Support vector machine mechanical fault diagnosis method based on voting mechanism

Abstract

The invention discloses a support vector machine mechanical fault diagnosis method based on a voting mechanism. The method comprises three different feature analysis algorithms: power frequency harmonic surface feature analysis algorithm, a multi-resolution band-pass filter tank spectrum analysis method and Hilbert-huang conversion feature analysis method; the feature extraction is performed on asound signal emitted under the mechanical equipment fault operation state, and then classification identification is performed by utilizing the support vector machine algorithm. The adopted three-feature analysis algorithms have good applicability for the mechanical noise and like unstable nonlinear signal, and the time frequency feature of the signal can be self-adaptively described in multiple aspects; the classification algorithm adopts a support vector machine multi-classifier based on the voting mechanism; the classifier decomposes the multi-classifier into multiple second-class classifiers by adopting a one versus rest method, and the classification method can achieve high identification rate and good algorithm robustness based on the voting mechanism and the confidence optimal criterion judgment method.

Description

technical field [0001] The invention belongs to the technical field of mechanical fault diagnosis, in particular to a support vector machine mechanical fault diagnosis method. Background technique [0002] Gas Insulated Switchgear (GIS) has been more and more widely used due to its high operational reliability. GIS has good insulation properties and is widely used in high-voltage transmission systems. With the rapid development of my country's electric power industry, GIS tends to develop with large capacity and high voltage. As an important equipment in the power system, once GIS fails, it will affect the normal power supply of the power system, resulting in huge economic losses and adverse social impact. Therefore, the operational reliability of GIS is very important. [0003] The failure rate of GIS is generally only 20%-40% of conventional equipment, but there are also defects that are not easy to find, there are few testing methods, and it is difficult to judge. Ope...

AI Technical Summary

Problems solved by technology

The technology of mechanical fault diagnosis based on GIS equipment noise has not been found in the relevant research literature so far.

Images