Markov Modeling of Signal Condition Transitions for Bearing Diagnostics Under External Interference Conditions

In real-world scenarios, the performance of rolling bearings can be significantly affected by frequent environmental disturbances, including high-energy fluctuations, transient noises, unintended collisions, or sudden variations in loading. These disturbances have the potential to obscure signs of damage, posing a challenge to the identification of failures. Traditional techniques for selecting demodulation band encounter challenges due to inherent complexity and parameter sensitivity. These methods, which heavily rely on postprocessing signal analysis involving interference components, prove ill-suited for complex scenarios, especially faced with multivariate random pulse noise. To address these challenges, this article proposes a method for Markov modeling of signal condition transitions to mitigate transient noise interference at its source. This method aims to improve insensitivity for external transient noise, enabling a more accurate and reliable selection of demodulation bands. The proposed method employs a signal transition model grounded in a Markov transition matrix, coupled with wavelet transforms and reconstruction. It improves the identification of anomalous signal components and leverages the transition matrix to track temporal state changes. In addition, an amplitude interference-limiting mechanism is designed to identify and mitigate transient noise that may adversely affect the demodulation band selection process. The study’s results demonstrate that this mechanism allows typical Kurtogram techniques to accurately identify bearing fault-characteristic bands, and the experimental results validate the effectiveness of the proposed methodology for fault diagnosis of rolling bearings.