Propeller cavitation degree identification method based on pulse frequency characteristic mode identification

Abstract

The invention discloses a propeller cavitation degree identification method based on pulse frequency characteristic mode identification. The method comprises the following steps: (1) acquiring a propeller noise signal; (2) importing the noise signal into a program, and calculating by using a rapid cyclic stationary characteristic function to obtain a cyclic density spectrum; (3) normalizing to obtain a cyclic coherence spectrum, and then further integrating and averaging to construct an enhanced envelope spectrum under logarithmic coordinates; (4) judging the characteristic frequency accordingto the obtained enhanced envelope spectrum, selecting integral multiples of the corresponding time period, and carrying out improved time domain averaging on the source data; (5) performing ensembleempirical mode decomposition to obtain a corresponding intrinsic mode function; (6) detecting and counting the pulse frequency of the intrinsic mode function by adopting a constant false alarm rate; and (7) taking the pulse frequency as a characteristic matrix, and performing BP neural network training identification to obtain judgment of a cavitation state. By utilizing the method, the statistical characteristics of the propeller in different cavitation states can be shown, and the estimation of the obtained state is more accurate.

Description

technical field [0001] The invention belongs to the field of signal processing and feature extraction, and in particular relates to a propeller cavitation degree identification method based on pulse frequency feature pattern identification. Background technique [0002] Cavitation is a common problem in propeller operation. The generation and development of cavitation not only affects the velocity distribution in the flow channel, deteriorates the working condition of the propeller and reduces its efficiency, but also affects its dynamic response. Long-term cavitation may also seriously damage the impeller and other flow-passing components. Finding appropriate fault diagnosis and identification methods is of great significance for effectively controlling propeller cavitation. [0003] Cavitation state identification is one of the difficulties in propeller state monitoring. At present, the commonly used rotating machinery fault detection methods in the field of signal proces...

AI Technical Summary

Problems solved by technology

[0005] Fault detection methods such as Fourier transform, short-time Fourier transform, wavelet transform, second-generation wavelet transform and multi-wavelet transform are all based on the assumption that the signal is a stationary signal, but in reality it is often a non-stationary signal, so these detection methods are all something unreasonable, unrealistic

On the other hand, due to theoretical limitations, these traditional detection methods are difficult to detect some important features of rotating machinery, such as blade passing frequency BPF, blade ratio frequency BRF, etc., which have great limitations

Images