Characteristic extraction of acoustic signals emitted from resistance spot welding process based on independent component analysis

In this paper, independent component analysis(ICA)and multisensor array technology were used to research the acoustic emission signals from resistance spot welding.An independent component analysis algorithm for the acoustic signals from spot welding was established.A multisensor array system with 8 microphones was used to measure the acoustic emission signals from spot welding synchronously.According to the calculation of the kurtosis coefficient, the acoustic signals were found to be non-Gaussian distribution.So the acoustic signals were decomposed by the independent component analysis.The results showe that the independent component analysis based on Fast ICA algorithm can extract the characteristic signals of the spot welding from superimposed noises obviously, which provides a new technology for the profound analysis of spot welding.Further more, this algorithm can be used in other welding methords.