In the field of process monitoring, in particular vibration monitoring, the majority of researchers use hybrid forms between different fault diagnosis methods. These forms generally include three main components, namely fault detection, feature extraction and fault identification, which sometimes require a lot of calculation. I really want to know the usefulness of hybrid methods by comparing with simple methods (such as: Wavelet analysis, Empirical mode decomposition, Minimum entropy deconvolution .... etc.) that are fast and efficient.