Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Gear reliability analysis system based on countermeasure network generated by boundary constraint

A boundary constraint, analysis system technology, applied in special data processing applications, instruments, electrical digital data processing, etc., can solve problems such as inability to guarantee learning distribution efficiency, prone to network loopholes, and inability to classify

Active Publication Date: 2019-01-15
CHONGQING TSINGSHAN IND +1
View PDF3 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Second, the data generated using GAN has no classification labels in the instances and cannot be directly used in classifiers
Although using instances in a class as input to a GAN can label the generated data, it cannot guarantee the efficiency of learning assignments due to the inadequacy of each class
Third, in traditional GANs, gradient descent is used to update the weights and biases of the two networks, which is prone to convergence difficulties due to the non-convex nature of the objective function
Furthermore, discriminators use strict objectives to identify raw and synthetic data, which is prone to network vulnerabilities

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
  • Gear reliability analysis system based on countermeasure network generated by boundary constraint
  • Gear reliability analysis system based on countermeasure network generated by boundary constraint
  • Gear reliability analysis system based on countermeasure network generated by boundary constraint

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment 1

[0039] Such as Figure 1-Figure 2 As shown, this embodiment discloses a gear reliability analysis system based on boundary constraint generation confrontation network, including: raw data sampling module, generation confrontation network module, mean covariance labeling module and classifier, wherein:

[0040] Raw data sampling module: used to obtain gear raw parameters and normalize them; the data collected in this example are 85 categories of gear parameter data, including gear tooth load, application factor K A , the internal dynamic factor K v , tooth width load factor K Hβ , K Fβ , End load factor K Hα , K Fα , meshing load factor K γ ;

[0041] Generative confrontation network module: including a generator and a discriminator, a boundary constraint module is also set between the generator and the discriminator, and the generator is based on the normalized original parameter D 0 and noise Z to generate data D G , data D G After being processed by the boundary con...

specific Embodiment 2

[0047] The difference in this embodiment is that the boundary constraint module uses a partial constraint model, such as image 3 (b), specifically described as:

[0048] Y=Y p +Y m ;

[0049]

[0050] Y i m =X i m ;

[0051] will generate data D G The parameter types in are divided into X p and x m Two parts, new data D B corresponds to Y in p and Y m , X i p means X p A parameter vector of the i-th type in X i m means X m The i-th type of parameter vector in Y i p means Y p The i-th type of parameter vector in Y i m means Y m A parameter vector of the i-th type in A p for X p Corresponding coefficient matrix, and 0≤A j ≤1, δ p for X p The corresponding deviation vector, Indicates that the corresponding elements in the vector are multiplied.

specific Embodiment 3

[0052] The difference in this embodiment is that the boundary constraint module adopts a multi-constraint model, such as image 3 As shown in (c), the specific description is:

[0053]

[0054]

[0055] will generate data D G The parameter types in are divided into X h and x l Two parts, in the new data D B corresponds to Y in h and Y l , X i h means X h A parameter vector of the i-th type in X i l means X l The i-th type of parameter vector in Y i h means Y h The i-th type of parameter vector in Y i l means Y l A parameter vector of the i-th type in A h for X h The corresponding coefficient matrix, and δ h for X h The corresponding deviation vector, A l for X l The corresponding coefficient matrix, and δ l for X l The corresponding deviation vector, Indicates that the corresponding elements in the vector are multiplied.

[0056] By constructing a specific system model for simulation verification, the system can generate more instance data ...

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 provides a gear reliability analysis system based on boundary constraint generation antagonism network, which is characterized by comprising an original data sampling module, a generation antagonism network module, a mean covariance marking module and a classifier, wherein, the original data sampling module is used for acquiring the original parameters of the gear and normalizing theoriginal parameters; the generating countermeasure network module comprises a generator and a discriminator, and a boundary constraint module is arranged between the generator and the discriminator.A mean covariance mar module is used for mar that final generated data of the generated antagonistic network module; classifier: selecting part of the original parameters and part of the generated data after tagging as the final composite data set for security classification. The effect is that more instance data can be generated and effectively marked, and the data generated by constrained GAN has better reliability classification ability.

Description

technical field [0001] The invention relates to mechanical equipment reliability analysis technology, in particular to a gear reliability analysis system based on boundary constraint generation confrontation network. Background technique [0002] The transmission gear provides the main transmission power for the vehicle and regulates the speed and direction of the vehicle's motion. Due to the complexity of the repair process of the gear, its maintenance and downtime costs are high. More than 60% of transmission failures are caused by gears, and gear reliability is a key indicator for evaluating gear safety. Therefore, accurate classification of the safety and reliability of gear parameters is the key to ensure the transmission operation. [0003] However, the lack of gear data with full-scale parameters leads to a high misclassification rate. In practice, gear parameters are obtained through four steps. First, initialize the gear parameters with empirical values. Second...

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): G06F17/50G06K9/62
CPCG06F30/17G06F30/20G06F18/24
Inventor 利节孙宇姜艳军龚为伦刘垚何宏黎陈瑶
Owner CHONGQING TSINGSHAN IND
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products