Cyclostationary Signal Analysis for Fault Diagnosis in Rotating Machinery

Rotating machinery such as bearings, gearboxes and pumps underpin critical sectors from manufacturing to energy generation. Faults in these components often manifest as periodic impulses embedded in vibration or current signals. Such signals are inherently non-stationary but exhibit statistical properties that vary periodically over time, a characteristic known as cyclostationarity. By exploiting cyclic features, cyclostationary signal analysis can unmask early fault indicators concealed by noise or operating variabilities. Core techniques include computation of cyclic autocorrelation functions and cyclic spectra, which reveal energy concentration at characteristic fault frequencies and their harmonics. Advances in algorithmic efficiency and noise resilience have extended practical deployment to real-time condition monitoring and predictive maintenance. This approach delivers global benefits: improving reliability, reducing unscheduled downtime and optimising lifecycle costs in industries ranging from wind energy to automotive manufacturing.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Cyclostationary Signal Analysis for Fault Diagnosis in Rotating Machinery publication trend

The graph below shows the total number of articles in cyclostationary signal analysis for fault diagnosis in rotating machinery across all publications each year (not limited to Nature Index journals).

Technical terms

Cyclostationarity: A property of signals whose statistical moments vary periodically over time, often arising from repetitive mechanical actions.

Cyclic spectral density: A frequency-domain function describing how signal energy at one frequency correlates with energy at another offset by a cycle frequency.

Spectral correlation: A measure of cyclic frequency components obtained by correlating spectral slices of a signal, used to detect periodic modulations.

Sparse representation: A signal modelling approach that represents data using a small number of basis functions or atoms, enhancing feature extraction and denoising.

References

1. Fast Spectral Correlation Based on Sparse Representation Self‐Learning Dictionary and Its Application in Fault Diagnosis of Rotating Machinery. Complexity (2020).
2. A noise-resistant Wigner-Vile spectrum analysis method based on cyclostationarity and its application in fault diagnosis of rotating. Journal of Vibroengineering (2020).
3. Methods of Hidden Periodicity Discovering for Gearbox Fault Detection. Sensors (2021).
4. Cyclostationary Analysis towards Fault Diagnosis of Rotating Machinery. Processes (2020).