Online monitoring method for flow state of acoustic emission technology in flow channel of water-turbine generator set

The invention discloses an online monitoring method for the flow state of a flow channel of a water-turbine generator set through an acoustic emission technology, and relates to the technical field of data encryption, comprising the following steps: collecting acoustic emission signals under different working conditions through an acoustic emission sensor, and analyzing an obtained noise spectrogram by using short-time Fourier transform; compensating the noise spectrogram through an adaptive noise contour compensation model to obtain a de-noising enhanced spectrogram; performing frequency domain feature extraction on the de-noised enhanced spectrogram to obtain morphological features; and setting a threshold value based on the information gain characteristic analysis result, comparing the threshold value with the real-time morphological characteristics to define the working state of the current water-turbine generator set flow channel, and performing real-time adjustment and maintenance. According to the meth