Research On Digital Twin System For Milling Tool Wear Monitoring

With the fast advance of new generation science and technology,there is a continuous transformation of more and more traditional manufacturing industries to digitalization and intelligence.At the same time,more and more enterprises choose to transform their original production routes.CNC machines are the most frequently used tools in the production sector,and the use of tools in the process has a major influence on the quality of the processing and the processing efficiency.But so far,there are still problems in the machining process such as unstable manufacturing process,poor timeliness of data,and weak visualization of tool wear monitoring.These problems are still in urgent need of a reasonable,effective,and efficient solution.In respond to the abstracted problems,a digital twin system for tool wear monitoring is introduced in this paper.The construction of a digital twin with co-existence and interaction in physical space and virtual space is therefore carried out.The majors of this thesis are as follows.(1)This paper introduces the overall scheme design of digital twin system for machine tool processing.With the help of the five-dimensional model of digital twinning,the framework and five-layer structure system of digital twinning system are built.After analyzing the subject task requirements and software features,the appropriate 3D modeling software UG and the digital twin system platform Unity3D are selected.Meanwhile,the machine tool equipment in the virtual space was built with the rendering function of 3ds Max software.According to the functional requirements of the twin system,collision detection,code analysis,machine tool motion simulation,dynamic cutting,data visualization,and other modules are developed,and the multi-source heterogeneous data is transmitted and managed using the MySQL database.(2)This paper briefly introduces the process of tool wear,including different forms of wear,types of wear,and stages of wear.After analyzing the principles and characteristics of the signals collected by various sensors,this paper chooses to collect force and acceleration signals in the experimental process.The specifics of the tool wear experiment are elaborated.After the experiments are carried out,the collected experimental data are organized and analyzed,including pre-processing operations and characterization of the experimental data.The features that were strongly correlated with tool wear were selected by the Pearson correlation coefficient method and were used as input to the neural network algorithm model.(3)According to the characteristics of the tool wear monitoring model in handling regression tasks.This paper provides an introduction to the constructs and the working strategies of BP neural network,CNN and LSTM.The data processed in the third chapter is used as the input of the algorithm model.The partitioned data sets are used to verify the performance of the model.Finally,the computational precision of the neural network model proposed in this paper is improved by comparing and analyzing the evaluation measures.(4)This paper applies the concept of the digital twin to a machine tool milling and machining scenario and uses it as a case study.First,a communication connection between the physical machine tool and the software platform of the digital twin system is developed.It is possible to interact,connect and control each other.Based on realizing machine tool communication,the realization process of twin data transmission and management is put forward.Taking the digital twining system of five-axis CNC machine tool milling as an example,two parts of function verification and online connection verification are carried out according to the content introduced in this paper.Functional verification is mainly the content of the second chapter.The online verification includes virtual and real interaction verification,the interactive verification between the twin system and tool wear monitoring,and the operation stability test of the twin system.At last,the feasibility and validity of the digital twin system of CNC machine tool machining for milling tool wear monitoring is verified by a case study.