| The punching die parts’ dimensional accuracy is high,and the application is very wide,which has become one of the pillar industries in the equipment manufacturing industry.With the rapid development of modern industry,the performance requirements of the punching die is also increasing.The quality of parts is impacted by the quality of the die.At present,it is difficult to scientific pre-judge the condition of punching die,usually the only thing to do is fault repair and regular maintenance of the die for the repair strategy.In this regard,this article summing up the previous application of acoustic emission technique,based on genetic algorithm-wavelet packet of acoustic emission signal processing,combined with the possibility of fuzzy math theory,the status determine of die was implemented,provide a theoretical foundation for the die-oline detection of die,which filled the gap.Its main contents and results can be seen as follow:First of all,In this paper,the basic theory such as the mechanism of acoustic emission signal generation,the characteristics and the impact factors of acoustic emission signal were introduced,combined with the type of acoustic emission signals, in accordance with the work mechanism and failure mechanism of die,the signal characteristics of die was analyzed.According to the current analysis method of characteristics signal of acoustic emission and its application status,the processing program of the acoustic emission signal of punching die was proposed.Secondly,the continuous wavelet transform,inverse wavelet transform,as well as the nature of it were guided through the introduction of the basic principles of wavelet analysis,and then the theory of wavelet packet transform was described in detail which used in this study,the wavelet basis and the largest wavelet scale decomposition were selected according to die,by which the wavelet packet energy feature vector extraction is obtained.And then,through the analysis of the basic genetic algorithm theory,the characteristics parameter optimization methods was discussed,with the application of statistical theory,the corresponding fitness function was designed,the technology of characteristic parameters automatic generation was analyzed,by which the wavelet packet energy feature vector was optimized.Thirdly,the basic method of fuzzy pattern recognition was given to discuss the recognition of the probability distribution function,diagnosis through fuzzy analysis, with the possibility theory,the mapping relationship between the fault characteristics and fault model was set up,the way to determine the status of the punching die was researched.Finally,to confirm the feasibility of the study,the experiment of determine the status of punching die was done.The acoustic emission signal of failure and normal die were tested,with the help of fault diagnosis software to deal with and analysis the acquisition signal.Experimental results show that in the identification of different states the GA characteristic parameters which more than 99%recognition rate be able to generated for the successive diagnosis,with the possibility theory of fuzzy reasoning,non-stationary signal can be diagnosed,and be able to determine the status in high precision.The results show that the frequency rang of acoustic emission signals is much wider than the frequency range of the mechanical vibration and noise that affected by processing conditions small,the use of wavelet packet and genetic algorithm can improve the resolution of status characteristic parameter,the membership function established by probability density function and the possibility can carry out to distinguish the state of punching die. |
PDF Full Text Request