Swing alarm judgment method for hydraulic turbine set

The invention discloses a swing alarm judgment method for a hydraulic turbine set which comprises the following steps: calculating a modulus value under an inherent frequency, namely a throw amplitude value of a large shaft at a measured point, if the throw amplitude value is smaller than an amplitude threshold I, judging that the hydraulic turbine set is normal, and if a read amplitude value is greater than or equal to the amplitude threshold I, executing the next judgment; and judging the type of a fault which has occurred or is about to occur in the hydraulic turbine set based on the swing frequency spectrum of the hydraulic turbine set. Signal data of a complete period can be obtained by a variable sampling rate method, disordered continuous time domain signals are decomposed into discrete regular frequency domain signals by a fast Fourier transform method, and only effective swing amplitude data are reserved and interference signals are removed through spectral analysis. The main throw characteristic of