Analysis of metrological characteristics of the thermovision technical diagnosis system containing a neural network

Abstract

The paper deals with the problem of determining the errors of the thermovision system of the electronic devices technical diagnostics using a neural network. The structure of the system is described, including measuring channels with a thermal imager, an external thermometer, a block of computational models of thermal states, a knowledge base and a software neural network thermogram analyzer. It is proposed to use for the analysis of device states the two-branch network consisting of the multilayer convolutional neural network and the fully connected network. A metrological model of measuring channels has been built. A classification of the components of the instrumental error is given and expressions for the multiplicative and additive components of the instrumental error are obtained. Particular attention is paid to the analysis of the methodological error caused by the use of an artificial neural network in the classification of failures in the device. Experimental studies were performed that confirmed the effectiveness of the proposed methodological and technical solutions.

About the authors

R. V. Girin

Samara State Technical University

Author for correspondence.
Email: romangirin@gmail.com
Russian Federation

References

Statistics

Views

Abstract: 67

PDF (Russian): 102

Dimensions

Article Metrics

Metrics Loading ...

PlumX

Refbacks

  • There are currently no refbacks.

Copyright (c) 2020 Samara State Technical University

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

This website uses cookies

You consent to our cookies if you continue to use our website.

About Cookies