Power station unit fault diagnosis method based on signal decomposition and weighted decision

The invention discloses a power station unit fault diagnosis method based on signal decomposition and weighted decision, and the method comprises the steps: S1, deploying an intelligent inspection robot: deploying the intelligent inspection robot in an operation region of a hydropower station unit; s2, collecting signals through a multi-sensor system of the inspection robot; step S3, a VMD signal decomposition step: a plurality of acquired signals of noise, vibration and temperature are decomposed by using a Variation Mode Decomposition signal decomposition technology; s4, a spectral analysis step: performing spectral analysis on each modal function of the vibration signal obtained by VMD decomposition; s5, a feature extraction step: aiming at a spectrum analysis result; the hydroelectric generating set fault diagnosis and analysis method has the advantages that noise, vibration and temperature signals are collected at the same time through an intelligent inspection robot of a multi-sensor system, and accurat