Integrated Visual Diagnosis System For CNC Tool Condition

In the context of the fourth industrial revolution,to transform China from a big manufacturing country to a strong one," Made in China 2025 " was proposed and carried out.It sets higher requirements for intelligent and automated manufacturing processes.Cutting tools' conditions are essential for product's quality and production's safety.Developing a set of CNC tool conditions integrated visual diagnosis system is of great significance for improving the level of intelligent of manufacturing industry.This paper analyzed the mechanism and shape of the tool wear during the cutting process.The visual diagnosis principle based on the tool main flank and the cutting-edge contour was studied.The visual diagnosis principle includes the image preprocessing method,the feature extraction method of color,texture and shape,and DNN classification model for tool condition based on momentum gradient descent method.A set of CNC tool condition integrated visual diagnosis system was designed for SINUMERIK 840 D.The hardware system was designed and tested from the embedded computer system,image acquisition device,lighting device and CNC communication module.Software system were researched and developed from lower machine software,PROFIBUS-DP communication program,image processing software and NC system integration software,and the code writing and debugging were completed.The transplant and programming of embedded operating system were realized,the writing and developing of device driver and application software were completed.The development of the PROFIBUS-DP slave program and the image processing software on cloud server were completed.The integration between the visual diagnosis system and NC system was realized through developing M instruction for visual diagnosis in the 840 D system and completing M signal decoding and writing of the PROFIBUS-DP master station software in the S7-300 PLC.The neural network of different topologies and algorithms was compared by experiment,which verified the accuracy and efficiency of DNN structure.Experiments were conducted in HTM125600 milling machining center,while the visual diagnosis system was embedded in the NC system as one of the CNC module.The results showed that the visual diagnosis system realized the tool image acquisition,the image preprocessing,feature extraction and condition recognition by call subroutine,and realized the integration with NC system,the system proved to be efficient.