VIBROACOUSTICAL SIGNALS TO EVALUATE TOOL PROFILES ON FRICTION STIR WELDING PROCESS

Abstract

This paper presents a study to evaluate the relationship between the tool design and vibroacoustical signals of friction stir welding process in the aluminum AA1050. The characterization in time and frequency domains of vibroacoustical signal and the statistical analysis have been carried out in order to correlate them with the design of two different tools. Vibroacoustical signals have been filtered in order to eliminate the noise of the friction stir welding machine. Additionally, the wavelet transform has been used as an efficient tool for filtering vibration signal waveforms. The statistical analysis has confirmed that the vibroacoustical signals were significantly affected by the tool design. It was found that in the frequency range of 0 to 100Hz there exists valid information to evaluate changes in the tool profile. Analysis techniques based on wavelet transform constitute a tool that can be efficiently employed in the evaluation of changes in the tool profile.