Source Separation method of Machine Faults Based on Post-Nonlinear Blind Source Separation

The existing separation of machine fault sources is only restricted to the linear mixing source signals. However, the observed signals do not suffice to this condition in the practical application. In view of this deficiency, a new mechanical source separation method based on nonlinear blind source separation (BSS) is proposed. The proposed method is compared with the linear BSS method of machine faults. The experiment result shows that the proposed method is very effective for nonlinear mixing mechanical sources. The linear BSS method of machine faults is only a special case of the proposed method. When source separation method based linear BSS cannot give a satisfactory separation performance, the nonlinear BSS method should be considered.