Fault monitoring method based on multi-classification algorithm

The invention discloses a fault monitoring method based on a multi-classification algorithm, and the method comprises the following steps: obtaining monitoring parameters of a high-voltage circuit breaker in various states, and extracting feature vectors; s2, using the feature vectors extracted in the step S1, using current and voltage data at different times as output values, setting classification model parameters, using fault positions as feature vectors to be imported into a value classification parameter model, and generating a classification model; and S2, carrying out 1-5-level grading on the output of the plurality of algorithm models in the step S2, firstly marking the mechanical fault output in the single classification algorithm model as G, marking the program fault as C, dividing the grade of the mechanical fault into 1-3 levels, and carrying out 1-3-level classification according to the position of the fault generation position and the fault generation frequency. According to the fault monitoring