact operation, the high voltage, high current, and strong electromagnetic environment are easy to cause damage to the system. Based on the actual parameters of typical 35 kV AC transmission lines and intermediate towers, this paper proposes a design scheme of the UAV inspection system based on mechanical arm. Moreover, the finite element model of UAV inspection system was established. The influence of electric field around the transmission line on the designed system was mainly studied. Finally, a test platform of UAV inspection system for 35 kV transmission line was built to carry out pressure test on the system. The results of simulation and test show that the working parameters of the prototype system adapt the design requirements. What's more, it can fit the technical requirements for the inspection of 35 kV high-voltage overhead transmission lines. The research results can provide technical supports for the design and protection of relevant components of the multi-rotor UAV inspection system on 35 kV transmission lines.

Aiming at the problem of poor decoupling effect of current loop caused by low carrier ratio and large digital control delay of high-speed permanent magnet synchronous motor, an improved deviation decoupling control strategy is proposed based on traditional deviation decoupling control.Firstly, the control sequence of current loop is analyzed to study the influence of angle delay caused by digital control delay.Secondly, on the basis of considering the additional coupling term between the shafts caused by the delay effect and the coupling term of the motor itself, according to the principle of traditional deviation decoupling control, the current control equation with coupling term is established, and a new decoupling term is derived, which is used as the system compensation to offset the mutual influence of the coupling term.The proposed control strategy is verified by building a three-phase high-speed permanent magnet synchronous motor simulation model, Compared with the traditional deviation decoupling control strategy, under the load mutation condition, the q-axis current fluctuation decreases by 54.6%~62.5%, and the d-axis current fluctuation decreases by 43.9%~60.3%, The three-phase current waveform is significantly improved, which improves the control performance of the system.

Taking the tandem robotic arm that does not satisfy Pieper′s criterion as the main object of study, an improved LM algorithm is proposed for the inverse solution of various types of tandem robotic arms with low accuracy, slow convergence speed and poor generality of the inverse kinematics solution.The robot motion model is established according to the improved D-H parameter method, and the objective function of the inverse solution problem is constructed with the goal of minimizing the positional error.An improved LM method is obtained by three strategies, such as nonlinear update of convergence factor, adjustment of convergence step by trust domain method and extension of iterative initial value, and it is used for the inverse kinematics solution.Finally, simulation tests are conducted in MATLAB environment, and the results show that the improved LM method has the characteristics of fast convergence, good stability performance and high accuracy, which proves the feasibility and effectiveness of the algorithm for the inverse.

Aiming at the problem that it is difficult to extract the fault features of the main reducer gear of shield machine, a fault diagnosis method based on parameter optimization variational mode decomposition (VMD) and kurtosis criterion is proposed.Particle swarm optimization (PSO) is used to determine the number of modal components K and penalty factor of VMD algorithm with minimum average envelope entropy as fitness function α the best combination of; Use the original vibration signal[K, α] the parameter combination is decomposed into several intrinsic mode function (IMF) components by VMD, and the best and second best IMF components are selected according to the kurtosis criterion for reconstruction to reduce noise interference; Finally, the time-domain and entropy theory features of the reconstructed signal are extracted, and the PSO-SVM and LightGBM are used for fault identification and classification.The analysis of the measured signal data of DDS test bench shows that the method can effectively extract the gear fault features and accurately determine the type of gear fault, which verifies the feasibility and effectiveness of the method.

In order to solve the problem that the enforceability of the initial scheduling scheme is reduced due to uncertain machine failure in the production process, a flexible job shop rescheduling method considering machine failure probability is proposed.The mixed driving strategy of event and cycle based on failure probability is adopted, a variety of rescheduling modes are used comprehensively, and the maximum completion time deviation, process start time cumulative deviation and process machine change are introduced as evaluation indexes.the scheduling scheme is solved by using the improved genetic algorithm combined with variable neighborhood search.Through the simulation analysis of the processed flexible job shop, the results show that compared with the single rescheduling method, the optimal rescheduling scheme has a great improvement in the performance evaluation index.the superiority of the proposed rescheduling method to deal with machine fault disturbance under different conditions is verified.

A path planning method based on the improved Hybrid A^* algorithm is proposed to address the efficiency of the path planning of mobile robots and the safety of the planned paths, which is grounded on the Ackerman six-wheel steering model.The number of node searches is reduced by improving the heuristic function in the Hybrid A^* algorithm and introducing a distance penalty function; guiding mobile robots as far away from obstacles as possible by constructing safe corridors; penalty terms for node forward, commutation and backward expansion are added to the cost function to ensure the enforceability and safety of the planned paths.The simulation shows that the path planning method based on the improved Hybrid A^* algorithm is applicable to the Ackerman six-wheeled mobile robot, which improves the efficiency of path planning and the planned paths are more secure and safe.

In order to ensure the consistency and stability of RV-80E reducer batch assembly process, the key technology for quality control of RV-80E reducer assembly process was studied and the assembly line based on machine vision was developed.Eight key matching methods of core parts and three complex dimension chains affecting the performance of machine were put forward by analyzing the principle of over-positioning structure of RV-80E reducer.The precision assembly process and quality control parameters of RV-80E were summarized based on the principle of achieving optimal overall performance.The RV-80E assembly line and machine vision system was designed and developed, the assembly line covers the key assembly processes of RV-80E, such as intelligent match, bushing precision pressing, needle roller installation, screw automatic tightening and so on, and the visual error-proof technology based on VisionPro was introduced in the key stations of the intelligent match and bushing precision pressing.Statistical analysis of 600 RV-80E reducers showed that the transmission error index conformed to a normal distribution with a Cpk index of 1.45.The production data showed that the assembly line based on machine vision technology improved the consistency and stability of the RV-80E assembly process.

To improve the working efficiency and stability of the manipulator, a time-optimal 3-5-3 polynomial interpolation trajectory planning algorithm based on the improved particle swarm optimization algorithm was proposed.The local and global search capability of the algorithm is optimized by dynamically adjusting the inertia weight and learning factor.Taking the UR5e robot as the research object, the joint position sequence corresponding to the path point at the end of a given manipulator was obtained through inverse kinematics, and the motion trajectory in joint space is fitted by piecewise polynomial.Then, under the velocity constraint, the improved particle swarm optimization algorithm is used to optimize the trajectory of the manipulator.Finally, the simulation and experimental analysis are carried out by MATLAB software.The results show that the improved algorithm has significantly improved convergence speed and optimization accuracy, and effectively reduces the time for the manipulator to complete the target task.

In order to realize the dynamic and static performance testing function of different spindles under simulated actual machining conditions, the performance evaluation indexes were obtained by analyzing the static and dynamic models of the spindles, and the V-block-like clamping mechanism was designed to realize the clamping of different spindles.The non-contact air-floating loading device was designed to realize static loading, which has the outstanding advantages of simple structure, small heat generation and stable load capacity compared with the commonly used drag loading and electromagnetic loading.The dynamic loading by milling process is closer to the actual working conditions.The selection of suitable sensors and dynamic test analyzer to achieve the collection, storage and timely analysis of radial displacement, dynamic and static load and speed data of the front end of the electric spindle.The key components of the experimental device were analyzed and calibrated, and the maximum linear displacement was 0.015 mm, which brought negligible error to the test system.The conclusion obtained verifies the reliability of the operation of the experimental device and meets the requirements of dynamic and static performance testing of machine tool electric spindles.

Aiming at the fact that the manipulator cannot completely establish an accurate mathematical model in practical work and the system has great uncertainty, an adaptive sliding mode robust controller based on RBF neural network compensation is proposed in this paper.First, using the computational torque control method, combined with the compensation controller based on the RBF neural network, the system can obtain a better tracking effect.Second, the sliding mode robust control is introduced, and the modeling error and external disturbance are adjusted through online training and parameter adjustment.The unknown uncertainty part of the manipulator is estimated and compensated in real time, which improves the tracking accuracy and the approximation speed.At the same time, the introduction of the saturation function reduces the chattering phenomenon in the sliding mode control.The simulation results show that the improved controller enhances the response speed of tracking manipulator joint position by about 4 s and 2 s respectively compared with the traditional controllers 2 and 3.Simultaneously, the improved controller accelerates the compensation speed of the system uncertainty and responds within about 1 s.

In order to explore the influence of amplitude, frequency and cutting speed on cutting force in elliptical vibration turning (EVT)316L stainless steel, a cutting simulation model was established by using finite element software ABAQUS, and the cutting processes under different amplitude, frequency and cutting speed were compared and studied. The results show that compared with conventional turning of 316L stainless steel, EVT can reduce Mises stress, cutting force and cutting temperature in the cutting process, which is an effective way to improve the processing of 316L stainless steel. Under EVT mode, the cutting force decreases with the increase of amplitude and frequency, and increases with the increase of cutting speed. Amplitude, frequency and cutting speed have more significant influence on the cutting force in the cutting speed direction.

To solve the problems of slow convergence and overestimation in deep reinforcement learning, an improved fusion method of deep reinforcement learning was proposed.The force in the artificial potential field was introduced as the reward value of deep reinforcement learning to guide the robot to avoid obstacles and move to the target, which improved the convergence of deep reinforcement learning network.Combining averaged DQN and double DQN method DDQN, the averaged DQN was used in the target network of DDQN to solve the overestimation problem.The low-level features were taken as input of the deep network to further improve the learning effect.Compared with other methods, the results of experiments show that the improved method can improve the convergence of the training network and solve the overestimation problem better.Using the improved method, the robot gets larger reward values and better planned paths.

In order to solve the problems of the traditional A^* algorithm, such as long path search time, too many inflection points and turns, unsafe search path, and uneven search path, an improved A^* and TEB integrated path planning algorithm is proposed.First, change the distance evaluation method of the heuristic function to improve the search calculation efficiency, and introduce the steering angle cost function to further optimize the evaluation function; When expanding nodes, automatic switching search neighborhood is enabled for safe expansion, and key nodes in the path are extracted to remove redundant nodes; Finally, TEB algorithm is used for local path planning of B-spline optimized path.The experimental results show that the fusion algorithm can effectively improve the planning efficiency, make the path more secure and smooth, and can successfully avoid dynamic obstacles.

Aiming at the problems of the ant colony algorithm in the path planning of mobile robot, such as the lack of stability, the slow convergence speed, the poor foresight and the tendency to fall into the local optimum, the traditional ant colony algorithm was optimized and improved. The transfer rule is optimized by improving the transfer probability so that the ants can accurately search for the next best grid position. Using the new heuristic information based on the infinite step principle to expand the field of vision and improve the visibility accuracy. In addition, new pheromone update rules are used to speed up the convergence and enlarge the search space. Simulation results show that the improved ant colony algorithm is superior to the traditional algorithm in path search efficiency and convergence speed.

The paper introduces a dedicated machine tool developed to effectively solve the problem of excessive material generated on both sides of mains bearing base on the traditional 4-cylinder, 6-cylinder and multiple cylinder engine in casting process. The use of this machine can prevent the interference resulting from the crankshaft rotating with the connecting rod during the operation of the engine. At the same time, the machine tool is equipped with sensors on key components to collect real-time information on the cutting cycle status. The actual operation results show that the machine tool is more stable, reliable, high-efficiency, low-noise, environmentally-friendly and easy-to-maintain than traditional dedicated machines.

In order to further improve the accuracy of tool wear prediction model and realize the online monitoring of tool machining process.A tool wear prediction model based on improved sand cat swarm optimization (ISCSO) and long short-term memory (LSTM) neural network is proposed.Using the acceleration vibration signal of the tool as the input sample, the LSTM is applied to predict the wear value of the milling tool.Aiming at the low convergence accuracy of the sand cat swarm optimization algorithm, chaotic mapping, nonlinear convergence factor and opposition point detection mechanism are introduced, and the parameters of long short-term memory are optimized using the improved sand cat swarm optimization algorithm.The experimental results show that the tool wear prediction accuracy of the ISCSO-LSTM model is significantly higher than the LSTM model.

In order to solve the problem that the traditional deep learning method can not mine the nonlinear mapping between the original vibration data and the state of rotating machinery, a deep reinforcement learning fault diagnosis method for rotating machinery based on the combination of stacking automatic encoder and deep Q network was proposed. Firstly, a "game" model of fault diagnosis was established to provide an interactive environment of observation, action and reward for the fault diagnosis agent. Secondly, the stackable automatic encoder adopted the fully connected model to learn the inherent characteristics of the fault diagnosis agent step by step. Then, by introducing memory playback, iterative update strategy and reward feedback mechanism, the nonlinear mapping between the original vibration signal and the fault mode was realised through the deep Q network. Finally, the experimental results show that the proposed method has high accuracy in the fault diagnosis of rolling bearing and hydraulic pump, which proves that the method can effectively achieve the end-to-end fault diagnosis of rotating machinery.

In view of the fact that the offline Cartesian space trajectory planning of the traditional manipulator cannot solve the problem of the contact with the environment, a real-time compliant trajectory planning method based on admittance control was proposed; The trajectory planning of manipulator in Cartesian space was divided into position planning and attitude planning, and the two were normalized by the time scale parameter s(t).The trajectory planning method was illustrated by taking the most common linear motion and circular motion as examples; An admittance controller was constructed, and the information obtained by the six-dimension force sensor was substituted into the controller to calculate the adjustment of the end pose; On the basis of trajectory planning, admittance control is introduced to solve the problem that trajectory planning cannot perceive the forces of the external environment in real time; Finally, the verification experiments of the manipulator were built, and the experiments of linear motion and circular motion were carried out; Experimental results show that the proposed method can not only achieve the high precision of the manipulator trajectory, but also actively comply the contact force of the external environment, which improves the stability and safety of the control system.

In order to improve the wear resistance of the surface of GCr15 steel ball, GCr15 steel ball, hard ball, grinding powder and grinding liquid were mixed to enhance the grinding process of steel ball.The cross section microstructure, grain size and Vickers microhardness of GCr15 steel balls were detected by optical metallography microscope, X-ray diffractometer and Vickers microhardness tester, and then the friction and wear tests were carried out on the GCr15 steel balls with different ball intensification lapping time by friction and wear testing machine, and the average friction coefficient, wear amount and wear width were analyzed.The effect of enhanced lapping time on the surface wear resistance of GCr15 steel ball was studied.The results show that when the processing time increases from 30 min to 90 min, the thickness of the surface layer increases, the degree of grain refinement gradually increases, the Vickers microhardness increases, and the wear resistance of the ball is enhanced.

Aiming at the problem of precision and efficiency of traditional calibration method of rotation center of five-axis machine tool, an automatic calibration method for the RTCP of a double-cradle five-axis machine tool with self-correction function is proposed, which used the rough automatic calibration using least square fitting, and the fine automatic calibration using "self-checked and self-correction function ".In this method, the tried-cut calibration idea is implemented, the accuracy of the existing RTCP parameters are self-checked by using the contact probe and the standard ball, and the RTCP parameters are corrected according to the error, and then checked again.The method is iteratively optimized and gradually approximates to the optimal solution to form a closed-loop calibration.After trial cutting and standard ball verification, the single-direction error is less than 0.01 mm, compared with the rough calibration error of 0.03~0.06 mm, the accuracy is improved by more than 66%, and the repeatability of multiple groups of experiments is stable.

Rolling bearing friction torque is the main factor leading to bearing heating and wear, In order to more accurately describe the changing law of bearing friction torque, it is based on the theory of pseudo-dynamics and rolling bearing thermal grid, based on elasticity Flow lubrication theory, establish a ceramic ball bearing force-thermal coupling model, analyze the influence of different working conditions and bearing structural parameters on the friction torque of ceramic ball bearings.The calculation results show that the friction torque of ceramic ball bearings is greatly affected by load, speed and pre-load.Reasonable selection of bearing structural parameters and lubricating oil temperature are helpful to reduce the overall friction torque of the bearing.

Aiming at the problem that the existing EPS motor test system structure is limited and the torque output ability of the measured motor under the flux weakening control is ignored, a PMSM performance test system for EPS based on LabVIEW is developed.The system is divided into three modules:active motor loading, measured motor control and data acquisition.By opening the parameters of the controller's flux weakening lead angle and introducing the order analysis technology, to test the torque-speed characteristics (T-N) and torque ripple (T-R) of PMSM for a certain type of EPS.The test results showed that the torque error of the motor is less than 5% at the measured speed point (less than 3% at 0 rpm).The torque ripple in the low rotation speed condition was mainly distributed in the 24th order, and the torque amplitude was less than 0.030 N·m.The other orders had no obvious torque fluctuation.The accuracy and reliability of the test bench were verified, and the product quality inspection was completed.

The existing thermal error prediction is mostly based on temperature data modeling, with a single characteristic dimension, and the characteristics of nonlinearity and coupling of thermal error, resulting in weak adaptability and low prediction accuracy of the prediction model.Aiming at above problems, a multi-source heterogeneous data acquisition scheme is designed, and a vertical machine tool thermal error prediction model based on multi-dimensional temperature, energy consumption data and artificial neural network is established.Build an experimental platform to compare the prediction accuracy of support vectors and random forest regression models.Compared with the traditional regression model, the DNN model has strong adaptability and higher prediction accuracy, and the average absolute error of Z-direction thermal error is 0.973 μm, which significantly improves the thermal error prediction accuracy while improving the adaptability of the prediction model.

To achieve the topology optimization of porous structures with a higher specific surface area to volume ratio, the local volume constraints are introduced to limit the material distribution in the design domain. Based on the variable density method of the modified solid isotropic penalty model, the constraint conditions of the periodic porous structure are constructed to establish a topology optimization mathematical model of the periodic porous structure with maximization of structural rigidity. The model is solved to obtain a two-dimensional periodic porous structure with a high specific surface area by the method of moving asymptotes. By measuring the surface area of the optimized porous unit cell and comparing it with the optimized porous structure without local volume constraints, the method can increase the specific surface area of the periodic porous structure effectively by sacrificing less structural performance.

In view of the chatter phenomenon in the process of spiral end milling, a new prediction model is proposed by using the three-dimensional cutting force model. Compared with the previous two-dimensional prediction model of end milling chatter, the new prediction model mainly considers the flexibility of the spindle tool holder, studies the dynamic stability of the spiral end milling cutter, and has higher accuracy. Through the finite element modeling of the spindle tool combination, the frequency response of the tool tip is obtained. Based on the three-dimensional cutting force model, the expressions of stable cutting depth and working speed are derived. The three-dimensional stability model is verified by cutting experiments. The results of experiments are in good agreement with the simulation results, which verifies the correctness of the established prediction model of end milling stability. Finally, based on different prediction models, the lobes of flutter stability domain are compared and analyzed. The results show that the improved prediction model has higher accuracy.

External disturbance and parameters mismatch degrade the performance of permanent magnet synchronous motor (PMSM) position servo system, in order to improve the performance, this paper proposed an improved position control method based on improved nonlinear backstepping active disturbance rejection control (ADRC).Firstly, in order to solve the problem that the control gain must be known for designing nonlinear state error feedback part in traditional ADRC controller and convergence performance needs to be further improved.An improved nonlinear controller is designed, based on backstepping control and continuous and smooth nonlinear function.Secondly, Landau adaptive observer is adopted to identify and compensate the disturbance and parameters, aiming at the limited performance of the equivalent disturbance estimation method used in traditional ADRC position controller.The simulation results demonstrate that the nonlinear backstepping active disturbance rejection control method proposed in this paper could effectively improves the tracking convergence speed and tracking accuracy of the PMSM position servo system, and enhance the adaptive ability to external disturbance and parameter uncertainty.

Aimed at the problems of high loss, low efficiency and serious rotor heating of high-end precision high-speed motorized spindle, this paper proposes a high-speed permanent magnet synchronous motor based on amorphous alloy stator core.According to the magnetization and loss characteristics of amorphous alloy stator core, the optimal design of high speed permanent magnet synchronous motor is studied, A physical analysis model of permanent magnet motor with rated power of 3.7 kW, rated speed of 7000 rpm and maximum speed of 15 000 rpm is established by using finite element method.The performance of the motor is compared with that of silicon steel (35W270) motor with the same structural parameters, such as loss and efficiency.The analysis results show that the theoretical iron loss of amorphous alloy used as the stator core of high-speed permanent magnet motor is 76.5% lower than that of silicon steel motor; Through the actual test of two prototypes, considering the deterioration of iron loss caused by iron core processing and assembly process, the measured iron loss decreased by 50.7%, the measured efficiency is 1.4% higher, and the temperature rise of stator winding decreased by 15.1 K under the same load and cooling conditions, which shows that the amorphous iron core has obvious advantages in the application of high-speed motorized spindle.

In order to improve the control performance of permanent magnet synchronous motor (PMSM), a vector control system with ADRC and improved ESO is designed.First of all, the traditional ESO is improved by using smooth nonlinear functions to improve its ability of signal smoothing.The improved ESO is paralleled with ADRC to improve the accuracy of signal processing.Then, the electromagnetic torque and load error, speed error and time integral of the system are actively compensated to the current loop to improve the response speed and anti-interference of the system.Finally, the crow algorithm is used to optimize the parameters of ADRC to reduce the complexity of manual parameter tuning.Simulation results in Simulink show that the improved system has better robustness and stability than ADRC and PI control.

Aim at the key feature of shear performance of aviation fasteners, a three-dimensional numerical model of Ti-6Al-4V fasteners under double shear conditions of blanks and heat treatment is established.The shear resistance of Ti-6Al-4V fasteners and their shear evolution process are correspondingly revealed.The results show that the model can accurately simulate the load-displacement curve of titanium alloy fasteners under double shear conditions, and the average error between the shear strength obtained and the experimental value is 0.5%.After heat treatment, the shear strength of Ti-6Al-4V fasteners is increased by 35%.In the process of shearing, the stress concentration phenomenon of the workpiece gradually extends from the contact of the lower knife surface to the entire shear surface until the workpiece fails to be sheared.This research can provide a theoretical basis and guidance for improving the shear strength of titanium alloy fasteners.