Research On Key Technologies For Efficient And Intelligent Monitoring Of Tool Damage On-machine

The intelligence level of high-end CNC machine tools in China is low.In the actual processing,most workers estimate whether the cutting tool is damaged by experience,which is easy to cause misjudgment,lead to tool waste and machining surface quality deterioration.Especially in the processing of complex parts such as aero-engine impeller,blade,turbine and high-precision gear,gearbox,it is easy to cause huge economic losses.At present,tool damage monitoring mainly includes sensor online monitoring method and machine vision detection method.The reliability of on-line monitoring method needs to be improved.Most of the machine vision detection methods need to dismantle the cutting tools and carry out off-line detection in the laboratory environment,which is inefficient and difficult to be truly applied to the production site.Therefore,it is urgent to develop an on-line high-efficiency intelligent monitoring system and method for tool damage,so as to promote the intelligent improvement of high-end manufacturing industry.First,a tool damage monitoring method combining on-line and in-position is proposed,and integrated and mobile machine vision in-position diagnosis systems are developed for single machine tool and workshop production line respectively..The multi-sensor fusion online early warning system is used to preliminarily predict the tool damage,and the machine vision diagnosis system is further used in the machining environment to detect the tool damage at an appropriate frequency and high precision,which realizes the efficient and reliable monitoring of the tool damage on the machine.Secondly,a tool damage dynamic monitoring method based on the fusion of vibration and acoustic emission sensor information decision level is proposed,which further improves the mapping accuracy of multi-sensor signal characteristics and tool damage,and realizes the highly reliable dynamic prediction of tool damage.The reliability of the method is more than 95% verified by PHM tool wear data set.Compared with the existing multi-sensor feature level fusion methods that have achieved better results,the prediction reliability is improved by more than 2%.Thirdly,a high-precision tool damage identification method based on visual feature migration and cutting edge reconstruction is proposed,which realizes the accurate identification and measurement of tool damage features,and solves the problem that it is difficult to identify tool damage information under the interference of oil and dust in the processing environment.The accuracy of this method can reach more than 98% verified by on machine tool damage image.Compared with the current popular local variance method and adaptive threshold method,the detection accuracy of the bottom edge damage is improved by more than 9%,and the detection accuracy of the side edge damage is improved by more than 15%.Finally,the developed integrated tool damage visual in-position high-precision diagnosis system is used to collect tool images at variable frequency,combined with the online high-reliability early warning system,and the tool damage on-machine monitoring experiment is carried out.The comprehensive monitoring results further show that: The tool damage monitoring system and method realize the efficient and intelligent monitoring of tool damage on the machine,improve the monitoring efficiency and reliability,and provide technical support for the improvement of the intelligent level of domestic CNC machine tools.