Research On Tool Wear State Based On Cutting Audio And Temperature Parameter Fitting

Machinery industry is the pillar industry of national economic development.With the progress of science and technology,machinery industry is developing towards the direction of intelligence.As an important part of advanced manufacturing technology and an urgent technology to be solved,tool condition intelligent monitoring technology has become a research hotspot at home and abroad.In this paper,the tool wear state in NC turning process is studied,and the following work is carried out:First of all,the cutting simulation temperature of the tool under different wear conditions is obtained by using the Third Wave Advant Edge software.Secondly,the acquisition system of cutting temperature signal based on thermocouple method and the acquisition and analysis system of cutting audio signal based on Lab VIEW platform are designed.Then,after analyzing the data of turning temperature,it is found that the temperature of turning tool tip is similar to that of simulation.And with the increase of tool wear,the cutting temperature also rises,especially in the stage of severe wear,the tool temperature rises sharply,which is consistent with the simulation results.After analyzing the sound signal in time domain and frequency domain,it is found that the sound signal of normal cutting is mainly concentrated in 1-5 k Hz,among which the audio signal in the frequency range of 1.4-1.7 k Hz and 2.7-5 k Hz has a good correlation with the tool wear state.After seven layers of multi frequency decomposition of the audio signal,it is found that the energy of the cutting audio signal of the tool under different wear conditions is mainly concentrated in the four frequency intervals of D3,D4,d5 and D6(340-5520 Hz).The frequency range with the best correlation is D4(1372-2760 Hz),followed by D3(2760-5520 Hz).Finally,the releef-f method is used for feature selection,and the selected correlation features are the mean square value of time-domain signal,and the percentage of A7,D7,D6,D5,D4,D3 band energy in total energy after multi frequency decomposition.By studying the fuzzy pattern recognition method,the fuzzy membership function model based on cutting temperature signal is established.The percentage of energy in total energy in the frequency band of D3 + D4(1372-5520 Hz)is selected to match with the cutting temperature to determine the tool wear state.The 7-9-3 BP neural network model is constructed.The energy percentage and mean value of cutting temperature of six frequency intervals(A7,D7,D6,D5,D4,D3)obtained by multi frequency decomposition are used as input vector after normalization and tool wear state as output vector to realize the recognition of tool wear state.Through the combination of simulation analysis,experimental research and theoretical analysis,this paper establishes a tool wear monitoring scheme based on cutting audio and temperature signal fitting.Compared with single signal monitoring system,the system can accurately identify the tool wear state,and the cost of temperature and audio monitoring equipment is low,which has great economic and application value,and provides a new method for tool real-time monitoring.