Peg-in-hole Assembly Task Of Nuclear Power Plant Operation And Maintenance Robot Based On Man-machine Cooperation

In the daily operation and maintenance process of nuclear power plant,a large number of high-precision assembly tasks such as shaft hole assembly are involved.As the last process of parts assembly,shaft hole assembly is an important part of operation and maintenance of nuclear power plant.Aiming at the shaft hole assembly task of nuclear power plant,which has the characteristics of complex working environment and high radiation dose,the traditional shaft hole assembly control method can not meet the requirements of nuclear power plant operation and maintenance.In order to improve the efficiency of daily operation and maintenance of nuclear facilities and reduce the radiation dose of staff,this paper studies the control method of shaft hole assembly of nuclear operation and maintenance robot based on human-computer cooperation.Firstly,the shaft hole assembly operation of nuclear power plant is analyzed.According to the actual shaft hole assembly requirements of nuclear power plant,the overall framework of the shaft hole assembly system of nuclear operation and maintenance robot is constructed.At the same time,the master-slave equipment is selected,the kinematics model of the masterslave equipment is established,and the forward and inverse kinematics of the manipulator and the workspace of the master-slave equipment are analyzed.Then,the control method of axis hole assembly for nuclear operation and maintenance robot based on human-computer cooperation is proposed.The assembly task of shaft hole is divided into search phase and insert phase.In the search phase,the master-slave heterogeneous teleoperation robot system is built.The master-slave manipulator controls the slave manipulator to place the pin clamped by the end gripper through the master-slave workspace mapping algorithm.Based on the image information and force feedback information,the pin is placed in the center of the hole.In the insertion stage,based on the deep reinforcement learning algorithm model,the Markov decision-making process is constructed according to the force feedback of the end of the slave manipulator and the position and pose of the pin.The mapping relationship between the current state and the action of the slave manipulator is obtained through training,so as to obtain the optimal control strategy of the axle hole assembly.According to the force feedback information,the position and attitude of the pin are adjusted in real time to control the pin insertion into the hole.Finally,the control method is verified by building an experimental platform in real scene and simulation environment and designing corresponding experiments.The experimental results show that the proposed method based on human-computer cooperation can effectively complete the high-precision assembly task in the process of nuclear operation and maintenance.The method has high stability and can improve the operation safety and efficiency of nuclear power plant.