The Structural Design And Trajectory Planning Of The Underwater Welding Robot

Nuclear power is an integral part of the power supply. Nuclear reactor construction has been adhering to a high standard. But with the time goes by some facilities would wear, even break.Once the leak of the radioactive material occurs, maintenance personnel will be directly in radioactive environments, which is very risky. Therefore, people have been working using robot technology for the inspection and maintenance of nuclear facilities. Some significant progress of the inspection of single test in certain structure has been made. The inspection and maintenance of steam generators and reactor shells by robot has been very mature.This paper investigated the current situation of nuclear power robots and underwater robots at home and abroad, studied their structure and motion parameters, advantages and disadvantages. Nuclear detection robot designed for specific environment was difficult to be devoted to other uses. Underwater robot mainly driven by hydraulic and dominated by master-slave, accuracy depended on the operator was used to observing, seizing and cannot be used for welding. Nuclear rescue robot belongs to mobile robot which was mainly to observe and measure which can't be used in underwater environment. Therefore it has a great significance to specially design a robot for underwater welding of nuclear reactor spent-tank.Based the actual work environment, the structure of a 6-DOF robot was designed. Each DOF was an independent revolute joint. Each joint was driven by electric drive for the hydraulic device may dirty the water and sealed by dynamic seal. Modularization design methodology was employed and each joint individually sealed improved the reliability of the robot. The links coordination was build. The robot's forward kinematics analysis was conducted using DH method, so did the inverse kinematics analysis of the robot using geometric method and inverse transform method, which lay the foundation for subsequent dynamics and trajectory planning of robot.The dynamic function of the robot was established using Lagrangian method and also the function was derived and simplified. The robot hydrodynamic model was studied. The equivalence of the Morison Equation and the Strip Theory was verified and the hydrodynamic function was got and derived, thus,the analytic solution was got and calculated using Mont Carlo method. The selection of motors and reducers and the design of robot control were all based on the calculation.Study of the control of the robot was conducted. After analyzed the spent-tank, it found that robot control system was not only to meet the accuracy of trajectory tracking but also needed force control function to avoid collisions which may damage the tank or equipment when catastrophe occurred. Impedance control based on position was proposed. The sensor can measure the force between the robot and the environment to adjust the trace of the end to avoid rigidity collision or miss the track. The controller was designed based on the control equation which simplified from robot dynamics function.Study of trajectory planning of the robot was described. Transformation function showed that trajectory planning of the wrist can replace the trajectory of the wilding torch. Analysis showed that the constraint conditions of the welding were speed and position, thus, cubic spine method can be used for trajectory planning in joint coordinate space. A modified BDJP method based on Taylor's BDJP method was proposed which combined the advantage of trajectory planning in joint coordinate space and Cartesian space. Taylor's BDJP method used the error of the midpoint of the joint space as the error of trajectory planning, which may cause mistake although it has a small amount of calculation. The modified BDJP method used the maximum error,thus,calculation is greater than the BDJP method's,but it's more accurate.In order to verify the correct of the calculation and analysis,a two-DOF platform controlled by dSPACE was built. Three experiences of the single joint servo control, trajectory tracking and impedance control were conducted. The effectiveness of control algorithms was demonstrated.