Research And Implementation Of Digital Twin-based Joint Robot Health Condition Assessment Technology

Joint robots are the basic equipment in intelligent manufacturing,and as joint robots are oriented to fast,accurate and intelligent development,their functions are becoming more and more complex,and unplanned stoppages caused by failures can seriously affect the production schedule of factories,but there is still a large gap between China and foreign countries in the intelligent operation and maintenance of related equipment,so it is of great practical significance to assess the health condition of joint robots.In this thesis,we propose a digital twin-based health condition assessment scheme for articulated robots with reference to the digital twin five-dimensional model,which is to solve the problems of poor model fidelity and low data utilization in the previous model-based or data-driven single methods.The work in this thesis is as follows.(1)A digital twin-based articulated robot health condition assessment architecture is designed.The joint robot health condition assessment idea is proposed and the main faults of the joint robot are analyzed,and the stator current is subsequently used to diagnose the joint robot motor bearing faults.And the joint robot scene sensing scheme is designed to collect and store the monitoring data persistently.(2)A health condition assessment method for jointed robots is proposed,and the effectiveness of the method is verified by experiments.A convolutional neural network is used for motor bearing fault diagnosis,and the accuracy of the model is improved by increasing the size of convolutional kernels in the first two convolutional layers.A recurrent neural network is used to predict bearing faults by predicting the current,and the accuracy of both recurrent and cycle prediction methods is compared.(3)An articulated robot digital twin model was constructed.Based on the objectoriented principle and using multi-domain modeling software,a digital twin model of the articulated robot was built by constructing the geometric and electrical models of the articulated robot to demonstrate the operating conditions of the articulated robot.(4)A remote operation and maintenance system for the jointed robot was built based on the above results.The remote operation and maintenance system is a platform for providing services to the outside world.The framework of the remote operation and maintenance system is designed first,then the data storage layer and the digital twin model of the joint robot are constructed,and finally the data management module is developed,and the RESTful style interface is designed to perform fault diagnosis,fault prediction and the invocation of the 3D model of the joint robot,and the system is deployed in the Linux server.