Research on Fault Diagnosis Method of Rolling Bearings Based on BKA-VMD-SVM

Rolling bearings are essential in mechanical systems and are crucial in industrial production across many sectors. However, extended operation and harsh environments often cause bearing failures, resulting in potential safety risks. Advancing fault diagnosis technology is essential to ensure equipment operates safely. To tackle this challenge, methods combining black winged kite algorithm optimization with variational mode decomposition and support vector machine are proposed. Black winged kite algorithm optimizes variational mode decomposition by determining the best parameters for decomposition count and penalty factors. Intrinsic mode functions from the decomposition are screened using a composite metric of envelope spectrum kurtosis and correlation coefficients, enabling reconstruction and noise suppression. Features are extracted from the processed signals, and black winged kite algorithm further optimizes support vector machine parameters to develop a rolling bearing fault diagnosis model. Experimental results show that the support vector machine optimized by the black winged kite algorithm improves classification accuracy in rolling bearing fault diagnosis, demonstrating its effectiveness in diagnosing faults.