Deep learning-based rail wagon bearing oil shedding fault detection method

A technology for fault detection and railway wagons, applied in neural learning methods, image data processing, biological neural network models, etc., can solve problems completed by artificial naked eyes, achieve reliable fault detection results, high real-time performance, and solve difficult-to-obtain problems Effect

Pending Publication Date: 2022-04-19
SOUTHEAST UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Purpose of the invention: The purpose of the present invention is to solve the problem that the existing railway freight car bearing oil throwing fault identification can only be completed by human eyes, and proposes a railway freight car bearing oil throwing fault detection method based on deep learning, which saves manpower and material resources , which is convenient for railway dynamic inspectors to quickly troubleshoot

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
  • Deep learning-based rail wagon bearing oil shedding fault detection method
  • Deep learning-based rail wagon bearing oil shedding fault detection method
  • Deep learning-based rail wagon bearing oil shedding fault detection method

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0045] A deep learning-based detection method for bearing oil rejection faults of railway wagons, comprising the following steps:

[0046] Step S1: Use a high-speed camera to obtain the whole vehicle image of the railway freight car when it is running. There are multiple parts such as side frame, middle part I, middle part II, coupler and coupler. Use the image naming rules to filter the parts, and only select the parts containing bearings. The image of the side frame part, the size of the image is 1400×1024×3.

[0047] Step S2: Use the Hough circle transform to detect the circle on the image of the above-mentioned side frame, and filter the interfering circle by setting different radius thresholds, so as to locate the bearing area of ​​the side frame, and cut the bearing area to form a fault detection The training set of the network. The radius threshold finally adopted by the present invention is 150-180 pixels.

[0048]Step S3: Based on the research of deep anomaly detect...

specific Embodiment approach 2

[0073] The difference between this embodiment and specific embodiment 1 is that the abnormality discrimination mechanism adopted in step S5 is reconstruction abnormality discrimination mechanism, and the abnormality score is expressed as:

[0074]

[0075] The reconstruction discriminative mechanism represents the difference between the input image and the reconstructed image. First, the input image is reconstructed through the encoder Encoder1 and the decoder Decoder, and the difference between the input image and the reconstructed image is calculated, and then normalized to obtain the abnormal score.

specific Embodiment approach 3

[0076] The difference between this embodiment and specific embodiment 1 is that the abnormality discrimination mechanism adopted in step S5 is a fusion abnormality discrimination mechanism, and the abnormality score is expressed as:

[0077] A t =A r +αA e

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 rail wagon bearing oil shedding fault detection method based on deep learning, and the method comprises the following steps: obtaining a bearing region image, and constructing a training set; building a bearing oil shedding fault detection network; the training set is preprocessed and sent to the fault detection network in batches for training, and a final fault detection model is obtained by adjusting parameters; and acquiring a to-be-detected image, processing the to-be-detected image, inputting the to-be-detected image into the fault detection model, and calculating an abnormal score to obtain a bearing oil shedding fault detection result. The automatic detection method is high in accuracy and real-time performance, and solves the problem of false detection and missing detection caused by visual fatigue due to the fact that faults can only be judged by identifying images through naked eyes of dynamic vehicle inspectors at the present stage. On the basis of the deep convolutional neural network, a structure of an auto-encoder and a generative adversarial network is adopted, and the combination of the auto-encoder and the generative adversarial network can reconstruct a specifically distributed picture more naturally, which has stronger feature extraction capability for fault detection.

Description

technical field [0001] The invention relates to the technical field of image recognition and fault detection of railway freight cars, in particular to a deep learning-based detection method for oil throwing faults of railway freight car bearings. Background technique [0002] With the development of the national economy and the continuous improvement of the demand for railway transportation, the scale of railway freight continues to expand, and the failure of railway freight cars directly affects the safety of train operation. As an important part of the side frame of railway freight cars, rolling bearings are prone to failures such as oil throwing and loss of gear keys, which will affect the running of the train in the slightest and cause traffic accidents in the worst case. [0003] Usually, the fault detection of railway wagons is accomplished through the combination of dynamic vehicle inspectors and high-speed camera capture systems. Dynamic vehicle inspectors check the...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/00G06T7/12G06V10/25G06V10/774G06N3/04G06N3/08
CPCG06T7/0004G06T7/12G06N3/08G06T2207/20084G06T2207/20081G06T2207/30268G06T2207/20132G06N3/045G06F18/214
Inventor 杨绿溪步兆军张颀李春国黄永明
Owner SOUTHEAST UNIV
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