Motor group fault diagnosis method based on binary deep neural network

A deep neural network and fault diagnosis technology, applied in the field of training fault diagnosis model using binary deep neural network, fault diagnosis of large-scale motor groups, and real-time fault diagnosis of large-scale motor groups, which can solve the fault diagnosis of large-scale motor groups and other problems to achieve the effect of saving forward propagation time, reducing costs, and saving data storage space

Inactive Publication Date: 2018-11-13
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF6 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is in order to overcome the defective of prior art, in order to solve the fault diagnosis

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Motor group fault diagnosis method based on binary deep neural network
  • Motor group fault diagnosis method based on binary deep neural network
  • Motor group fault diagnosis method based on binary deep neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0054] The present invention uses the proposed method and the mainstream deep neural network and traditional machine learning methods to carry out fault diagnosis experiment verification on a certain type of motor. Table 1 shows part of the attribute data. The data consists of four types of faults, each of which includes 850 training samples and 120 testing samples.

[0055] Table 1 Partial fault attribute data

[0056]

[0057] First, column standardize the data, reduce all data to [-1,1], and then perform One-Hot encoding on the output data according to the fault type. For example, currently there are five types of fault data, the One-Hot code of the first type of fault is [1,0,0,0], and the One-Hot code of the third type of fault is [0,0,1,0].

[0058] Secondly, the principal component analysis method is used to reduce the dimensionality of the original 97-dimensional fault data, and the eigenvalues ​​with a cumulative contribution rate of 85% are taken to reduce the da...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a motor group fault diagnosis method based on a binary deep neural network. In the method, a motor fault diagnosis model based on the binary deep neural network is trained through motor fault historical data, and the motor fault diagnosis model is deployed on an edge computing device for hoping to reduce real-time data I/O transmission load and large-scale deployment costof the edge computing device, and meanwhile, the method has higher diagnosis accuracy. With the method provided by the invention, under the premise of guaranteeing having the diagnosis accuracy comparable to that of a full precision deep neural network model, mass data transmission needed for a fault real-time diagnosis process is avoided, meanwhile, the binary deep neural network is used for substituting the full precision deep neural network at an edge end, large amounts of mathematical operations are changed into bit operations, lots of data storage space and forward transmission time are saved, cost for deploying the edge devices in large scale is reduced indirectly, and an effect of high precision, fast diagnosis and low deployment cost is realized.

Description

technical field [0001] The invention belongs to the technical field of industrial fault diagnosis, and relates to a fault diagnosis method for a large-scale motor group, in particular to a fault diagnosis model that uses a binary deep neural network to train a fault diagnosis model, and deploys the model on an edge computing device for large-scale A method for real-time fault diagnosis of motor groups. Background technique [0002] Most electrical equipment has a complex structure, complex and changeable working environment and working conditions, tight coupling and mutual influence between parts and components, and its faults are characterized by diversity, concealment, uncertainty, and complex causality. The current data-driven fault diagnosis technology is developing rapidly, which can make full use of fault sample data, combined with data mining and artificial intelligence methods that match the fault data, to mine various complex nonlinear relationships and potential co...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/62G06N3/08
CPCG06N3/084G06F18/2135G06F18/24
Inventor 李慧芳胡光政赵蕾蕾
Owner BEIJING INSTITUTE OF TECHNOLOGYGY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap