Abstract:

For early detection of boilers potential failure in power plant the high noise background fault self\|adaptive detection system is studied. A higher background noise fault signal detection algorithm is provided. Simulation experiment results, and points out that the efficiency of the method. The traditional cepstrum analysis is applied to signal based on FFT (Fast Fourier Transformation) transform, and the algorithm is based on the improvement CZT (Chirp Z\| inverse Transform) transform, get the failure characteristics of curve more stable, and more reliable. The experimental results show that this method is fast and effective, fault quoted accuracy over 99%.

Key words: high noise, adaptive detection, chirp z-transform(CZT), fast fourier transformation(FFT)algorithm, cepstrum analysis