Rolling bearing fault diagnosis method based on symbiotic organisms search

Bearings frequently operate under extreme conditions such as high temperatures, high speeds, and heavy loads, which complicates the detection of their vibration signals. This complexity often undermines the precision and reliability of traditional fault diagnosis methods. While techniques such as Empirical Mode Decomposition (EMD) and Variational Mode Decomposition (VMD) have proven somewhat effective for analyzing non-stationary signals, they are impeded by issues like sensitivity to noise, mode mixing, and the necessity to set hyperparameters. In particular, VMD requires the pre-determination of two hyperparameters: the penalty factor and the number of modes, where finding optimal values typically involves laborious experimentation, thereby hindering the method’s efficiency. To address these challenges, this paper introduces an innovative approach that enhances the optimization of VMD parameters using a combination of the Genetic Algorithm (GA) and the Symbiotic Organisms Search (SOS) algorithm. This synergy aims to automate and optimize the extraction of bearing fault features efficiently. Subsequently, the features are classified using a Support Vector Machine (SVM), significantly improving the precision and efficiency of fault diagnosis.