KR20240031447A - Condition based maintenance prediction system using image big data for vehicle body sub-units in railway vehicles - Google Patents

Abstract

The present invention relates to a state-based maintenance prediction system using image big data for a lower device of a railway vehicle body. The state-based maintenance prediction system using image big data for a lower device of a railway vehicle body comprises: a vehicle recognition unit recognizing a railway vehicle entering a detection position; a wheel detection unit photographing a wheel provided in a lower portion of a vehicle body of the railway vehicle to provide wheel detection data; a braking component detection unit photographing a braking component provided in the lower portion of the vehicle body to provide braking component detection data; and a collection detection unit providing a detection control signal to the wheel detection unit and the braking component detection unit when the entry of the railway vehicle is recognized through the vehicle recognition unit, receiving the wheel detection data to determine whether the wheel is damaged by using a laser profile and a 3D image, and receiving the braking component detection data to determine whether the braking component is damaged by using an area scan image. Therefore, a defect with respect to the railway vehicle component can be effectively detected, and a failure or the like can be predicted.

Description

Condition-based maintenance prediction system using video big data for railway vehicle underbody devices {CONDITION BASED MAINTENANCE PREDICTION SYSTEM USING IMAGE BIG DATA FOR VEHICLE BODY SUB-UNITS IN RAILWAY VEHICLES}

The present invention provides detection control signals to the wheel detection unit and braking component detection unit when the entry of a railway vehicle is recognized through the vehicle recognition unit, and receives wheel detection data from the resin detection unit using a laser profile and 3D image. By determining whether there is damage to the wheels, receiving braking part inspection data, and determining whether there is damage to braking parts using area scan images, it is possible to effectively detect abnormalities in parts provided in the lower part of the railway car body. This is about a condition-based maintenance prediction system using video big data for railway vehicle underbody devices that can not only effectively detect defects in parts, but also predict and predict failures.

As is well known, railway vehicles are used as a major means of transportation not only because they can transport large quantities of cargo or passengers, but also because they are highly safe. Due to their transport scale, these railway vehicles use various techniques to reduce the possibility of accidents occurring. After detection, countermeasures must be established.

Since these railway vehicles must drive at a planned speed in a specific section, the speed must be reduced or stopped at an accurate location as necessary. For this purpose, a braking device may be installed on the wheels provided at the bottom of the railway vehicle body. .

Conventionally, various brake systems have been applied to railway vehicles, but the mechanical brake system is largely divided into a tread brake system that brakes by pressing a wheel on the wheel tread of a railway vehicle, and a disc brake system that brakes by pressing a brake pad on the brake disc. , It can be classified into a drum brake system in which the brake drum expands by the operation of the brake lining (disc lining) and applies pressure to the wheel.

As described above, the wheels of railway vehicles and the wheel elements, disk pads, and disc linings provided on the wheels are subject to a lot of damage during the braking process of the railway vehicle. Damage to these parts is periodically done for the safe operation of the railway vehicle. It has to be inspected.

In addition, continuous inspection and detection of parts provided in railway vehicles, such as wheels, wheel rollers, disk pads, and disc linings, are essential for the safe operation of railway vehicles, and technology development for more efficient inspection and detection is essential. This is being demanded.

1. Korean Patent No. 10-0388535 (registered on June 10, 2003)

The present invention provides detection control signals to the wheel detection unit and braking component detection unit when the entry of a railway vehicle is recognized through the vehicle recognition unit, and receives wheel detection data from the resin detection unit using a laser profile and 3D image. By determining whether there is damage to the wheels, receiving braking part inspection data, and determining whether there is damage to braking parts using area scan images, it is possible to effectively detect abnormalities in parts provided in the lower part of the railway car body. We aim to provide a condition-based maintenance prediction system using video big data for railway vehicle underbody devices that can not only effectively detect defects in parts, but also predict and predict failures.

The purposes of the embodiments of the present invention are not limited to the purposes mentioned above, and other purposes not mentioned will be clearly understood by those skilled in the art from the description below. .

According to an embodiment of the present invention, a vehicle recognition unit that recognizes a railroad vehicle entering the inspection location; A wheel inspection unit that provides wheel inspection data by photographing the wheels provided on the lower body of the railway vehicle; A braking component inspection unit that photographs braking components provided in the lower part of the vehicle body and provides braking component inspection data; And when the entry of the railway vehicle is recognized through the vehicle recognition unit, a detection control signal is provided to the wheel inspection unit and the braking component detection unit, respectively, and the wheel inspection data is provided and wheel damage is detected using a laser profile and 3D image. Condition-based maintenance prediction using image big data for the lower body of a railway vehicle, including a collection and detection unit that determines whether the braking parts are damaged and determines whether the braking parts are damaged using the area scan image by receiving the braking part inspection data. A system may be provided.

In addition, according to an embodiment of the present invention, when the wheel detection data is provided from the wheel detection unit, the collection and detection unit compares the laser profile for the wheel with a reference wheel laser profile to determine whether there is a shape abnormality for the wheel. A condition-based maintenance prediction system using image big data for the lower body of a railway vehicle that detects scratches and cracks on the wheel by comparing the 3D image with the reference wheel infrared image can be provided.

In addition, according to an embodiment of the present invention, when the braking part detection data is provided from the braking part detection unit, the collection and detection unit compares the area scan image with the reference area scan image to determine the thickness and uneven wear of the braking part. A condition-based maintenance prediction system using video big data for railway vehicle underbody devices that detects abnormalities in braking parts, including falling out, can be provided.

In addition, according to an embodiment of the present invention, when the braking part detection data is provided from the braking part detection unit, the collection and detection unit uses a flood fill algorithm in 3D image data including the braking part. A condition-based maintenance prediction system using image big data for railway vehicle underbody devices that detects and separates braking component images can be provided.

In addition, according to an embodiment of the present invention, the collection and detection unit measures the length of the braking part to detect whether the braking part is abnormal after performing depth calibration on the separated braking part image. A condition-based maintenance prediction system that utilizes video big data for the underbody devices of railway vehicles can be provided.

In addition, according to an embodiment of the present invention, the vehicle recognition unit is a railroad car equipped with at least one car number recognition sensor capable of recognizing the car number of the railway car and at least one detection sensor that detects the entry of the railway car. A condition-based maintenance prediction system using image big data for vehicle underbody devices can be provided.

In addition, according to an embodiment of the present invention, the wheel detection unit includes at least one first area scan camera for capturing a wheel image of the vehicle wheel, and at least one laser profile sensor for measuring the laser profile for the vehicle wheel. and image big data for an underbody device of a railway vehicle that generates and provides the wheel detection data by including at least one infrared camera that captures the infrared image and a first light that irradiates light to an area where the wheel is provided. A condition-based maintenance prediction system utilizing can be provided.

In addition, according to an embodiment of the present invention, the braking component detection unit includes at least one second area scan camera for capturing a braking component image for the braking component, and a second area scan camera for irradiating light to an area where the braking component is provided. 2 A condition-based maintenance prediction system using image big data for the underbody device of a railway vehicle that generates and provides the braking component inspection data with lighting can be provided.

In addition, according to an embodiment of the present invention, the wheel inspection data and braking component inspection data detected by the condition-based maintenance prediction system are constructed as big data in response to vehicle face number recognition, and the machine is used using an artificial intelligence algorithm. A condition-based maintenance prediction system using video big data for railway vehicle underbody devices that learns and predicts abnormal wear and failure of vehicle parts can be provided.

The present invention provides detection control signals to the wheel inspection unit and component inspection unit when the entry of a railroad vehicle is recognized through the vehicle recognition unit, and receives wheel inspection data from the resin detection unit and uses a laser profile and 3D image to detect the vehicle. By determining whether there is damage to the wheels, receiving braking part inspection data, and determining whether there is damage to braking parts using area scan images, it is possible to effectively detect abnormalities in parts provided in the lower part of the railway car body. Not only can defects in parts be effectively detected, but also failures can be predicted and predicted.

Figure 1 is a block diagram of a condition-based maintenance prediction system using image big data for railway vehicle underbody devices according to an embodiment of the present invention;
Figure 2 is a diagram conceptually illustrating a condition-based maintenance prediction system using image big data for railway vehicle underbody devices according to an embodiment of the present invention;
Figures 3 to 9 are diagrams for explaining a condition-based maintenance prediction system using image big data for railway vehicle underbody devices according to an embodiment of the present invention;
10 to 13 are diagrams for explaining detection of damage to braking parts according to an embodiment of the present invention.

Advantages and features of the embodiments of the present invention and methods for achieving them will become clear by referring to the embodiments described in detail below along with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. The present embodiments are merely provided to ensure that the disclosure of the present invention is complete and to be understood by those skilled in the art. It is provided to fully inform those who have the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numerals refer to like elements throughout the specification.

In describing embodiments of the present invention, if a detailed description of a known function or configuration is judged to unnecessarily obscure the gist of the present invention, the detailed description will be omitted. The terms described below are terms defined in consideration of functions in the embodiments of the present invention, and may vary depending on the intention or custom of the user or operator. Therefore, the definition should be made based on the contents throughout this specification.

Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

Figure 1 is a block diagram of a condition-based maintenance prediction system using image big data for a railway vehicle underbody device according to an embodiment of the present invention, and Figure 2 is a block diagram of a condition-based maintenance prediction system for a railroad vehicle underbody device according to an embodiment of the present invention. It is a drawing conceptually illustrating a condition-based maintenance prediction system using image big data, and Figures 3 to 9 show condition-based maintenance prediction using image big data for the lower body device of a railway vehicle according to an embodiment of the present invention. It is a drawing for explaining the system, and Figures 10 to 13 are drawings for explaining whether or not a braking part is damaged according to an embodiment of the present invention.

Referring to Figures 1 to 13, the condition-based maintenance prediction system utilizing image big data for the underbody device of a railway vehicle according to an embodiment of the present invention includes a vehicle recognition unit 110, a wheel detection unit 120, and a braking system. It may include a parts detection unit 130, a collection and detection unit 140, etc.

The vehicle recognition unit 110 recognizes a railroad car entering the inspection location and is equipped with at least one car number recognition sensor capable of recognizing the car number of a railway car and at least one detection sensor that detects the entry of a railway car. It can be.

For example, as shown in FIG. 2, the vehicle recognition unit 110 may have a car sign recognition sensor and a detection sensor installed toward the entry direction of the railway vehicle on the path along which the railway vehicle moves along the rail, and the detection sensor may be installed as shown in FIG. When the entry of a railroad car is detected, the car number of the train can be detected through the car sign recognition sensor and the recognition signal can be provided to the collection and control unit 140.

Here, the maximum speed at which the railway vehicle enters is approximately 25 km/h, and the car designation recognition sensor and detection sensor can recognize and detect the entry and car registration of the railway car within the maximum speed.

The wheel inspection unit 120 provides wheel inspection data by photographing the wheels provided on the lower part of the vehicle body of the railway vehicle, and includes at least one first area scan camera that captures a wheel image of the wheels, and a laser profile for the wheels. Wheel detection data can be generated and provided by providing at least one laser profile sensor that measures, at least one infrared camera that takes infrared images, and a first light that irradiates light to the area where the wheel is provided.

For example, the wheel detection unit 120 can provide wheel detection data by photographing the wheel according to the detection control signal provided from the collection and detection unit 140. As shown in FIGS. 2 to 4, the railway vehicle On the path moving along this rail, a laser profile sensor can be installed in the center of the rail, a first area scan camera and an infrared camera can be installed on both sides, and a wheel provided on the lower part of the railway vehicle body. First lights that irradiate light to the portion may be installed in the center portion and both sides, respectively.

Here, the first area scan camera may include a 2D camera, a 3D camera, etc., and the first lighting may include an LED light, an infrared light, etc.

Accordingly, the wheel detection unit 120 generates a trigger signal when a detection control signal is provided from the collection and detection unit 140, and detects the wheels provided at the bottom of the car body of the incoming railway vehicle using a first area scan camera and a laser profile sensor. And after photographing and detecting through an infrared camera, the wheel detection data can be provided to the collection and detection unit 140. At this time, the first illumination may be emitted according to the time corresponding to imaging and detection through the first area scan camera, laser profile sensor, and infrared camera.

The measurement error of the first area scan camera and laser profile sensor as described above is ±0.5 mm, and its reliability is 98%, which can provide high detection performance for wheel shape measurement, and the wheel tread file is 10mm × 10mm ( Defects with a depth of 1 mm or more can be detected, and the reliability is 98%, providing high detection performance for wheel scratch measurement.

The braking component inspection unit 130 provides braking component inspection data by photographing the braking components provided in the lower part of the vehicle body, and provides braking component images for braking components (e.g., brake pads, disc pads, disc linings, etc.). Braking part detection data can be generated and provided by providing at least one second area scan camera that takes pictures and a second light that irradiates light to the area where the braking part is provided.

For example, the braking parts detection unit 130 can photograph braking parts according to the detection control signal provided from the collection and detection unit 140 and provide braking part detection data. As shown in FIG. 2, the railroad car On the path moving along this rail, a second area scan camera may be installed in the center of the rail to photograph and detect the disk pad, and a second area scan camera may be installed on both sides to photograph and detect the driver, respectively. In addition, second lights that irradiate light to the wheel portion provided in the lower body of the railway vehicle may be installed in the center and on both sides, respectively.

Here, the second area scan camera may include a 2D camera, a 3D camera, etc., and the second lighting may include an LED light, an infrared light, a laser module, a strobe light, etc.

Accordingly, the braking parts detection unit 130 generates a trigger signal when a detection control signal is provided from the collection and detection unit 140, and uses a second area scan camera to detect the disk pad and wheel control provided at the bottom of the car body of the incoming railway vehicle. After filming and inspection, the braking component inspection data can be provided to the collection and detection unit 140. At this time, the second light may be emitted according to the time corresponding to the shooting and detection through the second area scan camera.

The measurement error of the second area scan camera as described above is ±1 mm, and its reliability is 98%, which can provide high detection performance for measuring damage to braking parts, including thickness, uneven wear, and fall-off of braking parts. .

When the entry of a railroad vehicle is recognized through the vehicle recognition unit 110, the collection and detection unit 140 provides a detection control signal to the wheel inspection unit 120 and the brake component detection unit 130, respectively, and provides wheel inspection data. You can determine whether there is damage to the wheel using the laser profile and 3D image provided, and you can determine whether there is damage to the braking part using the area scan image by receiving the braking part inspection data.

When the wheel detection data is provided from the wheel detection unit 120, the collection and detection unit 140 compares the laser profile for the wheel with the reference wheel laser profile, and compares the wheel image for the wheel with the reference wheel image to determine the wheel. It is possible to detect shape abnormalities and detect scratches and cracks on the wheel by comparing the infrared image with the reference infrared image of the wheel.

For example, when wheel detection data is provided from the wheel detection unit 120, the collection and detection unit 140 can compare the laser profile for the wheel with the reference wheel laser profile as shown in FIG. 5, and through this, the wheel Shape abnormalities can be detected, and the wheel diameter wear data can be extracted and stored by comparing the wheel image with the reference wheel image, as shown in Figure 6 according to the wheel data request from the user terminal (or administrator terminal). As shown, the wheel diameter wear progress data can be generated in a pop-up window and displayed through a display device connected to the user terminal.

In addition, the collection and detection unit 140 can compare the 3D image with the reference wheel infrared image to extract wheel scratch data and store it, as shown in FIG. 7 according to the wheel data request from the user terminal (or administrator terminal). Trend chart data for each vehicle can be created in a pop-up window and displayed through a display device connected to the user terminal.

By statistically providing such wheel diameter wear progress data and vehicle-specific trend chart data according to railroad vehicles, management efficiency for the wheels of railroad vehicles can be improved.

In addition, when braking part detection data is provided from the braking part inspection unit 130, the collection and detection unit 140 compares the area scan image with the reference area scan image to determine the thickness of the brake part, uneven wear, and separation of the brake part. Any abnormalities can be detected.

For example, the collection and detection unit 140 compares the area scan image with the reference area scan image and detects brake parts including thickness, uneven wear, and separation of the wheel and disk pad as shown in Figures 8 and 9. Damage can be detected.

Meanwhile, when braking part detection data is provided from the braking part detection unit 130, the collection and detection unit 140 uses a flood fill algorithm to detect and detect braking part images from 3D image data including braking parts. It can be separated, and after depth calibration is performed on the separated braking part image, the length of the braking part can be measured to detect whether the braking part is abnormal.

For example, as shown in Figure 10, when braking component detection data is provided, the collection and detection unit 140 normalizes and pre-processes 3D image data from the corresponding detection data, and uses the flood fill algorithm to obtain the data in Figure 11. As shown in , the inspection target (i.e., brake parts such as wheel control, disk pad, etc.) can be separated from the entire input image. Here, the floodfill algorithm can separate the inspection target by recognizing the value within a preset range as one object based on one point as a standard.

In addition, through depth correction, it is possible to correct based on the depth information of the inspection object that is distorted by the sense of principal, and through this, the number of pixels can be calculated at the location where the dimensions (i.e., thickness, etc.) are to be measured in the detected object. The length can be measured by calculating. That is, as shown in FIG. 12, the thickness of the black portion of the wheel control element can be measured, and as shown in FIG. 13, the thickness of the black portion of the disk pad can be measured.

As described above, it is possible to detect whether the wheel control wheel and disc pad are damaged by comparing the measured thickness of the wheel control wheel and disc pad with the reference thickness value.

The detection accuracy of the vehicle underbody detection device according to the embodiment of the present invention as described above is shown in Table 1 below.

In other words, the vehicle body underbody detection device according to the embodiment of the present invention detects the height and width of the corresponding part of the cross section for the profile related to the wheel shape, but the measurement error is within ±0.5 mm, and the reliability is as high as 98%. It can provide detection performance, and in the case of friction surfaces related to wheel scratches, surface defects can be detected, and the crack depth can be detected over 1 mm and the length over 3 mm, and the reliability is 98%. It can provide high detection performance in %.

In addition, in the case of wheel elements and disc pads, damage including thickness, uneven wear, and fall-off is detected, but the measurement error is within ±1 mm and the reliability is 98%, providing high detection performance.

In the embodiment of the present invention as described above, the wheel inspection data and braking component inspection data detected by the condition-based maintenance prediction system are constructed as big data in response to vehicle face number recognition, and machine learning is performed using an artificial intelligence algorithm. This makes it possible to predict and predict abnormal wear and failure of vehicle parts.

Here, the artificial intelligence algorithm may include, for example, a CNN (convolution neural network) algorithm. The CNN algorithm is a useful algorithm for finding images, and learns images directly from data and classifies images using patterns. It has the advantage of being able to effectively process images by learning while maintaining the spatial information of the image.

Therefore, according to an embodiment of the present invention, when the entry of a railway vehicle is recognized through the vehicle recognition unit, a detection control signal is provided to the wheel detection unit and the braking component detection unit, respectively, and the wheel detection data is provided from the resin detection unit to create a laser profile. By determining whether there is damage to the wheels using infrared images and receiving braking part inspection data and determining whether there is damage to braking parts using area scan images, it effectively detects any abnormalities in parts provided in the lower part of the railway car body. It is possible to effectively detect defects in railway vehicle parts, as well as predict and predict failures.

In the above description, various embodiments of the present invention have been presented and explained, but the present invention is not necessarily limited thereto, and those skilled in the art will understand various embodiments without departing from the technical spirit of the present invention. It will be easy to see that branch substitutions, transformations, and changes are possible.

110: Vehicle recognition unit 120: Wheel detection unit
130: Braking parts inspection unit 140: Collection detection unit

Claims (9)

A vehicle recognition unit that recognizes railway vehicles entering the inspection location;
A wheel inspection unit that provides wheel inspection data by photographing the wheels provided on the lower body of the railway vehicle;
a braking component inspection unit that photographs braking components provided in the lower part of the vehicle body and provides braking component inspection data; and
When the entry of the railway vehicle is recognized through the vehicle recognition unit, a detection control signal is provided to the wheel inspection unit and the braking component inspection unit, and the wheel inspection data is provided to determine whether or not the wheel is damaged using a laser profile and 3D image. a collection and detection unit that determines whether the braking part is damaged by receiving the braking part inspection data and using an area scan image;
Condition-based maintenance prediction system using video big data for railway vehicle lower body equipment, including.
In claim 1,
The collection and detection unit,
When the wheel detection data is provided from the wheel detection unit, the laser profile for the wheel is compared with the reference wheel laser profile to detect shape abnormality for the wheel, and the 3D image is compared with the reference wheel infrared image. Detecting scratches and cracks on the wheel
Condition-based maintenance prediction system using video big data for railway vehicle underbody devices.
In claim 2,
The collection and detection unit,
When the braking component inspection data is provided from the braking component inspection unit, the area scan image is compared with the reference area scan image to detect abnormalities in the braking component, including thickness, uneven wear, and detachment of the braking component.
Condition-based maintenance prediction system using video big data for railway vehicle underbody devices.
In claim 3,
The collection and detection unit,
When the braking component detection data is provided from the braking component detection unit, the braking component image is detected and separated using a flood fill algorithm from the 3D image data including the braking component.
Condition-based maintenance prediction system using video big data for railway vehicle underbody devices.
In claim 4,
The collection and detection unit,
After performing depth calibration on the separated braking part image, the length of the braking part is measured to detect whether the braking part is abnormal.
Condition-based maintenance prediction system using video big data for railway vehicle underbody devices.
In claim 5,
The vehicle recognition unit,
At least one car number recognition sensor capable of recognizing the car number of the railway vehicle and at least one detection sensor that detects the entry of the railway car are provided.
Condition-based maintenance prediction system using video big data for railway vehicle underbody devices.
In claim 6,
The wheel inspection unit,
At least one first area scan camera for capturing a wheel image of the wheel, at least one laser profile sensor for measuring the laser profile for the vehicle wheel, and at least one infrared camera for capturing the infrared image, Generating and providing the wheel detection data by providing a first light that radiates light to the area where the wheel is provided.
Condition-based maintenance prediction system using video big data for railway vehicle underbody devices.
In claim 7,
The braking parts inspection department,
At least one second area scan camera for taking a braking part image of the braking part, and a second light for irradiating light to an area where the braking part is provided, to generate and provide the braking part inspection data.
Condition-based maintenance prediction system using video big data for railway vehicle underbody devices.
The method according to any one of claims 1 to 8,
The wheel inspection data and braking component inspection data detected by the condition-based maintenance prediction system are constructed as big data in response to vehicle face number recognition, and machine learning is performed using an artificial intelligence algorithm to prevent abnormal wear and failure of vehicle parts. predicting about
Condition-based maintenance prediction system using video big data for railway vehicle underbody devices.