Monitoring And Data Analysis Of Hardened Steel Mold Milling Process

The hardened steel material after quenching has a martensitic structure,high hardness,high strength,and low thermal conductivity,resulting in high cutting temperature and enhanced adhesion on the surface of the tool and chip,the surface of the tool,and the tool wears quickly.Therefore,a tool condition monitoring system is added during the processing process,which can prevent the machine from being stopped due to untimely tool replacement and reduce parts scrap.To this end,this article aims at monitoring the tool wear during the Cr12 Mo V milling process,using a variety of sensors and data acquisition methods to design and implement a low-cost tool condition monitoring system.The following research is performed:In terms of signal acquisition,the milling path of the reciprocating cutter is determined through theoretical analysis.The test platform built by cutting force signal,vibration signal and energy consumption signal is selected,and the test parameters and blunt standards suitable for the test conditions are set.In order to reduce the number of tests,an orthogonal test is designed based on the three cutting parameters.The optimal parameter combination is selected and introduced through the range analysis of the test results.The sensor signals corresponding to the tool wear status are collected to provide data support for subsequent processing and analysis.In terms of signal pre-processing,effective signal data is extracted through zero-point drift processing,interception method,and five-point three-time smooth filter processing,which reduces a lot of noise and redundant data in the original signal;analyzes the basic characteristics of wear to obtain different sensors correlation between signal and wear.In terms of signal analysis,the three monitoring characteristics of tool wear are analyzed,and the perceptual features related to the amount of tool die wear are extracted from the time domain,frequency domain,and time-frequency domain.A Pearson correlation coefficient analysis and kernel master are proposed.Component analysis(KPCA)combined with feature dimensionality reduction method finally identified feature combinations that can effectively reflect tool wear information,thereby improving the performance of subsequent network model recognition.In terms of status recognition,the BP neural network and Levy-GA-BP neural network model are used to complete the recognition of the tool wear status.The identification of different models is analyzed,and the results show that the Levy-GA-BP neural network-based prediction is better.With regard to the establishment of the monitoring system,the milling cutter wear monitoring system is developed using MATLAB software and the cutting database platform is improved.By entering the corresponding parameters,the expected monitoring effect is achieved.The monitoring system can be used for Optimization of cutting process parameters and tool design provide basic data.