Tool Wear Image Detection Based On Digital Twin Technology

In the process of metal cutting,tool wear always exists,which will also affect the machining accuracy and surface quality of the parts.Therefore,it is important to detect tool wear status.As a method to realize many advanced concepts such as industrial Internet and intelligent manufacturing,digital technology is increasingly used in intelligent factories.In this paper,aiming at the complex edge line of ball end mill and difficulty in collecting wear images,Unity3 D virtual reality engine is regarded as a development tool to detect tool wear characteristic values,and a tool wear detection twin system is established.The main research contents are as follows:Firstly,the general planning and functional design of the tool wear detection twin system are carried out.Regarding Pro/E and 3Ds Max as geometric modeling software,combined with the UR10 man-machine coordination manipulator parameters,a three-dimensional model is established.The model is built with Unity3 D virtual reality engine,and a virtual platform for tool wear detection is designed to finish the design and development of functions such as controlling position and gesture,planning trajectory,communicating data,online control and monitoring view angle.Secondly,the tool wear image acquisition planning is studied.Based on the D-H parameter method,the kinematics modeling of the UR10 man-machine collaborative manipulator is established,which is used as the object to the kinematics forward solution and inverse solution are derived,and the result is verified with MATLAB Robotic toolbox.Furthermore,combined with the edge line equation of the side edge and the one of the bottom edges of the ball end mill,the trajectory planning of the end of the manipulator is carried out,and a virtual control system is established.Next,the obtained tool wear images are processed to obtain the tool flank wear value.Aiming at the problem of the uncertainty of the initial angle of the tool,the method of identifying the characteristics of the bottom edge line is adopted to determine the initial deflection angle of the tool.The maximum between-class variance method(OTSU)is adopted to realize automatic threshold segmentation,which can divide the tool flank wear area by combining with Hough Transform,and the characteristic value of the tool wear area is extracted at the same time.Finally,the tool wear detection twin system is tested.Based on the TCP/IP protocol,the information interaction between the virtual scene and the physical scene is realized.Depending on the system requirements,the offline and online systems were tested respectively to verify the feasibility of the virtual-real interactive control system and the feasibility of the tool wear detection twin system.