Design And Key Technology Research Of Adaptive Robot For Pipe Grinding And Cleaning

At present,plastic pipes are widely used in many fields of industry and agriculture,such as water supply and drainage,gas,electric power and fire protection,and it is closely related to men’s life.However,in the process of working and production,plastic pipes are easy to a variety of defects,such as pockmarked surface,impurities,bulges and bubbles,etc.,while the internal space of the pipes is small,and the operators can’t efficiently and accurately detect the operation.In order to improve the quality of the pipe before they go out and convenient internal pipeline maintenance,reduce work accidents,this thesis designed a kind of adaptive pipe polishing cleaning robot with machine vision and inspection on wall of pipeline caused by defects,burr,concave and convex,pits and defects such as fold,directly on the polishing cleaning,in order to improve the quality of pipe.On the basis of the analysis and research of the current development and existing problems of the domestic and foreign pipeline robot,in line with the environment and characteristics of plastic pipeline processing,the entirety design of the self-adaptive pipeline grinding and cleaning robot is completed,including the architectural design of the mechanical system and control system.The modular design method is used to design the structure of the pipeline robot,and the walking drive device and the grinding and cleaning device are designed in detail.The self-adaptive reducing mechanism is designed by using the principle of shearing lifting and three jaw chuck reducing diameter,and the intensity,stress and transfiguration of the key units are analyzed by using ANSYS software.A control system equipped with a highdefinition camera is designed to realize the shooting,adaptive grinding and cleaning functions of the pipe wall.According to the structure of the designed pipeline robot,the performance of the pipeline robot,such as passability and obstacle surpassing ability,is analyzed.According to the characteristics of its walking and working,the force models of robot moving in different pipes,the force models of obstacle crossing and the grinding model were established,and the force,interference,motion characteristics and grinding operation were analyzed.Adams software is used to import the simplified pipeline robot model as a whole,and the integrated simulation method is used to analyze its performance in the pipeline,and the movement of the pipeline robot in the pipeline is studied.A defect detection system for pipeline robot based on machine vision is designed.Finished the hardware selection of the illuminant,HQ camera and wide-angle lens detection system;MATLAB was used to finish the visual calibration of defect detection system and achieve camera parameters.After the comparison and analysis of MSR,bilateral filtering and homomorphic filtering algorithms,the improved MSR algorithm is used to preprocess the collected pipe inner wall defect images,and the image enhancement of pipe wall defect is realized.The defect types of pipeline inner wall were studied,and the advantages of Canny edge detection and local automatic threshold segmentation were combined to realize image segmentation,and the defect types were identified by shape factor and gray contrast.Finally,the defect location is analyzed and solved by combining the defect information and camera parameters.According to the overall design of the adaptive pipe polishing and cleaning robot control,STM32F4 and Raspberry are selected as the controllers,and Altium Designer is used to design the hardware circuit of the minimum system,motor drive,power supply,communication and sensor module respectively.The fuzzy PID self-adaptive control system of pipeline robot is designed,and the system is emulated and analyzed by Simulink.The results show that the fuzzy PID adaptive control system has better control effect.On this basis,the software of the control system is designed,and the modules of walking drive,grinding and cleaning,image data acquisition and human-computer interaction interface are completed.