WO2023128125A1

WO2023128125A1 - Intelligent digital metal component, manufacturing method and artificial intelligence system comprising same

Info

Publication number: WO2023128125A1
Application number: PCT/KR2022/013049
Authority: WO
Inventors: 정임두; 서은혁
Original assignee: 울산과학기술원
Priority date: 2021-12-29
Filing date: 2022-08-31
Publication date: 2023-07-06
Also published as: KR20230101028A

Abstract

The present invention presents an intelligent digital metal component having a sensor for data collection embedded in the metal component, and a manufacturing method therefor. The intelligent digital metal component of the present invention comprises: a first metal layer in which metal powder is stacked; a sensor mounting space and fastening holes formed in the first metal layer; a sensor embedded in the sensor mounting space; a protection layer for sealing the upper part of the sensor mounting space; and a second metal layer formed by stacking metal powder on the first metal layer and the protection layer. The present invention further provides an artificial intelligence system, which analyzes sensor data detected by the sensor of the intelligent digital metal component, so as to monitor the fastened state of the intelligent digital metal component, the type of object applying impact from outside, or the like.

Description

Intelligent digital metal parts, manufacturing methods, and artificial intelligence systems including them

The present invention relates to an intelligent digital metal part having a sensor for data collection in the metal part, a manufacturing method thereof, and a fastening state of the metal part or an external impact object applied to the metal part based on real-time data collected from the sensor It is about an artificial intelligence system that can diagnose and monitor the type of

The 4th Industrial Revolution can be said to be the development of various technologies that could not be established only with the existing manufacturing process. For example, artificial intelligence and IoT (Internet of Things) technology combine real-time remote monitoring to create smart factories and self-driving cars. , it can be an important part of intelligent manufacturing technologies such as unmanned aerial vehicles. In addition, AI (Artificial Intelligence) technology resulting from various recent artificial neural network technologies is used to determine and predict the state of various mechanical devices.

However, since it is difficult to collect meaningful big data from mechanical systems or the physical environment of factories, there is a lack of research on applying AI technology to manufacturing. Nevertheless, various studies are being conducted to apply AI technology by combining sensors and semiconductors to various existing systems and factories for the efficiency of manufacturing systems. For example, a vibration sensor is attached to the surface of the pump to prevent oil refining efficiency from deteriorating due to pump performance degradation in the oil refinery system, and the replacement timing of the pump and pump components is determined by utilizing the data collected by the vibration sensor and AI technology. Predictable. In the past, parts such as pumps were replaced based only on experience or the lifespan of the pump, so compared to the past, the condition of mechanical parts can be efficiently and accurately monitored, and mechanical parts can be replaced at the right time according to the results. It became.

However, in the case of conventionally attaching a sensor to a mechanical system, considering that the system and the sensor are easily deteriorated by being exposed to the external environment, there are several problems as follows.

That is, the sensor itself can negatively affect the performance of the mechanical system.

In addition, since the sensor is easily separated or damaged by the movement of metal parts or the external environment, it is difficult to apply it to various environments or use it appropriately.

In addition, when a sensor is installed on a metal part, since most of the sensors must use a thin film sensor, there may be many restrictions on the shape of the sensor to be manufactured depending on the shape of the metal part to be installed as well as the increase in manufacturing cost of the sensor.

Therefore, it has been devised to solve the above problems, and an object of the present invention is to minimize or eliminate various external influences that occur when the sensor is mounted on the surface of various devices, etc. Intelligent sensor is mounted inside the part It is to provide digital metal parts.

Another object of the present invention is to provide a method for manufacturing a digital metal part in which a sensor is mounted.

Another object of the present invention is to provide an artificial intelligence system capable of monitoring and determining the fastening state of a metal part or the type of an external impact object applied to the metal part.

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

An intelligent digital metal component according to an embodiment of the present invention for achieving the above object includes a first metal layer on which metal powder is laminated; sensor mounting spaces and fastening holes formed in the first metal layer; a sensor built into the sensor mounting space; a protective layer sealing an upper portion of the sensor mounting space; It is characterized by comprising a second metal layer in which metal powder is laminated on the first metal layer and the protective layer.

The first metal layer and the second metal layer are formed by laminating metal powder by a laser powder bed melting (L-PBF) process.

The protective layer is made of the same material as the first metal layer and the second metal layer.

The sensor mounting space is formed in an area where the most stability is secured through at least one test among a tensile test, a FEM analysis test, and a microstructure analysis test for the intelligent digital metal part.

The sensor mounting space may be two or more, and the same sensor or a different sensor is embedded in each sensor mounting space.

A protrusion is formed on a surface of at least one of the first metal layer and the second metal layer.

The intelligent digital metal component is a T-shaped body including a first body and a second body made of the first metal layer and the second metal layer, and the sensor mounting space is formed in a central portion of the first body, The sensor is embedded in the sensor mounting space.

A method of manufacturing an intelligent digital metal part according to another feature of the present invention includes forming a first metal layer in which a sensor mounting space is formed at a predetermined portion while repeatedly applying metal powder; mounting a sensor in the sensor mounting space; covering an upper portion of the sensor mounting space with a previously prepared protective layer; and forming a second metal layer by repeatedly applying a metal powder to the upper surfaces of the first metal layer and the protective layer.

An artificial intelligence system according to another feature of the present invention is an artificial intelligence system that analyzes sensor data detected by a sensor embedded in the intelligent digital metal part and monitors the state of the intelligent digital metal part, and performs data preprocessing. preprocessing module; a learning module that learns the preprocessed preprocessed data; and an output module outputting the learned data.

The pre-processing module may include a data input unit receiving the sensor data (raw data); a first transform unit that converts the sensor data into a frequency domain by performing a Fast Fourier Transform (FFT) on the sensor data; A second conversion unit for converting the data converted into the frequency domain into data for machine learning, wherein the second conversion unit includes a half cut processor and a spectrogram, and converts the sensor data into a 10*5 image size spectrogram It is generated every predetermined time and provided to the learning module.

The learning module includes 6 convolutional layers; 1 transition layer; 1 pulley connected layer; and a CNN learning model including one softmax function unit, and converts the 10*5 image size into a 2*2 image and outputs it.

The output module is classified and expressed as a 3D visualization t-SNE graph.

According to the present invention, it is possible to easily embed a sensor capable of detecting various measurement values in a part or device having a predetermined shape, and various problems caused by the prior art in which a thin film sensor is mounted on the surface of a part or device It has the effect of providing intelligent digital metal parts that solve these problems.

According to the present invention, sensor data of intelligent digital metal parts is analyzed through a CNN learning model, so that the fastening state or external shock factors of intelligent digital metal parts mounted on a series of devices or systems can be easily monitored. there is. Therefore, without any negative impact on the performance of the device or system to which the sensor is attached, the effect of easily determining whether to inspect, repair, or replace the metal parts as well as the devices or systems equipped with metal parts can be expected. .

1 is a perspective view of an intelligent digital metal component according to an embodiment of the present invention.

Figure 2 is a plan view of Figure 1;

3 is a partially enlarged view of a sensor mounting space according to the present invention.

4 is a process diagram showing a manufacturing process of an intelligent digital metal part according to an embodiment of the present invention.

5 is a drawing of a metal part to be tested used in a test to form a sensor mounting space in the metal part of the present invention.

6 is a view showing the tensile test results of the metal part to be tested in FIG. 5;

FIG. 7 is a view showing a finite element method (FEM) analysis result of the metal part to be tested in FIG. 5 .

8 is a graph showing the result of analyzing the microstructure of the metal part to be tested in FIG. 5 .

9 is a block diagram of an artificial intelligence system for determining the state of an intelligent metal part of the present invention.

10 is a detailed configuration diagram of the CNN learning model of FIG. 9 .

11 and 12 are diagrams showing screw fastening states of metal parts and test results thereof.

13 and 14 are exemplary diagrams showing types of external objects applied to metal parts and diagrams showing test results thereof.

Since the present invention can apply various transformations and have various embodiments, specific embodiments will be illustrated in the drawings and described in detail. However, it should be understood that this is not intended to limit the specific embodiments of the present invention, and includes all conversions, equivalents, and substitutes included in the spirit and scope of the present invention. In describing the present invention, if it is determined that a detailed description of related known technologies may obscure the gist of the present invention, the detailed description will be omitted.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. These terms are only used for the purpose of distinguishing one component from another.

Terms used in the present invention are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, the terms "include" or "have" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

Spatially relative terms, such as below, beneath, lower, above, upper, etc., facilitate the correlation between one element or component and another element or component, as shown in the drawing. can be used to describe Spatially relative terms should be understood as encompassing different orientations of elements in use or operation in addition to the orientations shown in the figures. For example, when an element shown in the drawing is turned over, an element described as below or beneath another element may be placed above or above the other element. Accordingly, the exemplary term below may include both directions of down and above. Elements may also be oriented in other orientations, and thus spatially relative terms may be interpreted according to orientation.

As used in the present invention, an expression indicating a part such as “part” or “part” refers to a device in which a corresponding component may include a specific function, software which may include a specific function, or a device which may include a specific function. and software, but cannot necessarily be limited to the expressed functions, which are provided only to help a more general understanding of the present invention, and those with ordinary knowledge in the field to which the present invention belongs If so, various modifications and variations are possible from these descriptions.

In addition, it should be noted that all electrical signals used in the present invention, as an example, can be reversed in signs of all electrical signals to be described below when an inverter or the like is additionally provided in the circuit of the present invention. Therefore, the scope of the present invention is not limited to the direction of the signal.

Therefore, the spirit of the present invention should not be limited to the described embodiments, and it will be said that not only the claims to be described later, but also all modifications equivalent or equivalent to these claims fall within the scope of the spirit of the present invention. .

The present invention is to embed sensors in various metal-based parts, and monitor and diagnose the state of each part based on the data collected by the sensor. Hereinafter, the present invention will be described in more detail based on the embodiments shown in the drawings. let me explain

1 is a perspective view of an intelligent digital metal component according to an embodiment of the present invention, and FIG. 2 is a plan view of FIG. 1 . According to this embodiment, the metal part is a 'T' shaped metal part, but the shape of the metal part applied to the present invention is not limited thereto.

As shown in FIGS. 1 and 2, the intelligent digital metal part (hereinafter also referred to as 'metal part') is a first body 110a and a second body orthogonally intersecting at one end of the first body 110a. It is composed of a T-shaped body 110 made of a body (110b). In addition, a plurality of functional fastening holes 111a to 111e are formed in the first body 110a and the second body 110b to fasten fastening means such as screws. All of the fastening holes 111a to 111e may have different shapes. For example, it may be formed in a long hole shape extending along the longitudinal direction of the first body 110a, or may be formed in a substantially circular shape.

Also, the diameter of the circular fastening hole may be the same or different. That is, the shape of the fastening holes 111a to 111e may be variously changed according to the overall shape of the metal part. In addition, the fastening method of the metal part temporarily fastens appropriate screws to the fastening holes 111a to 111e, and in particular, while the screw is temporarily fastened to the long hole-shaped fastening hole, the metal part is horizontally moved to determine the fastening position, and then the screw When completely fastened, it is possible to firmly mount the metal part on a desired location of the target object.

As shown in FIG. 1 , fine protrusions 112 may be formed on the surface of the metal part. The fine protrusions 112 may improve bonding performance by increasing a bonding surface area when the surface of a metal part needs to be bonded to another material or object using a bond or cement.

1 and 2, the metal part of the present invention includes a sensor 120. The sensor 120 is for monitoring and diagnosing the state of the metal part on which the sensor is mounted. The figure shows that the sensor 120 is provided inside the first body 110a of the metal part. Specifically, a sensor mounting space 115 is provided in a predetermined area of the first body 110a, and the sensor 120 is installed in the sensor mounting space 115. That is, according to the present invention, the sensor 120 is not generally mounted on the surface of the structure of the metal part, but the sensor 120 inside the metal part. Therefore, the present invention has the advantage of completely eliminating the problem of loss or deterioration due to external environmental factors when the sensor is mounted on the surface and exposed to the external environment as it is. In addition, since there is no need to manufacture the sensor in the form of a thin film, the time and cost required for manufacturing the sensor can be reduced.

In this embodiment, the sensor 120 may be a strain gage. However, the sensor 120 is not limited to a strain gauge, and other types of sensors for measuring physical quantities may be provided. In addition, if necessary, two or more sensor mounting spaces may be provided in the body of the metal part, and the same sensor or different sensors may be provided in each sensor mounting space.

3 is a partially enlarged view of a sensor mounting space according to the present invention. As shown therein, a sensor mounting space 115 is formed in the body 110 forming the metal part. In addition, a sensor 120 is installed in the sensor mounting space 115, and the sensor 120 is a sensor for measuring the degree of deformation of a metal part, and may be a strain gauge sensor.

And the body 110 of the metal part is formed of a metal layer on which metal powder is stacked. In the embodiment, a metal part is manufactured by using an additive manufacturing method using a selective laser melting (SLM) method. Reference numeral 116 may be referred to as a part formed by the SLM method, and in the manufacturing process, the first metal layer 116a and the second metal layer 116b are formed in order according to the process order. The manufacturing process will be discussed in detail below.

A protective layer 130 is formed on top of the sensor 120 . The protective layer 130 serves to protect the sensor 120 as well as to prevent metal powder from collapsing on the upper side of the sensor mounting space during the molding of the metal part or after the molding is completed. The material of the protective layer 130 may be a metal of the same material as the metal powder (ie, the first metal layer and the second metal layer), and is melted together with the metal powder by laser irradiation so as to be integrated. However, other materials that can suitably melt with metal powder may be used.

The metal part of the present invention is molded and manufactured by the SLM process. Although the SLM process is not shown in the drawings, it may be performed in the following order.

A thin metal layer is repeatedly laminated by applying metal powder to the build plate by a recoater. Lamination is repeated until the shape (height) of the metal part to be molded is formed. The metal parts shown in FIG. 1 and the like can be molded by such an SLM process. As described above, the metal part of FIG. 1 may be in a state in which the sensor mounting space 115 and the fastening holes 111 are formed.

Next, we look at the process of manufacturing intelligent digital metal parts by mounting sensors on metal parts made by the SLM process.

4 is a process diagram showing a manufacturing process of an intelligent digital metal part according to an embodiment of the present invention. 4 is a process of manufacturing an intelligent metal part by mounting a sensor on a metal part molded by a 3D printing process, laser powder bed fusion (L-PBF).

In FIG. 4 (a), the body of the metal part in which the sensor mounting space 115 is formed is manufactured. A method of determining the formation position of the sensor mounting space 115 in the metal part will be described in detail below. The process (a) may include a process of repeatedly printing the metal powder until the sensor mounting space 115 is formed at a height sufficient to mount the sensor. At this time, a hole should be formed on the side of the metal part so that the wire of the sensor can be wired.

And, as shown in (b), the sensor 120 is mounted in the sensor mounting space 115. Here, if the sensor mounting space 115 is not formed in the process (b), a process of removing metal powder from the sensor mounting space 115 may be accompanied.

And, as shown in (c), the protective layer 130 prepared in advance is covered on top of the sensor 120. The protective layer 130 is pre-manufactured in an appropriate size so as to cover only the upper portion of the sensor mounting space 115, and in particular, the protective layer 130 serves to prevent deterioration of the sensor 120 due to laser processing.

Then, the body 110 and the protective layer 130 are integrated by irradiating a laser as in (d). Through this L-PBF process, as shown in (e), an intelligent digital metal part having a built-in sensor capable of measuring physical strain data can be manufactured.

On the other hand, the manufacturing process of the intelligent digital metal part according to the present invention can also be carried out in the following way.

Specifically, the first metal layer 116a (FIG. 3) is formed by applying a metal powder to the build plate using a recoater and repeatedly stacking the metal layers. At this time, when the first metal layer 116a is formed, the sensor embedded space 115 in which the sensor 120 is to be embedded and the fastening holes 111 are also formed. Then, the sensor 120 for measuring the physical quantity is inserted into the sensor built-in space 115, and the upper part of the sensor built-in space 115 is covered with a protective layer 130 prepared in advance. Thereafter, a metal layer is repeatedly stacked on the first metal layer 116a and the protective layer 130 to form a second metal layer 116b (FIG. 3).

Through the process of forming the first metal layer 116a and the second metal layer 116b in this order, it is possible to manufacture a metal part with a built-in sensor.

The following describes the process of determining the sensor mounting space where the sensor is to be mounted in the intelligent digital metal part.

In the present invention, when an external force is applied to a metal part, it is necessary to mount the sensor at a position where various characteristics are minimal. To this end, in the present invention, a sample having the same shape as the metal part was first molded, and various tests were conducted on this sample.

Referring to (a) and (b) of FIG. 5 , Part 1 to Part 3, which are candidate positions of the sensor mounting space, were selected in the T-shaped first body of the metal part to be tested. Part 1 and Part 3 are near the fastening holes in the upper and lower directions from the center of the first body, and Part 2 is the central portion of the first body in which no fastening holes are formed. Specifically, Part 1 may be a 15 mm part above the sensor mounting space, Part 2 may be a sensor mounting space, and Part 3 may be a 17 mm part below the sensor mounting space. Here, the sensor mounting space is a position determined by various test results described below. In other words, it can be said that the location of Part 2 was judged more optimal as a sensor mounting space than the locations of Part 1 and Part 3.

The tensile test was performed by applying tension until the metal part to be tested was fractured. When the tension is applied in this way, as can be seen in FIG. 6, it appears that the possibility of breakage is relatively high in Part 1 and Part 3 located at the upper and lower parts, except for Part 2, which corresponds to almost the center of the metal part to be tested.

In the digital image correlation (DIC) tensile test, Part 1 showed local deformation until fracture occurred, and Part 3 also showed deformation, although smaller than Part 1. However, it was confirmed that Part 2 maintains its initial state with relatively little deformation.

As shown in FIG. 7a, FEM analysis was performed on two lines (line 1 and line 2), and the experimental results of FIG. 7b show that the degree of uniform deformation of Part 2 (sensor embedded region) was less than 0.01. , it can be seen that it is better than Part 1 and Part 3.

8 is a graph showing the result of analyzing the microstructure of the metal part to be tested in FIG. 5 . For the microstructure analysis, electron backscatter diffraction (EBSD) and electron channeling contrast imaging (ECCI) analysis were performed using a field emission scanning electron microscope (SEM), and the results of Part 1, Part 2, and Part 3 were analyzed. The orientation angles were compared, respectively.

As can be seen in (a), (b), and (c) of FIG. 8, the experimental results show that the orientation angle error is 60° in Part 1 (a drawing) and Part 3 (c drawing), whereas Part 2 (b drawing) In , the orientation angle error was found to be relatively insignificant.

As such, the present invention determines Part 2 of the metal part to be tested, whose stability has been substantially confirmed through the above experiments, as a mounting space in which the sensor 120 is to be mounted, and manufactures an intelligent digital metal part.

9 is a block diagram of an artificial intelligence system for determining the state of an intelligent metal part of the present invention. The artificial intelligence system 200 collects sensor data from intelligent digital metal parts and accurately determines the state of the intelligent digital metal parts based on the collected sensor data.

As shown, the artificial intelligence system 200 includes a preprocessing module 210 for data preprocessing, a learning module 220, and an output module 230 that outputs the learning result as visualized image data by classifying it into a 3D visualization t-SNE graph. ).

The pre-processing module 210 is a data input unit 211 into which sensor data, that is, raw data, is input from the sensor 120 mounted on the intelligent digital metal part, and performs fast Fourier transform (FFT) on the input raw data. It may include a first transform unit 212 that transforms into the frequency domain and a second transform unit 213 that converts the data converted into the frequency domain into data for machine learning. Here, the FFT algorithm of the first transform unit 212 may be a Discrete Fourier Transform (DFT) method that decomposes the discrete input signal into frequencies. The Discrete Fourier Transform (DFT) can be defined as:

And the second conversion unit 213 includes a half cut 213a and a spectrogram 213b. The spectrogram 213b is generated every 6 seconds as a 10*5 image size spectrogram through a half-cut processor from raw data, and is input as input data to the learning module 220 at a later stage.

The reason why the preprocessing module 210 is applied in the present invention is that it is difficult to learn raw data collected by the sensor 120 using machine learning, that is, a Convolution Neural Network Model (CNN) model.

In FIG. 9 , the learning module 220 may use a CNN learning model. The CNN learning model 220 of this embodiment includes a convolution layer 221, a transition layer 222, a fully connected layer 223, and a softmax function. section 224. The CNN learning model 220 having such a configuration learns the sensor data detected by the sensor 120 embedded in the intelligent digital metal part to learn the current state of the intelligent digital metal part, for example, the screw is not fastened or loosely fastened. The location of the damaged screw can be known, and the type of external impact object can also be known.

10 is a detailed configuration diagram of the CNN learning model of FIG. 9 . The CNN learning model 220 applied to this embodiment is based on DenseNet including 6 convolutional layers, 1 conversion layer, a fully connected layer, and a softmax layer. In addition, each convolution layer includes a Rectified Linear Unit (ReLU) function, and a conversion layer is placed in the middle of the CNN learning model to reduce the width, vertical size, and number of feature maps.

Referring to FIG. 10, a 10*5 image is finally converted into a 2*2 image through a convolution layer and a conversion layer, and the state of a metal part is classified through a fully connected layer and a soft max layer.

11 and 12 are diagrams showing screw fastening states of metal parts and test results thereof. As shown in FIG. 11, (a) is the normal state in which the mode screws are completely fixed, (b) is the state in which all screws are loose (abnormal state 1), (c) is the left side based on the drawing The test was conducted based on the state where there is no screw (Abnormal state 2) and (d), the left screw is loosely fastened and the right screw is not (Abnormal state 3).

The test was performed through the processes of the preprocessing module 210, the learning module 220, and the output module 230, and the test result is expressed as a t-SNE 3D plot as shown in FIG. 12. Through this, the user can easily check the fastening state of the metal part.

As another example, it is also possible to recognize an object impacting a metal part. 13 is a view showing types of external objects applied to metal parts. Looking at this, (a) shows an example of applying an impact to a metal part using a hand, (b) a hammer, and (c) a spanner.

The test was performed through the processes of the preprocessing module 210, the learning module 220, and the output module 230, and the test result can be expressed as a t-SNE 3D plot as shown in FIG. 14. Referring to FIG. 14, it is expressed differently depending on the type of object applied to the metal part, so that the user can also easily check the type of object applied to the metal part.

As such, the present invention embeds a sensor inside a metal part using L-PBF (Laser Powder Bed Fusion) technology, performs FFT (Fast Fourier Transform) and image processing on sensor data collected in real time from the sensor, and then It was trained on the CNN learning model. In addition, the condition or malfunction of metal parts can be diagnosed and predicted by the learning result of the CNN learning model, and the learning result is expressed as t-stochastic Neighbor Embedding (t-SNE).

Although the above has been described with reference to the illustrated embodiments of the present invention, these are only examples, and those skilled in the art to which the present invention belongs can variously It will be apparent that other embodiments that are variations, modifications and equivalents are possible. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

It can be used for intelligent digital metal parts where sensors are mounted inside the part.

Claims

A first metal layer on which metal powder is laminated;

sensor mounting spaces and fastening holes formed in the first metal layer;

a sensor built into the sensor mounting space;

a protective layer sealing an upper portion of the sensor mounting space; and

Characterized in that it comprises a second metal layer in which metal powder is laminated on the first metal layer and the protective layer, the intelligent digital metal component.
According to claim 1,

The first metal layer and the second metal layer,

An intelligent digital metal part formed by laminating metal powder by the laser powder bed melting (L-PBF) process.
According to claim 1,

Wherein the protective layer is made of the same material as the first metal layer and the second metal layer.
According to claim 1,

The sensor mounting space,

An intelligent digital metal part formed in an area where the most stability is secured through at least one test of a tensile test, a FEM analysis test, and a microstructure analysis test for the intelligent digital metal part.
According to claim 1,

The sensor mounting space may be two or more,

The sensor is an intelligent digital metal part in which the same sensor or a different sensor is embedded in each sensor mounting space.
According to claim 1,

A protrusion is formed on the surface of at least one of the first metal layer and the second metal layer, the intelligent digital metal part.
According to claim 1,

The intelligent digital metal part,

A T-shaped body including a first body and a second body made of the first metal layer and the second metal layer,

The sensor mounting space is formed in the central portion of the first body,

An intelligent digital metal part in which the sensor is embedded in the sensor mounting space.
Forming a first metal layer in which a sensor mounting space is formed in a predetermined area while repeatedly applying metal powder;

mounting a sensor in the sensor mounting space;

covering an upper portion of the sensor mounting space with a previously prepared protective layer; and

The method of manufacturing an intelligent digital metal part, characterized in that carried out including the step of forming a second metal layer by repeatedly applying a metal powder on the upper surface of the first metal layer and the protective layer.
According to claim 8,

The first metal layer and the second metal layer,

A method for manufacturing an intelligent digital metal part formed by laminating metal powder by a laser powder bed melting (L-PBF) process.
An artificial intelligence system for monitoring the state of an intelligent digital metal part by analyzing sensor data detected by a sensor embedded in an intelligent digital metal part comprising the configuration of any one of claims 1 to 7,

a pre-processing module for pre-processing the sensor data;

a learning module for learning the preprocessed data preprocessed by the preprocessing module; and

An artificial intelligence system for determining the state of an intelligent digital metal part, characterized in that it comprises an output module that outputs the data learned by the learning module.
According to claim 10,

The preprocessing module,

a data input unit receiving the sensor data (raw data);

a first transform unit for transforming the sensor data into a frequency domain by performing a Fast Fourier Transform (FFT) on the sensor data; and

A second conversion unit for converting the data converted into the frequency domain into data for machine learning;

The second conversion unit includes a half cut processor and a spectrogram,

An artificial intelligence system for determining the state of an intelligent digital metal part, generating the sensor data as a spectrogram of 10*5 image size every predetermined time and providing it to the learning module.
According to claim 11,

The learning module,

6 convolutional layers;

1 transition layer;

1 pulley connected layer; and

A CNN learning model including one softmax function,

An artificial intelligence system for determining the state of an intelligent digital metal part that converts the 10*5 image size into a 2*2 image and outputs it.
According to claim 10,

The output module,

An artificial intelligence system for determining the condition of intelligent digital metal parts, classified and expressed as a 3D visualization t-SNE graph.