JP7194295B1 - How to build a digital twin model - Google Patents

Abstract

A method for building a digital twin model is provided. A simplified geometry is shaped and sized to correspond to a member of a feed system. After sampling the simplified geometric body to obtain second position data, the material data of the member, the second position data, and the second size data of the simplified geometric body are obtained by a modal analysis method. calculating a set of model eigenvalues and a set of model eigenvectors based on, and determining that a set of actual eigenvectors of the member is similar to the set of model eigenvectors by a modal verification method, the simplified geometric body is defined as the digital twin model of the member. The amount of data of the second position data and the second size data is much smaller than the amount of data of the first position data and the first size data of the component image. This accelerates the model building speed and greatly reduces the amount of data. [Selection drawing] Fig. 2

Description

The present invention relates to the field of digital twins, and more particularly to methods for building digital twin models.

In recent years, the application of the digital twin technology to industry has been gradually spreading. Digital twin technology is used to build a virtual model of a physical object, and there is a relationship between the physical object and the virtual model. The virtual model generates feedback after a series of processing, analysis and judgments are made on the data immediately returned by the sensing unit.

Chinese Patent Application Publication No. 112292702A Taiwan Patent Application Publication No. 668584 Chinese Patent Application Publication No. 112487584

However, the amount of data of the virtual model is generally very large, and a huge amount of data calculation is required to obtain the feedback result of the virtual model. Therefore, not only does such a virtual model require a huge amount of computational processing resources, but it is also unsuitable for evaluating whether or not the physical member can be applied to mechanical devices with different specifications (for example, See conventional Patent Documents 1, 2, and 3).

Therefore, the inventor of the present invention thought that the above-mentioned drawbacks could be improved, and as a result of earnest studies, the present inventors came up with the proposal of the present invention that effectively solves the above-mentioned problems with a rational design.

The present invention has been accomplished through intensive research by the inventors in view of the above problems.

A first object of the present invention is to provide a method for constructing a digital twin model that greatly reduces the amount of data of the virtual model and accelerates the construction speed of the virtual model.

A second object of the present invention is to provide a method of building a digital twin model that significantly reduces the required amount of computational resources.

A third object of the present invention is to provide a method of constructing a digital twin model that can be applied to evaluate whether the virtual model to be generated can be applied to a mechanical device with different specifications of the members of the physical object. is.

A method for constructing a digital twin model of one aspect of the present invention to solve the above problems is applied to construct a digital twin model of at least one member of a feed system, said member being a corresponding set of actual models. and a set of actual eigenvectors, the method of constructing the digital twin model is executed by at least one processing unit, receives user settings via a user interface, and based on the user settings to set a geometric body image, wherein the contour of the simplified geometric body of said geometric body image corresponds to the contour of said member, said user settings relating to the shape and size of said simplified geometric body; (A); sampling the simplified geometric body of the geometric body image to obtain second position data (B); obtaining material data of the member from a database (C); (D) calculating a set of model eigenvalues and a set of model eigenvectors based on the second size data, the second position data and the material data of the simplified geometric body by modal analysis; modal verification; (E) determining the similarity between the set of actual eigenvectors and the set of model eigenvectors by the method; and if determining that the set of actual eigenvectors is similar to the set of model eigenvectors. , defining the simplified geometric body as the digital twin model of the member, and defining the set of model eigenvalues and the set of model eigenvectors as twin dynamics of the member; The amount of data of the second size data is less than the amount of data of the first size data of the member, the amount of data of the second position data is less than the amount of data of the first position data of the member, the first size data, and The first position data is stored in the database and obtained from a member image of the member.

In one aspect of the present invention, the set of actual eigenvalues and the set of actual eigenvectors are calculated by the modal analysis method based on the first size data, the material data, and the first position data. .

In one aspect of the present invention, step (B) includes step (B1) of discretizing the simplified geometric body into a plurality of second image blocks, and pixel coordinates of each vertex of each of the second image blocks as and (B2) defining as the second position data.

In one aspect of the present invention, the method for obtaining the first position data from the member image includes the step (G) of discretizing the member image into a plurality of first image blocks; and (H) defining pixel coordinates of vertices as the first position data.

At least the following matters will be clarified by the description of the present specification and drawings.

1 is a block diagram illustrating a system for building a digital twin model according to one embodiment of the invention; FIG. 1 is a flowchart illustrating a method for building a digital twin model according to one embodiment of the invention; 4 is a flowchart illustrating a method for obtaining first location data according to one embodiment of the present invention; 4 is a flowchart illustrating a method for obtaining second location data according to one embodiment of the present invention; 1 is a schematic configuration diagram showing a feed system of a machine according to one embodiment of the present invention; FIG. FIG. 6 is a schematic block diagram showing a component image of a workbench of the feed system of the machine of FIG. 5 according to one embodiment of the present invention; FIG. 7 is a schematic diagram of the member image of FIG. 6 processed by discretization according to one embodiment of the present invention; 1 is a schematic configuration diagram showing a geometric body image according to an embodiment of the present invention; FIG. 9 is a schematic diagram of the geometric body image of FIG. 8 processed by discretization according to an embodiment of the present invention; FIG.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. It goes without saying that the present invention is not limited to the following examples, and can be arbitrarily modified without departing from the gist of the present invention.

First, referring to FIGS. 1 to 9, a method for constructing a digital twin model (hereinafter abbreviated as construction method) and a system for constructing a digital twin model (hereinafter abbreviated as system 1) according to an embodiment of the present invention. ) will be explained in detail.

This construction method is executed by system 1 . The system 1 is applied to digital twin technology, construction of a virtual model of at least one member 21 of the feed system 2 of the machine and detection of its twin dynamics. The member 21 is, for example but not limited to, a bearing, a ball screw, a transmission member such as a rotary table or a linear slide rail, or a workbench 22 . In order to clearly explain the spirit of the present invention, an example in which the member 21 is a workbench 22 will be described below.

The system 1 may be embodied in a server unit or distributed in a plurality of servers communicating with each other. The system 1 comprises at least one processing device and at least one storage communicating with each processing device. A plurality of software are installed in this system 1, at least one storage, at least one processing device, and operationally jointly these software are a position sampling unit 11, a database 12, a non-simplified modal analysis unit 13, a geometric It comprises a body setting unit 14, a position sampling unit 15, a simplified modal analysis unit 16, a similarity determination unit 17, and a database .
A position sampling unit 11 and a non-simplified modal analysis unit 13 communicate with a database 12, a geometric body setting unit 14, a position sampling unit 15, a simplified modal analysis unit 16, a similarity determination unit 17, and a database. 18 communicate with each other, similarity determination unit 17 communicates with unsimplified modal analysis unit 13 and simplified modal analysis unit 16 communicates with database 12 .

The method of constructing the digital twin model of workbench 22 (ie, the construction method of the present invention) is not limited, but includes, for example, the following steps.

First, in step S<b>11 , the non-simplified modal analysis unit 13 obtains first size data, material data and first position data of the workbench 22 from the database 12 .
The first size data, material data, and first position data of the workbench 22 are stored in advance in the database 12, and the database 12 records the correspondence relationship between the first size data, the material data, and the first position data. there is The first size data is not limited, but is constructed or set, for example, when drawing the member image IM1 of the workbench 22 by graphic software (but not limited to, for example, AutoCAD (registered trademark)) installed in the system 1. be. This component image IM1 is a three-dimensional image in which the image V1 of the workbench 22 is represented. Although the first size data is not limited, for example, the length L1 (eg, 730 mm) along the axial direction D1 of the image V1, the width W1 (eg, 375 mm) along the axial direction D2, and the height H1 (eg, 375 mm) along the axial direction D3 of the image V1. For example, 170 mm), the diameter of the through-hole, and the depth of the groove, each size is not limited to the actual size or aspect ratio size. The axial directions D1-D3 are perpendicular to each other. Material data include, but are not limited to, density and Young's modulus, for example. Although the first position data is not limited, it is acquired by sampling pixel coordinates from the member image IM1, for example.

A method of acquiring the first position data is realized by a finite element method or a continuum method. In the example of the finite element method (see FIGS. 1, 3, 6 and 7), first, in step S31, the position sampling unit 11 acquires the member image IM1 of the workbench 22 from the storage. Then, in step S32, the position sampling unit 11 uses graphics software (e.g., without limitation, AutoCAD) or computer-aided engineering (CAE) software (without limitation, e.g., ANSYS (registered A plurality of first image blocks B1 (or subregion or element).
The mesh shape of the first image block B1 is not limited, but may be triangular or square, for example. In this embodiment, the mesh shape of the first image block B1 is triangular. Then, in step S33, the position sampling unit 11 defines the pixel coordinates of each vertex P1 (or called node or discrete point) of each first image block B1 as the first position data of the worktable 22; Finally, the position sampling unit 11 stores the first position data in the database 12 .

After the unsimplified modal analysis unit 13 obtains the first size data, the material data and the first position data, in step S12, the unsimplified modal analysis unit 13 executes the CAE software installed in the system 1. is used to determine a set of actual eigenvalues (i.e., actual eigenvalue data) and a set of actual eigenvectors ( That is, actual eigenvector data) are calculated. The set of actual eigenvalues are the eigenfrequencies of the worktable 22 and the set of actual eigenvectors are the modals of the worktable 22 . The set of actual eigenvalues and the set of actual eigenvectors are the dynamics of worktable 22 .

In step S12, after knowing the geometric shape, the first size data, the first position data, the material data (e.g., density, Young's modulus), and the formula of density, etc., the image V1 discretized by the modal analysis method We obtain equation (1).

On the other hand, in step S13, the geometric body setting unit 14 receives user settings through the user interface, and based on the user settings, sets a geometric body image IM2 corresponding to the contour of the workbench 22 (see FIG. 8). ).
User settings relate to the shape and size of the simplified geometric image V2 displayed in the geometric image IM2. The user interface is not limited, for example, the geometry setting unit 14 may be provided in combination with CAE software and displayed on a display device communicating with the processing device. By way of example, a user may follow the general contours of workbench 22 (e.g., resembling a rectangular body) by means of an input device (e.g., but not limited to, a keyboard, mouse, or touch panel of a display unit) that communicates with the processing unit. Based on this, one shape option is selected from a plurality of shape options for the virtual model provided on the user interface (rectangular body option), and the first size data for workbench 22 (but not limited to, for example, an image Based on the length L1, width W1, and height H1 of V1), enter the required size for the rectangular body-shaped simplified geometric body image V2 (for example, but not limited to, the length along the axis direction D1 L2 is 730 mm, the width W2 in the axial direction D2 is 375 mm, and the height H2 in the axial direction D3 is 170 mm).
These inputs regarding shape and size are the user settings and are transmitted to the geometry setting unit 14 . At this time, the geometric body setting unit 14 defines a rectangular simplified geometric body image V2 as a digital twin model of the workbench 22 based on this user setting, and the size of the simplified geometric body image V2 (ie, second size data). The simplified geometric body image V2 is taken as a simplified virtual model of the work table 22, and the shape and structure of the simplified geometric body image V2 have many structural features that do not significantly affect the dynamic characteristics in the work table 22. (For example, but not limited to, through holes, concave grooves, and convex ribs) are omitted, so that the data amount of the second size data of the simplified geometric body image V2 is the first size of the workbench 22 Much less than the amount of data in the data.

Subsequently, in step S14, the position sampling unit 15 obtains the geometric body image IM2 by the geometric body setting unit 14, performs sampling (or discretization) on the geometric body image IM2, and obtains a simplified geometric body image Obtain second position data of V2.
Although the sampling method is not limited, it is implemented by, for example, the finite element method or the boundary element method. In the example of finite element sampling (see FIGS. 1, 4, 8 and 9), position sampling unit 15 divides simplified geometric body image V2 in geometric body image IM2 into a plurality of second images in step S41. After spatial discretization into blocks B2, pixel coordinates of each vertex P2 of each second image block B2 are defined as second position data in step S42. In this embodiment, the shape of the second image block B2 is square, but in other embodiments, the second image block B2 may have the same shape as the first image block B1, and furthermore, the shape of the second image block B2 is square. The size may be the same as or different from the size of the first image block B1. The shape and structure of the simplified geometric body image V2 do not significantly affect the dynamic characteristics in the workbench 22. Since many structural features are omitted, the second position data of the simplified geometric body image V2 is The amount of data is much less than the amount of data for the first position data of the worktable 22 .

After that, the simplified modal analysis unit 16 obtains the material data from the database 12 in step S15, and obtains the second size data from the geometric body setting unit 14 and the second size data from the position sampling unit 15 in step S16. After obtaining the two-position data, adopt the same method as the non-simplified modal analysis unit 13, and according to the modal analysis method, the simplified geometric body based on the second size data, the second position data and the material data. Compute a set of model eigenvalues and a set of model eigenvectors of image V2. The model eigenvalue is the eigenfrequency of the simplified geometric image V2, and the model eigenvector is the modal of the simplified geometric image V2.

Once unsimplified modal analysis unit 13 has computed and obtained the set of actual eigenvalues and the set of actual eigenvectors, simplified modal analysis unit 16 computes the set of model eigenvalues and the set of actual eigenvectors. After calculating and obtaining the model eigenvectors, the similarity determination unit 17 obtains the set of actual eigenvectors by the unsimplified modal analysis unit 13 and the simplified modal analysis unit 16 in step S16. to obtain the set of model eigenvectors. A modal verification method is then used to determine the similarity between the set of actual eigenvectors and the set of model eigenvectors. Modal verification methods are not limited, but may be, for example, modal reliability criteria, average phase shift methods, or modal phase collinearity methods.

次いで、ステップＳ１８において、類似性判断ユニット１７が前記１セットの実際の固有ベクトルが前記１セットモデル固有ベクトルに類似していると判断した場合、前の簡略化幾何学体画像Ｖ２が作業台２２と同等であることを示す。この際、類似性判断ユニット１７はステップＳ１９において、簡略化幾何学体画像Ｖ２を作業台２２のデジタルツインモデルとして定義し、前記１セットモデル固有値及び前記１セットモデル固有ベクトルを作業台２２のツイン動特性として定義する。
また、ステップＳ１９において、類似性判断ユニット１７が幾何学体設定ユニット１４が幾何学体画像ＩＭ２及びその第二サイズデータをデータベース１８に保存したことを通知し、位置サンプリングユニット１５が第二位置データをデータベース１８に保存したことを通知し、簡略化されたモーダル解析ユニット１６がツイン動特性及び材質データをデータベース１８に保存したことを通知する。なお、データベース１８には幾何学体画像ＩＭ２、第二サイズデータ、第二位置データ、材質データ、及びツイン動特性の対応関係も記録されている。

Then, in step S18, if the similarity determination unit 17 determines that the set of actual eigenvectors is similar to the set of model eigenvectors, the previous simplified geometric body image V2 is equivalent to the workbench 22. indicates that At this time, in step S19, the similarity determination unit 17 defines the simplified geometric body image V2 as a digital twin model of the workbench 22, and converts the one set model eigenvalue and the one set model eigenvector into twin motions of the workbench 22. Define as a characteristic.
Further, in step S19, the similarity determination unit 17 notifies that the geometric body setting unit 14 has stored the geometric body image IM2 and its second size data in the database 18, and the position sampling unit 15 receives the second position data has been stored in the database 18 and the simplified modal analysis unit 16 has stored the twin dynamics and material data in the database 18 . The database 18 also records correspondence relationships among the geometric body image IM2, the second size data, the second position data, the material data, and the twin dynamic characteristics.

Conversely, if the similarity determination unit 17 determines in step S18 that the set of actual eigenvectors is not similar to the set of model eigenvectors, then the simplified geometric body image V2 at that time is 22 , the similarity determination unit 17 does not define the simplified geometric body image V2 as a digital twin model of the workbench 22 , and the set of model eigenvalues and the set of model eigenvectors as the workbench 22 's Not defined as twin dynamics.

Through the process of steps S13 to S16 described above, the present invention greatly reduces the data amount of the virtual model and accelerates the construction speed of the virtual model. At the same time, the required amount of computing resources is also greatly reduced, making the workbench 22 adaptable for assessing the applicability of machines of different specifications.

The virtual model with the reduced data amount is maintained at the same level as the workbench 22 through the verification process of steps S17 and S18 described above.

The execution order of steps S13-S16 in the above embodiment is independent of steps S11-S12, however, the invention is not limited to this flow chart example. In other embodiments, steps S13-S16 can be performed at any point in time before step S17.

Additionally, although the above embodiments describe the workbench 22 as an example, in practice the system 1 and method of the present invention may be applied to other members of the feed system 2 (eg, but not limited to screws). It can be applied to build a digital twin model and its twin dynamics, or it can also be applied to build a digital twin model and its twin dynamics of a component of a device other than the feed system 2 of a mechanical device.

Although databases 12 and 18 in the above embodiment are constructed separately, the invention is not limited to this embodiment. Other embodiments may be modified to merge databases 12 and 18 together.

The present invention is not limited to the above-described embodiments, but can be modified in various ways within the scope of the claims, and can be obtained by appropriately combining technical means disclosed in different embodiments. is also included in the technical scope of the present invention.

1 System for Building Digital Twin Models 11 Location Sampling Unit 15 Location Sampling Unit 12 Database 18 Database 13 Unsimplified Modal Analysis Unit 14 Geometric Body Setting Unit 16 Simplified Modal Analysis Unit 17 Similarity Determining Unit 2 Feeds System 21 Member 22 Worktable B1 First image block B2 Second image block D1 Axial direction D2 Axial direction D3 Axial direction H1 Height H2 Height IM1 Member image IM2 Geometric body image L1 Length L2 Length P1 Vertex P2 Vertex V1 Image V2 Simplified geometric body image W1 width W2 width

Claims (5)

A method of building a digital twin model applied to build a digital twin model of at least one component of a feed system, comprising:
the member has a corresponding set of actual eigenvalues and a set of actual eigenvectors;
The method of constructing the digital twin model is performed by at least one processing unit,
receiving user settings through a user interface, setting a geometric body image based on said user settings, a simplified geometric body contour of said geometric body image corresponding to said member contour, and step (A), wherein the settings relate to the shape and size of said simplified geometry;
(B) sampling the simplified geometric body of the geometric body image to obtain second position data;
a step (C) of acquiring material data of the member from a database;
(D) calculating a set of model eigenvalues and a set of model eigenvectors based on the second size data, the second position data and the material data of the simplified geometric body by modal analysis;
(E) determining the similarity of the set of actual eigenvectors and the set of model eigenvectors by a modal verification method;
If the set of actual eigenvectors is determined to be similar to the set of model eigenvectors, define the simplified geometric body as the digital twin model of the member, and (F) defining model eigenvectors as twin dynamics of said members;
the amount of data of the second size data is smaller than the amount of data of the first size data of the member;
the amount of data of the second position data is less than the amount of data of the first position data of the member;
The first size data and the first position data are stored in the database and obtained from the member image of the member,
How to build a digital twin model. 3. The set of actual eigenvalues and the set of actual eigenvectors are calculated by the modal analysis method based on the first size data, the material data, and the first position data. 2. A method of constructing a digital twin model according to 1. The method of constructing a digital twin model according to claim 1, wherein the step (B) and the method of obtaining the first position data from the member image are performed by a finite element method or a continuum method. The step (B) includes step (B1) of discretizing the simplified geometric body into a plurality of second image blocks, and defining pixel coordinates of each vertex of each of the second image blocks as the second position data. The method of building a digital twin model according to claim 1, characterized in that it comprises a step (B2) of: The method for obtaining the first position data from the member image includes the step (G) of discretizing the member image into a plurality of first image blocks, and calculating the pixel coordinates of each vertex of each first image block as the first image block. and (H) defining as single position data.

Images