Short-time Fourier transform for fault detection of rolling bearings

Fault diagnosis includes vibration method, noise method, temperature method, etc. Vibration method is often used as the main means of equipment monitoring and diagnosis. Traditional Fast Fourier Transform (FFT) provides assistance for fault diagnosis, while Short Time Fourier Transform (STFT) has more significant advantages in bearing vibration signal analysis. It can simultaneously display both time and frequency domains, accurately capture transient characteristics and local frequency changes, and effectively address the complexity and non stationarity of bearing fault signals. This article extracts and pre-processes horizontal and vertical vibration signals, calculates the mean, root mean square value and kurtosis through time-domain analysis, and plots them; Generate spectrograms using STFT in the frequency domain to pinpoint the characteristic fault frequencies for diagnosing fault types; Subsequently, the signal envelope will be extracted through envelope analysis, and the envelope spectrum will be calculated to further assist in diagnosis.