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

Convolutional neural network deep learning-based environmental corrosion severity recognition method

A convolutional neural network and deep learning technology, applied in the field of testing and evaluation of electrical service environment, can solve the problems of high cost and low efficiency, and achieve the effect of reducing cost and shortening recognition time.

Active Publication Date: 2022-02-18
CHINA NAT ELECTRIC APP RES INST
View PDF8 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method has the disadvantages of high cost, low efficiency, and over-specialized operation.

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
  • Convolutional neural network deep learning-based environmental corrosion severity recognition method
  • Convolutional neural network deep learning-based environmental corrosion severity recognition method
  • Convolutional neural network deep learning-based environmental corrosion severity recognition method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0048] The present invention discloses a method based on convolutional neural network depth learning environmental corrosion, which is characterized in that:

[0049] Step S1, data acquisition of multiple different electrical service environments, the data acquisition method of each electrical service environment is: placing the copper test piece in the electrical service environment, and conducts copper test pieces in multiple inspection time points Detection; with a total of N groups of real test data, each set of measured data contains an electrical service environment, a measured corrosion, and placed copper test piece placed in the electrical service environment photo data and service time. However, the photo data can characterize the corrosion state of the copper test piece, the service time is placed from the copper test piece in the electrical service environment to the time point, the photo data and the service time are normalized. deal with;

[0050] Step S2, the neural ...

Embodiment approach

[0059] like figure 1 As shown in step S2, the construction of the neural network depth learning model includes:

[0060] Step S2-1, build a photo data input module: Enter the photo data of the input neural network depth learning model, input to the convolution group of the pre-training convolutional neural network, and add a flatten layer in sequence after the convolution group, first The DENSE full connect layer and the DropOut layer are outputted image feature information; wherein the number of neurons and activation functions of the first DENSE full connection layer are set to 256 and the RELU function, and the Dropout layer has a Dropout ratio of 50%. That is, that is: Image feature data stretches into one dimensional data; the first DENSE full connection layer is used to extract the global information of the photo, and the role of the DROPOUT layer is to reduce the network over fitting.

[0061]Step S2-2, build a service time input module: Enter the service time of the input ...

Embodiment 3

[0081] Based on the above-described embodiments or of the first embodiment, the third embodiment also employs the following preferred embodiments:

[0082] The normalization process of the photo data is: decoding all of the photos of the copper test piece into an RGB pixel grid, and adjusts the height and width of each photo to the same size, and the height and The width is within 400 pixels, and each pixel value (0 to 255) of the photo is removed from 255, and each pixel value is normalized to the [0, 1] interval to obtain the N. Set photo data of data on 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 discloses an environmental corrosion severity recognition method based on convolutional neural network deep learning. The method comprises the following steps: carrying out the training through a neural network deep learning model, and building the incidence relation between the photo data and service time of a copper test piece and the actually measured corrosion severity level of an electric appliance service environment; an environment corrosion severity grade recognition model is obtained; and collecting photo data after the copper test piece is placed in the service environment of the target electric appliance for a preset service time, so that the predicted corrosion severity grade of the service environment of the target electric appliance can be quickly recognized through an environment corrosion severity grade recognition model. The application stage shown in the step S3 can be completed on site in the service environment of the target electric appliance, and the test sample does not need to be transported to a laboratory for testing, so that the time for recognizing the predicted corrosion severity grade of the service environment of the target electric appliance can be greatly shortened, and the cost for transporting the test sample and manually analyzing and judging is reduced.

Description

Technical field [0001] The present invention relates to testing and evaluation of electrical service environments, specifically a method of environmental corrosion, severe recognition method based on convolutional neural network depth learning. Background technique [0002] With the rapid economic development of China, China Electrical Appliances has moved to the world, as well as the implementation of high-quality development strategies in China, such as sea wind power, sea photovoltaic products, etc., are rapidly building, whether it is in the ocean islands, offshore land And other parts of the world, the metal components of electrical products will be corroded by the effects of air corrosion pollution substances, so that the performance and life of electrical products have decreased significantly, and even threatens the user's property and personal safety. Therefore, it is of great significance for the corrosion rate of electrical equipment and the environmental corrosion of e...

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): G06T7/00G06V10/764G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06T7/0004G06N3/08G06T2207/20081G06T2207/30164G06N3/047G06N3/045G06F18/2415Y04S10/50
Inventor 李淮彭煌向利祁黎赵雪茹揭敢新张晓东王俊
Owner CHINA NAT ELECTRIC APP RES INST
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com