Abstract: An effective approach was developed to extract diagnostic information used to identify expulsion from electrode force curve by using wavelets analysis method for aluminum alloy shock wave resistance spot welding.The irregular signal singularity of electrode force curve was detected by using db5 wavelet.The detection result shows that the location and intensity of the signal singularity can be detected accurately from high frequency reconstructed signal of wavelet decomposition structure.That is to say, the expulsion can be identified easily.The diagnostic information is the global maximum value of high frequency reconstructed signal.The recognition method is that if the global maximum value exceeds the threshold value built by analyzing a number of testing data, the expulsion will occur.The testing result proves that the method is correct, reliable and credible.Therefore, the signal characteristic of electrode fore curve can be transformed into numerical characteristic that can be identified by computer.