Parameter-adaptive fast iterative filtering decomposition for fault diagnosis of rolling bearings in spinning frames

Abstract Rolling bearings in spinning frames are essential components that frequently develop faults because of the severe operating conditions, including high-speed rotation and fluctuating loads. In response to these challenges, this study proposes a parameter-adaptive fast iterative filtering decomposition (PA-FIFD). The PA-FIFD technique is developed to adaptively optimize decomposition parameters (i.e. mask length) according to signal characteristics, thereby enabling the precise extraction of fault-induced impulsive components from complex noise backgrounds. Experiments using real-world vibration data from spinning frame bearings validate the superior performance of the proposed method. Compared to several existing methods, the proposed approach achieves superior accuracy in feature extraction. Moreover, the framework demonstrates remarkable computational efficiency, making it highly practical for industrial condition monitoring.