As one of the important functional components of earth-pressure balance shield, the screw conveyor often needs to be adjusted in design to meet new engineering requirements. However, the design adjustment is easy to lead to excessive changes in product indicators, which will have a negative impact on the reliability, safety and maintainability of products. In order to maximize the reuse of the existing design scheme verified by engineering and reduce the negative impact of design adjustments, an intelligent decision method for the adaptive design of shield screw conveyor was proposed. Firstly, the product indicators and design parameters of the shield screw conveyor were collected to construct the product data set, and the dependency matrix and the correlation matrix between product indicators and design parameters were established through the dependency analysis and correlation analysis, respectively; then, based on the correlation between product indicators and design parameters, the hierarchical clustering was carried out to obtain different product element clusters, and the influence scope of product indicator changes in the process of design adjustment was calibrated; and then, the key design parameters corresponding to each product indicator in the different clusters were identified by combining the dependency matrix and hierarchical clustering results; finally, the neural network was used to construct the regression prediction model of product indicator-design parameter, and the design parameter adjustment schemes with relatively good adaptability to new product indicator requirements were obtained through search. Through the comparative analysis of the demand value and predicted value of the product indicator of shield screw conveyor, it was found that its mean absolute percentage error (MAPE) was 3.964 7%, and the achievement degree was 96.04%, which basically met design requirements. The results show that the proposed method can accurately calibrate the potential influence range of the product indicator change of shield screw conveyor, effectively identify the key design parameters on which the product indicator depends, and realize the intelligent solution of key design parameters. It can not only provide decision support for the adaptive design of shield screw conveyors, but also provide technical reference for the product adaptive design to meet personalized and differentiated requirements.

In order to solve the problem that the traditional detection method can only detect the overall axial displacement of the corrugated compensator in the industrial pipeline, but can not detect the axial displacement between its adjacent corrugations (the axial displacement between adjacent corrugations varies unevenly, and if it exceeds a certain limit, the corrugated compensator will be damaged), an axial dimension detection method for the corrugated compensator based on the image recognition is proposed. Firstly, the Sobel operator was used to detect the edge of the corrugated compensator image smoothed by the Gaussian filter, and the edge features of the original image were preserved to the greatest extent. Then, the grayscale and Otsu threshold segmentation processing for the image were carried out. After inverting the image and filling holes, the ideal binary image with target grayscale value of 1 and background grayscale value of 0 was obtained. Finally, based on the particularity of the corrugated compensator shape, the upper, lower, left and right edge feature points of the part with grayscale value of 1 in the corrected binary image were calculated, and the actual dimensions of the corrugated compensator and its adjacent corrugations were obtained through the proportional relationship. The experimental results showed that the proposed method could accurately detect the overall axial dimension of the corrugated compensator and the axial dimension between its adjacent corrugations under different compression levels. The dimension detection method effectively solves the problem of damage to the corrugated compensator caused by the uneven axial displacement between adjacent corrugations, which is of great significance to ensure the smooth operation of the corrugated compensator and maintain production safety.

The numerous quality inspection items and long inspection time in the production process of automobile combination instruments have restricted the further improvement of production efficiency to a certain extent. To this end, a support vector machine (SVM) classification prediction method based on the improved max distance synthetic minority over-sampling technique (MDSMOTE) was proposed. Firstly, the feature selection for the original automobile combination instrument production data was carried out combined with the expert experience, and the class imbalance rate I_R was introduced into the MDSMOTE to expand the selected feature data; then, the error penalty factor C and the kernel parameter γ of the SVM were optimized by the particle swarm optimization (PSO) algorithm; finally, an optimized SVM classification prediction model was established to make classifications for the automobile combination instruments. Compared with the prediction results of other classification prediction models on different data sets, the SVM classification prediction model based on the improved MDSMOTE was superior to other models in terms of the evaluation indexes as accuracy, F value and geometric mean value. The proposed method shows strong generalization ability and stability in the classification of automotive instrument products, which can provide an effective reference for the improvement of production efficiency of instrument manufacturers.

The lifting arm is the key component for realizing the cam opening movement on the dobby loom. In order to meet the requirements of automatic manufacture for the lifting arm of dobby loom, a novel assisting-riveting parallel robot is design to achieve the high-speed and high-precision grasping and placing of parts and components during the riveting process of the lifting arm, so as to significantly improve the production efficiency and ensure the processing quality. Firstly, the topological structure characteristics of the novel assisting-riveting parallel robot mechanism were introduced, and the type and number of the degree of freedom of robot mechanism were analyzed by combining the screw theory and the modified G-K (Grübler-Kutzbach) formulation. Secondly, the position closed-loop vector equations of this robot mechanism were established, and the analytical expressions of forward and inverse position solutions were derived. The velocity and acceleration mapping models were established by the derivation method. Thirdly, in order to effectively evaluate the kinematic transmission performance of this robot mechanism, the global comprehensive performance indices were defined based on the global mean value and fluctuation of condition number of the kinematic Jacobian matrix; at the same time, combined with the spatial structure characteristics of hook joint, considering the rod interference and the transmission performance and structural compactness of the mechanism, the constraints of various angles and rod lengths were obtained, and the dimensional synthesis model of this robot was established. On this basis, the optimal dimensional parameters of robot were obtained by means of the particle swarm optimization (PSO) algorithm for optimization and solution, and then the three-dimensional virtual prototype model of the robot was established. Finally, in view of the engineering requirements of the assisting-riveting operation for the dobby loom lifting arm, the virtual simulation verification for the robot was carried out by means of the multi-body simulation software. The results indicated that the optimized assisting-riveting parallel robot had excellent comprehensive performance and met the requirements of engineering applications. The research results can provide an effective solution for the automatic manufacture of the dobby loom lifting arm, and lay a theoretical foundation for the solid prototype manufacturing and experimental study of assisting-riveting robots.

At present, maintenance operations such as derusting, painting and testing of petrochemical storage tanks in my country are usually completed manually, which is inefficient and dangerous. The wall-climbing robot can complete the above maintenance work instead of manual work, but its frequently-used contact permanent magnet adsorption method will affect its obstacle-surmounting operation. Therefore, aiming at the wheel-legged wall-climbing robot with obstacle-surmounting ability, a design and optimization method for improving the adsorption capacity of its permanent magnet adsorption device was proposed. Firstly, based on the principle of obstacle-surmounting and permanent magnet adsorption of the wheel-legged wall-climbing robot, four design schemes of non-contact permanent magnet adsorption device were proposed. The two-dimensional static magnetic field simulation analysis for the permanent magnet adsorption device was carried out by the Ansoft Maxwell software, and its optimal structure was obtained. Then, through force analysis, the structural parameters affecting the magnetic adsorption force of the permanent magnet adsorption device were determined. Finally, based on the principle of control variable method, the influence of structural parameters of permanent magnet adsorption device and steel wall on the magnetic adsorption force was analyzed through simulation, and the optimal value of each structural parameter was determined. The results showed that when the radius of steel wall was 10-14 m, the gap between adjacent permanent magnet adsorption units and the number of permanent magnet adsorption unit had a relatively great influence on the magnetic adsorption force; when the gap between adjacent permanent magnet adsorption units was 2-4 mm and the number of permanent magnet adsorption unit was 3-5, the utilization rate of magnetic adsorption force was relatively high. Experiments under certain conditions had verified the accuracy of optimization results. After optimization,the measured value of the magnetic adsorption force of the permanent magnet adsorption device was about 16.45% higher than the simulated value on average, and the optimized device could effectively ensure the stable movement of the wheel-legged wall-climbing robot on the steel wall of petrochemical storage tanks. The research results can provide reference for the structural design and optimization of the magnetic adsorption device.

In order to reduce the particle size and distribution width of sepiolite ore powder after grinding, a spiral stirring mill was developed based on laminated crushing theory. The sepiolite grinding was simulated by EDEM （engineering discrete element method) to study the motion law of the dielectric sphere in grinding cavity. The optimal process parameters were obtained by taking the velocity gradient, rotational kinetic energy of dielectric sphere and wear amount of stirring mill as evaluation indexes. The self-developed spiral stirring mill was used to carry out sepiolite grinding experiment, and the particle size of the sepiolite ore powder after grinding was detected. The simulation and experimental results showed that sepiolite ore powder with relatively small particle size and distribution width could be obtained when the rotating speed of mill and diameter of dielectric sphere were relatively small, and the wear amount of the mill was relatively small. The research result can provide reference for the optimization of sepiolite grinding equipment and the determination of optimal process parameters.

In order to solve the problem of cutterhead fracture during the operation of a disc ditcher, the response surface analysis method and sensitivity analysis method were used to analyze and optimize the performance of the ditcher. Firstly, the simulation model of the ditcher was established, and the dynamic simulation of the cutter of the ditcher was carried out by using the finite element analysis method. Through the response surface analysis, it was concluded that the stress concentration at the edge of the cutterhead was the main reason for the fracture of the cutterhead; secondly, through sensitivity analysis, the sensitivity of the maximum stress leading to the fracture of the cutterhead to the stress of each cutter was obtained, and a multi-condition optimization model based on different values of stress on cutter, cutter material elastic modulus and cutterhead thickness was established; finally, the cutterhead material was selected according to the optimization model, and the optimization scheme was obtained, that was, the cutterhead performance was the best when the cutterhead material was Q245. It was concluded from the user's feedback that the improved scheme could effectively solve the problem of cutterhead cracking of ditcher and improve the service life of the ditcher.

In order to study the temperature distribution of the electromagnetic micro hammer peening mechanism in the working process and improve its power density, first of all, based on the existing electromagnetic micro hammer peening mechanism, the relationship between its thermal effect and output power was analyzed from the perspective of energy, and it was clarified that the thermal effect would limit the maximum output power of mechanism. Then, the CFD (computational fluid dynamics) simulation analysis of the temperature field of electromagnetic micro hammer peening mechanism under air-cooling and water-cooling mode was carried out by the COMSOL Multiphysics software, and the equivalent convective heat transfer coefficients under different inlet boundary conditions were determined; at the same time, according to the relationship between the equivalent convective heat transfer coefficient and the inlet boundary condition, a transient temperature analysis model for this mechanism was established. Finally, a temperature measurement experimental platform for the electromagnetic micro hammer peening mechanism was built, and the temperature rise characteristics and the steady-state temperature characteristics of the mechanism were experimentally studied. The experimental results showed that the proposed thermal effect simulation analysis method for the electromagnetic micro hammer peening mechanism was relatively reasonable and accurate, which could provide a reference for the temperature control and structure optimization of micro hammer peening mechanisms.

The main feed water regulating valve (MFWRV) installed in the secondary circuit of the nuclear power plant regulates the water level in the steam generator by controlling the feedwater flow. Therefore, the flow characteristics of MFWRV affect the operation safety and efficiency of the nuclear power plant. Firstly, the dimensionless unbalance degree δ of circumferential arrangement of throttling window was proposed, and the circumferential arrangement of throttling window of MFWRV was designed; secondly, the numerical simulation method verified by grid independence and flow experiment was used to analyze the velocity and pressure distribution characteristics of the flow field in the valve; finally, the influence of the unbalance degree of the circumferential arrangement of the throttling window on the unbalance torque of the valve core and the flow coefficient of the valve was discussed. The results showed that the uneven pressure distribution at the bottom of the valve core would produce unbalanced torque to the valve core, increase the friction between the valve core and the sleeve, and then aggravate the abrasion between the valve core and the sleeve, which might cause the action stagnation of the valve core and shorten the service life of the valve core and the sleeve; when the valve opening was less than 40%, adjusting the circumferential arrangement of the throttling window had little effect on the unbalanced torque of the valve core; when the valve opening was greater than 40%, the unbalanced torque of the valve core could be effectively changed by adjusting the circumferential arrangement of the throttling window; under the same conditions, by increasing the unbalance degree of the circumferential arrangement of the throttling window from 0 to 0.5, the maximum unbalance torque of the valve core could be reduced by 22.10%, while the flow coefficient of the valve could be reduced by only 2.74%. The research results are conductive to the design and optimization of MFWRV and similar valves.

In order to improve the stability of magnetic levitation conveyor belt and reduce the influence of unmodeled dynamics and unknown external interference on the control performance of magnetic levitation system, a control strategy combining model reference sliding mode control and sliding mode control based on improved reaching law was proposed based on the improved ESO (extended state observer) technology. Firstly, the sliding mode design of the reference model was carried out. On this basis, according to the requirements of fast response and robustness of the magnetic levitation system, the traditional reaching law was improved by combining the power reaching law and exponential reaching law, and then a sliding mode control based on the new reaching law was designed; secondly, a new nonlinear function was designed to improve ESO. Based on the improved ESO, the disturbance and state of the system were observed and estimated, and the observation results were added to the new sliding mode controller to compensate the external interference, so as to improve the control performance of the new sliding mode controller. The simulation results showed that compared with the traditional sliding mode control based on exponential reaching law, the overshoot of air gap output of magnetic levitation system was reduced by 15.15%, and the system had higher robustness; compared with the sliding mode control method based on the improved reaching law, the proposed controller could make the system free of chattering and have better tracking performance. Under the model reference sliding mode control based on the improved ESO, the magnetic levitation system can operate stably and has good control performance. The research results have certain reference value for the levitation control of magnetic levitation conveyor belt.