A crayfish optimised wavelet filter and its application to fault diagnosis of machine components

Industrial machinery relies heavily on accurate fault diagnosis to maintain its reliability and operational efficiency. However, analyzing vibration signals for early fault detection is complex, as these signals often suffer from low signal-to-noise ratios, background noise, and random interferences. While wavelet filters are frequently employed for identifying informative frequency bands, determining the optimal filter parameters is crucial for accurately extracting repetitive transients related to faults. This research proposes a novel approach that utilises a crayfish optimization algorithm (COA) to adaptively optimise the wavelet filter. The COA employs correlated kurtosis (CK) as a fitness function to guide the optimization process. This method addresses limitations related to inaccurate CK period estimation through a continuous update mechanism, ensuring more precise parameter selection. Through its application to various industrial cases, the proposed methodology demonstrates its superiority over existing techniques in accurately extracting informative frequencies. This advancement enables more precise fault diagnosis, leading to improved maintenance strategies and minimized downtime in industrial environments.