Penetration State Recognition Based On The Double-sound-sources Characteristic Of VPPAW And Hidden Markov Model

With the advantages of high-energy-density,strong-penetration-ability,low-cost and so on,VPPAW(variable polarity plasma arc welding)has been widely used in aluminum alloy structural parts welding of aerospace industry.Plasma arc and weld pool interact on each other because of the special ‘Keyhole Effect' in PAW,which makes the keyhole quite sensitive to weld parameters and the quality of VPPAW is of uncertainty to some extent.Penetration control is one of the most critical issues of the quality control in VPPAW.In this article,a series of work is conducted to study the relationship between the welding sound and penetration states,which are No-keyhole,Keyhole and Cutting Modes.The HMM(Hidden Markov Model)penetration-states-recognition model,based on the special doublesound-sources characteristic of VPPAW,is built.Firstly,the spatial distribution of sound mean energy is measured and the optimal acquisition angle is determined as 150°.Then,the mechanism,based on the special double-sound-source characteristic of VPPAW,is proposed and demonstrated by a series of phenomena about sound of VPPAW,which shows the sound,caused by the second sound source,consists of pulse component and turbulence component.The characteristic illustrates that the source of power,structure of sound source and sound generating mechanism are quite similar to those of speech.According to the similarity,MFCC(Mel-Frequency Cepstral Coefficients)and HMM,the common technology in the field of speech recognition,are employed in VPPAW sound characterization and penetration state recognition respectively.A two-steps recognition strategy,adapted to VPPAW conditions,is proposed,which means one HMM recognition model,based on characteristic parameter in time domain,is built to figure out No-keyhole State and then the other HMM recognition model,based on MFCC,is built to distinguish the other two states.The recognition rate could reach up to 99.6%.The work shows that the acoustic signal of VPPAW is similar to speech to some extent,which establishes the theoretical basis for the application of speech process technology in VPPAW,and the strategy is also practical,especially for MFCC and HMM.