Fault diagnosis method and system for small samples of power equipment based on virtual-real twin space

Abstract

The invention discloses a small-sample fault diagnosis method and system for power equipment based on a virtual-real twin space, which belongs to the field of fault diagnosis of power equipment. Accurate solid space; according to the simulation model of the equipment to be diagnosed, the fuzzy mirror space is obtained by simulation; the training set of twin space is obtained by space hybridization between the training set of precise solid space and the sample set of fuzzy mirror space; the training sample set of twin space is used as the precise solid space As a supplement to the training set, the fault type and location are used as diagnostic labels, which are input into the deep convolutional neural network for training, and the fault identification and location are performed on the accurate entity space verification set to verify the diagnostic effect. The invention utilizes power equipment simulation to enhance the data of measured samples, can give full play to the auxiliary function of power simulation in diagnosis practice, and effectively improves the accuracy rate in the case of small sample diagnosis.

Description

technical field [0001] The invention belongs to the field of fault diagnosis of electric equipment, and more specifically relates to a fault diagnosis method and system for a small sample of electric equipment based on a virtual-real twin space. Background technique [0002] With the development of modern equipment toward high precision, intelligence, and high integration, the manufacturing industry has continuously increased requirements for equipment reliability and operational safety. However, at present, the ability of intelligent diagnosis of power equipment faults is weak, and the lack of fault samples of power equipment further limits the development of intelligent diagnosis of equipment. However, the fault destructive tests on transformers to obtain fault samples required for diagnostic research cannot be realized on a large scale, either from the perspective of economy or safety. Digital twin technology can use real-world data to construct a digital space model of ...

AI Technical Summary

Problems solved by technology

However, at present, the ability of intelligent diagnosis of power equipment faults is weak, and the lack of fault samples of power equipment further limits the development of intelligent diagnosis of equipment

However, the fault-destructive testing of transformers to obtain fault samples required for diagnostic research cannot be realized on a large scale, either from the perspective of economy or safety.

Images