KR102120128B1 - System for monitoring vehicle manufacturing processes - Google Patents

Abstract

According to an embodiment of the present invention, a system for monitoring a vehicle manufacturing process, which is capable of detecting defective parts by monitoring production status of mass-produced automobile parts, comprises: a vehicle manufacturing process monitoring apparatus which detects a defect of each type based on the process data collected from the vehicle manufacturing process including manufacturing of parts and assembly through the parts for each part of the vehicle, and outputs the current state of the production lines where the defect is detected; and a management server which builds big data based on the history of the process data and the defect detection of each type, and provides critical value setting information connected to the defect detection based on the big data. The process data include the reflection light intensity of the light irradiated to the parts and the assembly unit based on the assembly between parts, the three-dimensional image of the parts and the assembly unit, and the weight of the parts and the assembly unit. The vehicle manufacturing process monitoring apparatus can determine the defect type based on at least one of the reflection light intensity, the three-dimensional image, and the weight.

Description

Vehicle manufacturing process monitoring system {SYSTEM FOR MONITORING VEHICLE MANUFACTURING PROCESSES}

The present invention relates to a vehicle manufacturing process monitoring system.

Industrial development was achieved as some industries, such as IT, semiconductor, shipbuilding, and automobiles, led exports, and the manufacturing industry served as a driving force for economic growth and exports. Automobiles, which are representative mechanical devices used throughout the industry today, have many parts built in, and different parts implement organic operation. In particular, in the case of an aggregate in which a large number of parts are combined, such as an automobile, each part must be manufactured in a predetermined size, strength, and shape in order to operate different parts organically. Each part of a car manufactured through mass production is manufactured according to a predetermined standard. However, if an inadvertent product is manufactured and distributed or assembled inadvertently, it is impossible to operate organically between each part, resulting in damage or accident, and this may affect the reputation of the manufacturer.

The background technology of the present application is disclosed in Korean Patent Publication No. 10-2019-0009769.

An object of the present invention is to provide a vehicle manufacturing process monitoring apparatus and method capable of detecting defective parts by monitoring the production state of automobile parts that are mass-produced to solve the problems of the prior art described above.

However, the technical problems to be achieved by the embodiments of the present application are not limited to the technical problems as described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, a vehicle manufacturing process monitoring system according to an embodiment of the present application is based on process data collected in a vehicle manufacturing process including manufacturing of parts for each part of a vehicle and assembly through parts. Vehicle manufacturing process monitoring device that detects defects by type and outputs the status of the production line in which defects are detected, and builds big data based on the process data and the history of defect detection by type, based on the big data, And a management server providing threshold setting information associated with the defect detection, wherein the process data includes the amount of reflected light of light irradiated to the assembly unit based on assembly between the parts and parts, a three-dimensional image of the parts and the assembly unit, and The weight of the component and the assembly unit, and the vehicle manufacturing process monitoring device may determine the defect type based on at least one of the reflected light amount, the three-dimensional image, and the weight.

According to an embodiment of the present application, the vehicle manufacturing process monitoring apparatus, a sensor unit for collecting process data in a vehicle manufacturing process including the manufacturing and assembly of parts for each part of the vehicle, based on the process data It may include a failure analysis unit for detecting defects by defect type and a display unit for outputting the status of the production line in which the defect type and defects are detected.

According to one embodiment of the present application, the sensor unit is provided for each of the production line, the sensor unit, a light emitting sensor for irradiating light to the surface of the component and assembly unit, the component and assembly unit for the irradiated light It may include a light receiving sensor for receiving the reflected light reflected from the measuring the amount of reflected light, an image sensor for obtaining the three-dimensional image of the component and the assembly unit and a weight sensor for measuring the weight of the component and the assembly unit. .

According to one embodiment of the present application, the defect type may include injection failure, assembly failure, stain failure, crack failure, paint failure, and foreign substance failure.

According to an embodiment of the present application, the defect analysis unit detects the injection defect and the foreign substance defect based on the 3D image and the weight, and detects the assembly defect based on the reflected light amount and the 3D image, The spot defect may be detected based on the three-dimensional image, and the crack defect and the paint defect may be detected based on the amount of reflected light.

According to an embodiment of the present disclosure, the failure analysis unit may determine a response type for each defect type for the component and the assembly unit according to whether a predetermined step-by-step threshold value for each of the reflected light amount and the weight is exceeded.

According to one embodiment of the present application, the defect analysis unit calculates the similarity with the normal 3D image through the artificial neural network model using the 3D image as an input, and determines whether the preset step-by-step threshold for the similarity is exceeded. Accordingly, it is possible to determine a corresponding type for each defect type for the component and the assembly unit.

According to one embodiment of the present application, the corresponding type includes a first type that performs a close inspection on the detected defects of the parts and assembly units, and a repair and repair operation on the detected defects of the parts and assembly units. It can include two types and a third type that destroys the parts and assembly units.

According to one embodiment of the present application, the defect analysis unit receives threshold setting information from a management server of a vehicle manufacturing process, wherein the management server builds big data based on the process data and a history of defect detection by type, Based on the big data, the threshold setting information associated with the defect detection may be provided.

According to one embodiment of the present application, the defect analysis unit may update the threshold value for each of the reflected light amount and weight based on the threshold setting information.

According to one embodiment of the present application, the display unit may output numerical values of the reflected light amount, weight, and similarity exceeding the preset step-by-step threshold, and display the determined defect type on the parts and the assembly unit differently for each type. have.

The above-described problem solving means are merely exemplary and should not be construed as limiting the present application. In addition to the exemplary embodiments described above, additional embodiments may exist in the drawings and detailed description of the invention.

According to the above-described problem solving means of the present application, it is possible to provide a vehicle manufacturing process monitoring apparatus and method capable of detecting defective parts by monitoring the production state of automobile parts that are mass-produced.

1 is a view showing the configuration of a vehicle manufacturing process monitoring system according to an embodiment of the present application.
2 is a view showing the configuration of a vehicle manufacturing process monitoring apparatus according to an embodiment of the present application.
3 is a diagram illustrating an example of collecting process data of a vehicle manufacturing process monitoring apparatus according to an embodiment of the present application.
4 is a diagram illustrating an example of a defect type display of a vehicle manufacturing process monitoring apparatus according to an embodiment of the present application.
5 is a view showing a flow of a vehicle manufacturing process monitoring method according to an embodiment of the present application.

1 is a view showing the configuration of a vehicle manufacturing process monitoring system according to an embodiment of the present application.

Referring to FIG. 1, a vehicle manufacturing process monitoring system may include a vehicle manufacturing process monitoring device 100 and a management server 200. The vehicle manufacturing process monitoring device 100 detects defects by type based on process data collected in a vehicle manufacturing process including manufacturing of parts for each part of a vehicle and assembly through parts, and the status of a production line in which defects are detected Can output The management server 200 may build big data based on process data and a history of defect detection for each type, and provide threshold setting information associated with defect detection based on the big data. Specifically, the vehicle manufacturing process monitoring apparatus 100 detects a defect of an assembly unit assembled between parts and components on a production line, provides a countermeasure according to the defect, and the management server 200 is a reference for detecting the defect That is, it is possible to provide setting information of a grain boundary value that can be judged as defective.

2 is a view showing the configuration of a vehicle manufacturing process monitoring apparatus according to an embodiment of the present application.

Referring to FIG. 2, the vehicle manufacturing process monitoring apparatus 100 may include a sensor unit 110, a defect analysis unit 120, and a display unit 130. The sensor unit 110 may collect process data in a vehicle manufacturing process including manufacturing parts for each part of a vehicle and assembling through parts. The sensor unit 110 may be provided for each production line 10 of vehicle manufacturing to collect process data.

3 is a diagram illustrating an example of collecting process data of a vehicle manufacturing process monitoring apparatus according to an embodiment of the present application.

Referring to FIG. 3, the sensor unit 110 may include a light emitting sensor 111, a light receiving sensor 112, an image sensor 113, and a weight sensing sensor 114. The light emitting sensor 111 may irradiate light to the surface of the moving parts and the assembly unit on the production line 10, and the light receiving sensor 112 receives the reflected light reflected from the parts and the assembly unit for the irradiated light The amount of reflected light can be measured. The image sensor 113 can acquire a three-dimensional image of the component and assembly unit, and the weight detection sensor 114 can measure the weight of the component and assembly unit. Information measured and obtained by each configuration of the sensor unit 110 may be included in the process data. That is, the process data may include the amount of reflected light of the light irradiated with the assembly unit based on the assembly between parts and components, the three-dimensional image of the parts and assembly units, and the weight of the parts and assembly units.

The vehicle is made up of various parts. In the production of finished cars, not only domestic parts but also overseas parts are used, but it is a trend to localize the parts required for possible vehicles. Therefore, in order to produce a more complete vehicle, it is necessary to detect defective parts and assembly units from the production stage of each part. Referring to FIG. 3, process data corresponding to the fuel injector 1 may be collected through each configuration of the sensor unit 110.

Referring to FIG. 3, the light emitting sensor 111 may irradiate light toward a component or an assembly unit. Hereinafter, parts will be described as an example for convenience of description. The light emitting sensor 111 may irradiate light having a predetermined wavelength, such as infrared light or visible light. The light emitting sensor 111 may be, for example, an LED (LED) as a light emitting device. However, the present invention is not limited thereto, and the light emitting sensor 111 may be formed of various light emitting devices. The light receiving sensor 112 may be arranged to be spaced apart from the light emitting sensor 111 at a predetermined interval. The reflected light reflected from the component corresponding to the light irradiated toward the component of the light receiving sensor 112 may be received. As illustrated in FIG. 3, the light emitting sensor 111 and the light receiving sensor 112 may be provided as a pair and provided in plural. Therefore, it is possible to detect a more accurate defect through reflected light for various angles.

The light receiving sensor 112 may measure (sensing) the intensity of the reflected light changed by the crack defect and the coating defect described later in comparison with the intensity of the light emitted from the light emitting sensor 111 as the degree of reflected light. That is, when light is irradiated from the light emitting sensor 111 toward the component, the light receiving sensor 112 may measure (sensing) the intensity of reflected light compared to the intensity of light emitted from the light emitting sensor 111 as the degree of reflected light. . Through the degree of reflected light measured by the light receiving sensor 112, it is possible to determine whether a component is defective or painted.

The defect analysis unit 120 may detect defects for each defect type based on process data. For example, the defect analysis unit 120 may determine a defect type based on at least one of the amount of reflected light, the 3D image, and the weight. The defect type is a classification of parts and components that may occur in the assembly unit, and the defect type may include injection failure, assembly failure, stain defect, crack defect, paint defect, and foreign substance defect.

The defect analysis unit 120 may determine the type of defect by combining each of the reflected light amount, the 3D image, and the weight, or at least one or more. For example, the defect analysis unit 120 may detect an injection defect and a foreign object defect based on a three-dimensional image and weight (part, assembly unit). In the case of a defective injection, it may occur when the part is not cut completely during the injection process or the finishing process such as excessive cutting is not completed. Therefore, a difference in weight may occur compared to a normal component or an assembly unit, and a difference from a normal component may be found through a 3D image. In addition, a foreign body defect is a foreign substance attached to the surface of a component, and since the weight and the 3D image of the component may differ from a normal component or assembly unit, a foreign body defect can be detected through the 3D image and weight. have. The detection of the foreign material defect and the assembly defect, stain defect, crack defect, and coating defect described later may be judged step by step based on a predetermined threshold value for each process data or similarity using an artificial neural network model. The criteria for judging such a stepwise defect will be described later.

The defect analysis unit 120 may detect assembly failure based on the amount of reflected light and the 3D image. Illustratively, the assembly failure is one of the defects that may occur in the assembly unit due to assembly between parts, and assembly failure may occur due to factors such as separation between parts and assembly that is not completely fastened. The amount of reflected light of the phosphorus assembly unit is different from the amount of reflected light of the normal assembly unit, and the three-dimensional image is also different from the normal assembly unit to detect assembly failure. For example, when a part of the assembly unit protrudes due to an assembly failure in which the assembly is not fully engaged, reflected light may be received as light is irradiated to the projected portion. At this time, the reflected light corresponding to the protruding part may be more or less than the reflected light of the corresponding part of the assembly unit normally assembled. Therefore, the failure analysis unit 120 may determine whether assembly failure has occurred in the corresponding assembly unit through reflected light on all parts of the assembly unit.

In addition, the defect analysis unit 120 may further utilize a 3D image of the assembly unit to detect a more accurate assembly defect. The defect analysis unit 120 may analyze a three-dimensional image of all surfaces of the assembly unit to detect assembly failure due to factors such as separation between parts and assembly that is not fully fastened. The failure analysis unit 120 may be detected as an assembly failure when both the threshold value of the 3D image and the threshold value of reflected light are exceeded. The threshold value for the 3D image is a threshold value for the similarity between the 3D image of the assembly unit to be judged and the normal 3D image of the normal assembly unit, and the threshold value of the reflected light amount may be a preset step-by-step threshold value. Determination of defects using these threshold values will be described later.

The defect analysis unit 120 may detect unevenness of parts and assembly units based on the 3D image. Since the three-dimensional image is a photograph of the surfaces of the parts and the assembly unit, the unevenness in the surface treatment process of the parts and the assembly unit can be detected through the three-dimensional image. In the manufacturing process of the parts and assembly units, there may be surface treatment processes such as chemical treatment, etching, and cleaning on the surfaces of the parts and assembly units. In this case, when the surface treatment is not properly performed, a stain defect in which a stain is present on the surface may occur, for example, chemicals remain. Therefore, the defect analysis unit 120 may detect a defect in the unevenness of parts and assembly units through a 3D image.

In addition, the defect analysis unit 120 may detect a crack defect and a paint defect based on the amount of reflected light of the light irradiated to the component and the assembly unit. Crack defects and paint defects can also be detected from the surfaces of parts and assembly units. In the case of a crack, a part or assembly unit may break or crack in the manufacturing process. The defect analysis unit 120 may detect a crack defect through the amount of reflected light for the crack. In addition, a defective coating may also cause a difference in the amount of reflected light in a portion where the coating is clogged or insufficient, so that a defective coating may be detected through the defective analysis unit 120.

The failure analysis unit 120 may determine a response type for each defect type for the component and the assembly unit according to whether a predetermined step-by-step threshold is exceeded for each amount of reflected light and weight. For example, the step-by-step threshold may have a higher threshold as the step increases. Therefore, the higher the step, the higher the threshold for detection. That is, it can be interpreted that the higher the level, the stronger the degree of defects in the parts and assembly units. In addition, the response type means a processing method for a component or assembly unit in which a defect has occurred, and the response type may be different according to a threshold value for each step. For example, when the step threshold is set to 3 steps, the step 2 threshold may be greater than the step 1 threshold and less than the step 3 threshold. According to the above, in the case of crack failure, it can be detected based on the amount of reflected light. For example, when the amount of reflected light for a certain assembly unit exceeds a threshold of 2 stages (less than 3 stages), the failure analysis unit 120 may determine a response type corresponding to stage 2 for the assembly unit.

In addition, the failure analysis unit 120 calculates the similarity with the normal 3D image through an artificial neural network (ANN) model that uses a 3D image as an input, and whether a preset step-by-step threshold for the similarity is exceeded According to this, it is possible to determine a corresponding type for each defective type for the component and the assembly unit. The artificial neural network may be constructed through learning to input a normal 3D image for each component and a normal 3D image for each assembly unit. The artificial neural network may be a deep learning artificial neural network, but is not limited thereto, and the similarity may be calculated through other artificial intelligence models. The defect analysis unit 120 may calculate the similarity with the normal 3D image for each component and assembly unit through an artificial neural network model that inputs a 3D image of the component and assembly unit on the production line 10. When the threshold value for each step according to the similarity is set to 3 steps, a relative difference in steps may occur as described above.

According to the foregoing, as in the case of detecting an injection defect and a foreign object defect based on a 3D image and a weight, a defect is detected in consideration of at least two of a 3D image, a weight, and a reflected light amount among defect types. The defect analysis unit 120 may determine that any one of the factors (3D image, weight, and reflected light amount) for determining the defect type exceeds the threshold value and is considered to be defective. In addition, when the threshold value of each step exceeded by each of the two factors for determining the defect type is different, a corresponding type may be determined based on exceeding the threshold value of a high step. Exemplarily, when the 3D image and the weight both exceed the threshold value, an injection failure is detected, the 3D image exceeds the 1st threshold, and the weight exceeds the 3rd threshold, corresponding to 3rd stage The type of response can be determined.

The response type is the first type that performs a detailed inspection on the detected defects of the parts and assembly units, the second type that performs repairs and repairs on the detected defects of the parts and assembly units, and the second type that destroys the parts and assembly units. It can include three types. The first type and the second type have room for reuse of the parts and assembly units, but in the case of the third type, the corresponding parts or assembly units may be destroyed by determining that the parts and assembly units are fatal defects.

The defect analysis unit 120 may receive threshold setting information from the management server 200 of the vehicle manufacturing process. The management server 200 may build big data based on process data and a history of defect detection by type, and provide the threshold setting information associated with the defect detection based on the big data. In addition, the failure analysis unit 120 may update the threshold values for each of the reflected light amount and the weight based on the threshold setting information.

Specifically, the management server 200 may collect process data for each component and assembly unit, and big data the history of defect detection according to the process data. Therefore, data for determining defects according to process data may be provided. The management server 200 may provide threshold setting information based on the big data. The threshold setting information is information for setting a threshold value that can be judged as defective, and accumulates information (reflected light amount, weight, and similarity according to a 3D image) when judged to be defective, resulting in a more sharp and accurate defect determination. It can be created to update the threshold to be done. Parts and assembly units may have errors in measurement by the sensor unit 110 depending on the position, posture, and arrangement on the production line 10. Accordingly, the result of performing the defect detection for all parts and the assembly unit is recorded, and reinforcement learning on thresholds for determining normal and defective based on information determined as normal and information determined as defective. As this is achieved, errors in judgment due to errors can be minimized, and more accurate judgment of defects can be made by establishing a strict criterion for defect judgment through an updated threshold.

4 is a diagram illustrating an example of a defect type display of a vehicle manufacturing process monitoring apparatus according to an embodiment of the present application.

The display unit 130 may output a defect type and a status of a production line in which the defect is detected. Referring to FIG. 4, the display unit 130 may output numerical values of reflected light amount, weight, and similarity exceeding a preset step-by-step threshold, and display the determined defect type on the parts and assembly units differently for each type. Therefore, the operator can easily identify whether the defect occurred in the current production line or the assembly unit, the cause of the defect, and the value judged to be defective, and can easily check the countermeasures against the defect.

5 is a view showing a flow of a vehicle manufacturing process monitoring method according to an embodiment of the present application.

The vehicle manufacturing process monitoring method illustrated in FIG. 5 is performed by the vehicle manufacturing process monitoring apparatus described with reference to FIGS. 1 to 4. Therefore, even if it is omitted below, the contents described with respect to the vehicle manufacturing process monitoring apparatus through FIGS. 1 to 4 may be applied to FIG. 5 as well.

Referring to FIG. 5, in step S510, the sensor unit 110 may collect process data in a vehicle manufacturing process including manufacturing parts for each part of a vehicle and assembling through parts. The sensor unit 110 may be provided for each production line 10 of vehicle manufacturing to collect process data. The sensor unit 110 may include a light emitting sensor 111, a light receiving sensor 112, an image sensor 113, and a weight sensing sensor 114. The light emitting sensor 111 may irradiate light to the surface of the moving parts and the assembly unit on the production line 10, and the light receiving sensor 112 receives the reflected light reflected from the parts and the assembly unit for the irradiated light The amount of reflected light can be measured. The image sensor 113 can acquire a three-dimensional image of the component and assembly unit, and the weight detection sensor 114 can measure the weight of the component and assembly unit. Information measured and obtained by each configuration of the sensor unit 110 may be included in the process data. That is, the process data may include the amount of reflected light of the light irradiated with the assembly unit based on the assembly between parts and components, the three-dimensional image of the parts and assembly units, and the weight of the parts and assembly units.

In step S520, the defect analysis unit 120 may detect a defect according to the defect type based on the process data. For example, the defect analysis unit 120 may determine a defect type based on at least one of the amount of reflected light, the 3D image, and the weight. The defect type is a classification of parts and components that may occur in the assembly unit, and the defect type may include injection failure, assembly failure, stain defect, crack defect, paint defect, and foreign substance defect. The defect analysis unit 120 may determine the type of defect by combining each of the reflected light amount, the 3D image, and the weight, or at least one or more. For example, the defect analysis unit 120 may detect an injection defect and a foreign object defect based on a three-dimensional image and weight (part, assembly unit). In the case of a defective injection, it may occur when the part is not cut completely during the injection process or the finishing process such as excessive cutting is not completed. Therefore, a difference in weight may occur compared to a normal component or an assembly unit, and a difference from a normal component may be found through a 3D image. In addition, a foreign object defect is a foreign substance attached to the surface of a component, and since the weight of the component and a 3D image may differ from a normal component or assembly unit, a foreign object defect can be detected through the 3D image and weight. have. The detection of the foreign material defect and the assembly defect, stain defect, crack defect, and coating defect described below may be judged step by step based on a predetermined threshold value for each process data or similarity using an artificial neural network model.

The defect analysis unit 120 detects an injection defect and a foreign object defect based on the 3D image and the weight, detects assembly failure based on the reflected light amount and the 3D image, and detects a defect due to the 3D image. It can detect and detect a crack defect and the coating defect based on the amount of reflected light. In addition, the failure analysis unit 120 may determine the response type for each defect type for the component and the assembly unit according to whether the preset threshold value for each of the reflected light amount and the weight is exceeded.

In addition, the defect analysis unit 120 calculates the similarity with the normal 3D image through the artificial neural network model using the 3D image as an input, and parts and assembly units according to whether the preset threshold for the similarity is exceeded. It is possible to determine the response type for each defect type. The corresponding type destroys the first type, which performs a detailed inspection on the detected defects of the parts and assembly units, and the second type, which repairs and repairs the detected defects of the parts and assembly units, and the parts and assembly units. It may include a third type. The first type and the second type have room for reuse of the parts and assembly units, but in the case of the third type, the parts or assembly units may be destroyed by determining that the parts and assembly units are fatal defects.

The defect analysis unit 120 may receive threshold setting information from the management server 200 of the vehicle manufacturing process. The management server 200 may build big data based on process data and a history of defect detection by type, and provide the threshold setting information associated with the defect detection based on the big data. In addition, the failure analysis unit 120 may update the threshold values for each of the reflected light amount and the weight based on the threshold setting information. Specifically, the management server 200 may collect process data for each component and assembly unit, and big data the history of defect detection according to the process data. Therefore, data for determining defects according to process data may be provided. The management server 200 may provide threshold setting information based on the big data. The threshold setting information is information for setting a threshold value that can be judged as defective, and accumulates information (reflected light amount, weight, and similarity according to a 3D image) when judged to be defective, resulting in a more sharp and accurate defect determination. It can be created to update the threshold to be done.

In step S530, the display unit 130 may output the status of the production line in which the defect type and the defect are detected. The display unit 130 may output numerical values of reflected light amount, weight, and similarity exceeding a preset step-by-step threshold, and display the determined defect type on the parts and assembly units differently for each type. Therefore, the operator can easily identify whether the defect occurred in the current production line or the assembly unit, the cause of the defect, and the value judged to be defective, and can easily check the countermeasures against the defect.

According to an embodiment of the present application, a vehicle manufacturing process monitoring method may be implemented in a program instruction form that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, or the like alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present invention, or may be known and available to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs, DVDs, and magnetic media such as floptical disks. -Hardware devices specifically configured to store and execute program instructions such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language code that can be executed by a computer using an interpreter, etc., as well as machine language codes produced by a compiler. The hardware device described above may be configured to operate as one or more software modules to perform the operation of the present invention, and vice versa.

The foregoing description of the present application is for illustration, and a person having ordinary knowledge in the technical field to which the present application belongs will understand that it is possible to easily change to other specific forms without changing the technical spirit or essential characteristics of the present application. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined form.

The scope of the present application is indicated by the claims below, rather than the detailed description, and it should be interpreted that all modifications or variations derived from the meaning and scope of the claims and equivalent concepts thereof are included in the scope of the present application.

100: vehicle manufacturing process monitoring device
110: sensor unit
111: luminescence sensor
112: light receiving sensor
113: image sensor
114: weight sensor
120: defect analysis unit
130: display unit
200: management server

Claims (11)

In the vehicle manufacturing process monitoring system,
A vehicle manufacturing process monitoring device that detects defects by type based on process data collected in a vehicle manufacturing process including manufacturing parts of each part of a vehicle and assembling through parts, and outputs a status of a production line in which defects are detected; And
And a management server for constructing big data based on the process data and a history of defect detection by type, and providing threshold setting information associated with the defect detection based on the big data.
The process data includes the amount of reflected light of light irradiated to the assembly unit based on the assembly between the parts and the parts, a three-dimensional image of the parts and the assembly unit, and the weight of the parts and the assembly unit,
The vehicle manufacturing process monitoring apparatus determines the defect type based on at least one of the reflected light amount, the three-dimensional image, and the weight,
The vehicle manufacturing process monitoring device,
A sensor unit that collects process data in a vehicle manufacturing process including manufacturing parts for each part of a vehicle and assembling through parts;
A defect analysis unit that detects defects by defect type based on the process data; And
And a display unit for outputting the status of the production line in which the defect type and defect are detected
The defective type,
Including poor injection, poor assembly, uneven stain, poor crack, poor painting, foreign material defect,
The defect analysis unit,
Determining the response type for each defect type for the component and the assembly unit according to whether or not the preset threshold value for each of the reflected light amount and the weight is exceeded,
The corresponding types include a first type that performs a precise inspection on detected defects of the parts and assembly units, a second type that performs repairs and repairs on the detected defects of the parts and assembly units, and the parts and assembly units. Includes a third type that destroys,
The defect analysis unit,
The similarity with the normal 3D image is calculated through the artificial neural network model using the 3D image as an input, and the response type for each defect type for the component and assembly unit is determined according to whether a predetermined step-by-step threshold is exceeded for the similarity Decide,
Threshold setting information is received from the management server of the vehicle manufacturing process,
The management server builds big data based on the process data and the history of defect detection by type, and provides the threshold setting information associated with the defect detection based on the big data,
The threshold value for each of the reflected light amount and weight is updated based on the threshold setting information,
The artificial neural network model
Reinforcement learning to adjust the threshold value for each step based on process data when it is determined to be normal for the part and assembly unit by the big data and process data when it is detected as defective,
The defect analysis unit
A vehicle manufacturing process monitoring system that updates the threshold value based on the reinforcement learning. delete According to claim 1,
The sensor unit is provided for each of the production line,
The sensor unit,
A light emitting sensor that irradiates light to the surface of the component and assembly unit;
A light receiving sensor for receiving the reflected light reflected from the component and the assembly unit for the irradiated light and measuring the reflected light amount;
An image sensor that acquires the three-dimensional image of the component and assembly unit; And
And a weight sensor for measuring the weight of the parts and the assembly unit. delete According to claim 1,
The defect analysis unit,
The injection defect and the foreign substance defect are detected based on the 3D image and the weight, the assembly failure is detected based on the reflected light amount and the 3D image, and the stain defect is detected based on the 3D image. And detecting the crack defect and the coating defect based on the reflected light amount. delete delete delete delete delete According to claim 1,
The display unit,
A vehicle manufacturing process monitoring system that outputs numerical values of the reflected light amount, weight, and similarity exceeding the preset step-by-step threshold, and displays the determined defect type on the parts and assembly units differently by type.

Images