GPT-based elasticizer digital twinning mechanism model construction method and system

Through the GPT deep learning model, the historical digital twin mechanism model of the add-on machine is converted into natural language text and the target model is automatically generated, which solves the problems of low efficiency and high cost of manual modeling in the existing technology, and realizes the prediction of the add-on machine failure Efficient construction and automatic generation of models.

CN119089795AActive Publication Date: 2024-12-06WUXI HONGYUAN ELECTROMECHANICAL TECH +1
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411262145.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-10
Publication Date
2024-12-06
Estimated Expiration
2044-09-10

AI Technical Summary

Technical Problem

The construction of existing bullet-adding machine fault prediction models relies on manual operation, which is costly and inefficient, making it difficult to meet practical application needs, especially when design or process changes, manual modeling is required.

Method used

Using the GPT-based digital twin mechanism model construction method, the digital twin mechanism model of the historical bomber is converted into natural language text through the 'model-program-text' forward multimodal mapping mechanism, and the deep learning model is used for training to achieve Automatic generation of target models.

Benefits of technology

It improves the efficiency of constructing mechanism models, reduces costs, and realizes the rapid transformation and automatic generation of digital twin mechanism models, solving the problems of low efficiency and high cost of traditional manual modeling.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089795A_ABST
    Figure CN119089795A_ABST
Patent Text Reader

Abstract

The invention discloses a GPT-based elasticizer digital twinning mechanism model construction method and system, and the method comprises the steps: S1, obtaining a plurality of historical elasticizer digital twinning mechanism models, and converting the historical elasticizer digital twinning mechanism models into natural language texts through a model-program-text forward multi-modal mapping mechanism; s2, constructing a GPT deep learning model, and training the GPT deep learning model by taking the natural language text obtained in S1 and the corresponding computer program as samples; S3, obtaining a demand of a target elasticizer digital twinning mechanism model, converting the demand into a natural language text in a preset format, inputting the natural language text into the trained GPT deep learning model, and obtaining a target elasticizer digital twinning mechanism model; and converting the output of the GPT deep learning model into a model to obtain a target elasticizer digital twinning mechanism model. The method is high in construction efficiency and low in cost.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of fault prediction of complex textile equipment, and in particular to a method and system for constructing a digital twin mechanism model of a texturing machine based on GPT. Background Art

[0002] The textile industry is related to people's livelihood and is a national basic industry. One of the important standards for measuring the national textile level is high-end textile complex equipment. The texturing machine is a highly integrated advanced engineering and technical machinery. This type of equipment is the key to breaking through the intelligent transformation of the textile industry. The production mode of the texturing machine is usually 24x7, and it is in an overloaded and high-intensity state for a long time. The damage or performance degradation of its key components will not only reduce the production quality, but also cause unexpected downtime of the equipment, causing huge economic losses to the company. Therefore, it is very important to use digital means to predict the faults of the texturing machine. The realization of fault prediction of the texturing machine requires not only the collection of real-time and comprehensive operating data, but also the establishment of a digital model that can truly reflect the internal mechanism.

[0003] When constructing a mechanism model, it is necessary to fully consider the physical coupling relationship between the components within the complex system and its corresponding degradation mechanism in order to form a digital model that can accurately map the physical degradation entity. However, the construction of such mechanism models is currently highly dependent on manual operations, which is not only costly but also inefficient. In particular, when the design, process or working conditions change, manual modeling needs to be repeated, which is difficult to meet the needs of practical applications. Summary of the invention

[0004] In view of the problems existing in the prior art, the purpose of the present invention is to provide an efficient method and system for constructing a digital twin mechanism model of a texturing machine based on GPT.

[0005] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0006] A method for constructing a digital twin mechanism model of a texturing machine based on GPT includes the following steps:

[0007] S1. Obtain several digital twin mechanism models of historical texturing machines, and convert them into natural language texts through a "model-program-text" forward multimodal mapping mechanism, wherein the forward multimodal mapping mechanism is used to first convert the digital twin mechanism model of the historical texturing machine into a computer program, and then convert the computer program into a natural language text in a preset format;

[0008] S2. Construct a GPT deep learning model, and use the natural language text and the corresponding computer program generated in step S1 as samples to train the GPT deep learning model, wherein the GPT deep learning model is used to generate a corresponding computer program according to the natural language text of the digital twin mechanism model of the texturing machine;

[0009] S3. Obtain the requirements of the digital twin mechanism model of the target texturing machine, convert the requirements into natural language text in a preset format, input it into the trained GPT deep learning model, and convert the output of the GPT deep learning model into a model to obtain the digital twin mechanism model of the target texturing machine.

[0010] Furthermore, step S1 specifically includes the following steps:

[0011] S11. Obtain several digital twin mechanism models of historical texturing machines, convert the digital twin mechanism models of historical texturing machines into computer programs in a preset programming language using a model conversion tool, conduct tests, and delete or modify computer programs that fail the tests;

[0012] S12. According to the digital twin mechanism model of the historical texturing machine, extract the key textual knowledge of the texturing machine, build a textual knowledge database of the texturing machine, and map the textual knowledge in the texturing machine textual knowledge database with the key information of the texturing machine in the computer program to form a textual knowledge-computer program mapping relationship;

[0013] S13, parsing the computer program obtained in step S11, extracting key information of the texturing machine therein, and using natural language generation technology, combined with a text knowledge-computer program mapping relationship, converting the key information of the texturing machine in the computer program into a natural language text in a preset format;

[0014] S14. Verify the natural language text obtained in step S13, and modify and adjust the inconsistencies or errors therein.

[0015] Furthermore, the natural language generation technology is used to convert the key information of the texturing machine in the computer program into a natural language text in a preset format in combination with the text knowledge-computer program mapping relationship, specifically including:

[0016] Through a pre-written text template, using rules or machine learning algorithms, according to the text knowledge-computer program mapping relationship, the key information of the texturing machine in the computer program is matched and filled with the text template, and a natural language text in a preset format is generated, and the text is checked after filling in; wherein, the text template is a template that describes the different components and working states of the texturing machine.

[0017] Furthermore, step S2 specifically includes the following steps:

[0018] S21, taking the natural language text and computer program of each historical texturing machine digital twin mechanism model as sample pairs, forming a sample set, and dividing it into a training set and a test set;

[0019] S22. Construct a GPT deep learning model, wherein the GPT deep learning model includes a text encoder and a program parser connected to each other, wherein the text encoder is used to encode the input natural language text of the digital twin mechanism model of the texturing machine into a series of vector representations, wherein the vector representation captures the semantic and syntactic information in the natural language text; and the program parser is used to generate a corresponding computer program according to the vector representation;

[0020] S23. Use the training set of step S21 to pre-train the GPT deep learning model, and after the pre-training is completed, use the test set to test the GPT deep learning model, and finally fine-tune the GPT deep learning model according to the test results until the GPT deep learning model meets the preset requirements.

[0021] Furthermore, step S3 specifically includes the following steps:

[0022] S31. Obtaining the requirements of the digital twin mechanism model of the target texturing machine, filling the requirements of the digital twin mechanism model of the target texturing machine into a pre-written text template to obtain a natural language text in a preset format;

[0023] S32, inputting the natural language text obtained in step S31 into the trained GPT deep learning model, and the output is a computer program of the digital twin mechanism model of the target texturing machine;

[0024] S33. Convert the computer program of the digital twin mechanism model of the target texturing machine into the digital twin mechanism model of the target texturing machine by using a model conversion tool.

[0025] A digital twin mechanism model construction system for a texturing machine based on GPT, including:

[0026] A forward mapping module is used to obtain a number of historical texturing machine digital twin mechanism models and convert them into natural language texts through a "model-program-text" forward multimodal mapping mechanism, wherein the forward multimodal mapping mechanism is used to first convert the historical texturing machine digital twin mechanism model into a computer program, and then convert the computer program into a natural language text in a preset format;

[0027] A deep learning model acquisition module, used to construct a GPT deep learning model, and use the natural language text and the corresponding computer program generated in step S1 as samples to train the GPT deep learning model, wherein the GPT deep learning model is used to generate a corresponding computer program according to the natural language text of the digital twin mechanism model of the texturing machine;

[0028] The target model construction module is used to obtain the requirements of the digital twin mechanism model of the target texturing machine, convert the requirements into natural language text in a preset format, input it into the trained GPT deep learning model, and convert the output of the GPT deep learning model into a model to obtain the digital twin mechanism model of the target texturing machine.

[0029] Furthermore, the forward mapping module specifically includes:

[0030] A model conversion unit is used to obtain several digital twin mechanism models of historical texturing machines, convert the digital twin mechanism models of historical texturing machines into computer programs in a preset programming language using a model conversion tool, and perform tests, and delete or modify computer programs that fail the tests;

[0031] A text mapping relationship establishment unit is used to extract key text knowledge of the texturing machine according to the digital twin mechanism model of the historical texturing machine, build a text knowledge database of the texturing machine, and map the text knowledge in the text knowledge database of the texturing machine with the key information of the texturing machine in the computer program to form a text knowledge-computer program mapping relationship;

[0032] A text conversion unit is used to parse the computer program obtained by the model conversion unit, extract the key information of the texturing machine therein, and convert the key information of the texturing machine in the computer program into a natural language text in a preset format by using natural language generation technology and combining the text knowledge-computer program mapping relationship;

[0033] The verification unit is used to verify the natural language text obtained by the text conversion unit, and to modify and adjust the inconsistencies or errors therein.

[0034] Furthermore, the text conversion unit specifically includes:

[0035] A program parsing subunit is used to parse the computer program acquired by the model conversion unit and extract key information of the texturing machine therein;

[0036] The text generation subunit is used to match and fill in the key information of the texturing machine in the computer program with the text template through a pre-written text template, using rules or machine learning algorithms, according to the text knowledge-computer program mapping relationship, to generate a natural language text in a preset format, and check it after filling in; wherein the text template is a template that describes the different components and working states of the texturing machine.

[0037] Furthermore, the deep learning model acquisition module specifically includes:

[0038] A sample set generation unit, used for taking the natural language text and computer program of each historical texturing machine digital twin mechanism model as sample pairs, forming a sample set, and dividing it into a training set and a test set;

[0039] A deep learning model construction unit, used to construct a GPT deep learning model, wherein the GPT deep learning model includes a text encoder and a program parser connected to each other, wherein the text encoder is used to encode the input natural language text of the digital twin mechanism model of the texturing machine into a series of vector representations, wherein the vector representation captures the semantic and syntactic information in the natural language text; and the program parser is used to generate a corresponding computer program according to the vector representation;

[0040] The model training unit is used to pre-train the GPT deep learning model using the training set generated by the sample set generation unit, and after the pre-training is completed, use the test set to test the GPT deep learning model, and finally fine-tune the GPT deep learning model according to the test results until the GPT deep learning model meets the preset requirements.

[0041] Furthermore, the target model building module specifically includes:

[0042] A target model text acquisition unit is used to acquire the requirements of the digital twin mechanism model of the target texturing machine, fill the requirements of the digital twin mechanism model of the target texturing machine in a pre-written text template, and obtain a natural language text in a preset format;

[0043] A target program acquisition unit is used to input the natural language text acquired by the target model text acquisition unit into the trained GPT deep learning model, and the output is a computer program of the digital twin mechanism model of the target texturing machine;

[0044] The target model acquisition unit is used to convert the computer program of the digital twin mechanism model of the target texturing machine into the digital twin mechanism model of the target texturing machine by using a model conversion tool.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] (1) The present invention uses the GPT deep learning model to solve the rapid conversion of the digital twin mechanism model from text to the corresponding computer program;

[0047] (2) The present invention realizes the automatic generation of digital twin mechanism models. Compared with traditional manual modeling, it improves the efficiency of constructing mechanism models and reduces the cost of constructing mechanism models. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of a method for constructing a digital twin mechanism model of a texturing machine based on GPT provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0050] Embodiment 1

[0051] The embodiment of the present invention provides a method for constructing a digital twin mechanism model of a texturing machine based on GPT, such as Figure 1 As shown, the following steps are included:

[0052] S1. Obtain several historical digital twin mechanism models of texturing machines and convert them into natural language texts through the "model-program-text" forward multimodal mapping mechanism.

[0053] Among them, the forward multimodal mapping mechanism is used to first convert the digital twin mechanism model of the historical texturing machine into a computer program, and then convert the computer program into a natural language text in a preset format.

[0054] Specifically, step S1 includes the following steps:

[0055] S11. Obtain several historical texturing machine digital twin mechanism models, convert the historical texturing machine digital twin mechanism models into computer programs in a preset programming language (such as C / C++) using a model conversion tool, and test them, and delete or modify computer programs that fail the test. The model conversion tool can be a tool such as Simulink Coder or Embedded Coder, which converts the model into a program and verifies the generated program to ensure that it can accurately reflect the behavior of the digital twin mechanism model of the texturing machine.

[0056] S12. According to the historical digital twin mechanism model of the texturing machine, extract the key textual knowledge of the texturing machine, build a textual knowledge database of the texturing machine, and map the textual knowledge in the textual knowledge database with the key information of the texturing machine in the computer program to form a textual knowledge-computer program mapping relationship. Among them, the key textual knowledge of the texturing machine can be the structure, working principle, key parameters and their influence of the texturing machine. The key information of the texturing machine in the computer program can be variables, functions and expressions in the program.

[0057] S13, parsing the computer program obtained in step S11, extracting the key information of the texturing machine therein, and using natural language generation (NLG) technology, combined with the text knowledge-computer program mapping relationship, converting the key information of the texturing machine in the computer program into a natural language text in a preset format. Specifically, the key information of the texturing machine in the computer program can be matched and filled with the text template according to the text knowledge-computer program mapping relationship through a pre-written text template, using rules or machine learning algorithms, to generate a natural language text in a preset format, and then checking after filling; wherein the text template is a template that describes the different components and working states of the texturing machine. The inspection mainly performs grammatical and semantic checks on the generated text to ensure its accuracy and readability.

[0058] S14, verifying the natural language text obtained in step S13, and modifying and adjusting the inconsistencies or errors therein. The verification mainly compares the original model, the generated computer program and the converted text to ensure that the mapping relationship between them is accurate.

[0059] S2. Build a GPT deep learning model, and use the natural language text and corresponding computer program generated in step S1 as samples to train the GPT deep learning model.

[0060] This step specifically includes:

[0061] S21. The natural language text and computer program of each historical texturing machine digital twin mechanism model are taken as sample pairs to form a sample set, which is divided into a training set and a test set.

[0062] S22. Construct a GPT deep learning model, which is used to generate a corresponding computer program based on the natural language text of the digital twin mechanism model of the texturizing machine. The GPT deep learning model is a Transformer structure, specifically including a connected text encoder and a program parser, the text encoder is used to encode the input natural language text of the digital twin mechanism model of the texturizing machine into a series of vector representations, and the vector representation captures the semantic and syntactic information in the natural language text; the program parser is used to generate a corresponding computer program based on the vector representation.

[0063] S23. Use the training set of step S21 to pre-train the GPT deep learning model, and after the pre-training is completed, use the test set to test the GPT deep learning model, and finally fine-tune the GPT deep learning model according to the test results until the GPT deep learning model meets the preset requirements.

[0064] S3. Obtain the requirements of the digital twin mechanism model of the target texturing machine, convert the requirements into natural language text in a preset format, input it into the trained GPT deep learning model, and convert the output of the GPT deep learning model into a model to obtain the digital twin mechanism model of the target texturing machine.

[0065] This step specifically includes:

[0066] S31, obtaining the requirements of the digital twin mechanism model of the target texturing machine, and filling the requirements of the digital twin mechanism model of the target texturing machine in a pre-written text template to obtain a natural language text in a preset format. The text template is the template in step S13.

[0067] S32. Input the natural language text obtained in step S31 into the trained GPT deep learning model, and the output is a computer program of the digital twin mechanism model of the target texturing machine.

[0068] S33, converting the computer program of the digital twin mechanism model of the target texturing machine into the digital twin mechanism model of the target texturing machine by using a model conversion tool. The model conversion tool is a tool of the type of Simulink or the like.

[0069] In addition, after obtaining the digital twin mechanism model of the target texturing machine, it is also necessary to evaluate the accuracy of the digital twin mechanism model of the texturing machine, including the common index evaluation of the texturing machine mechanism and the characteristic index evaluation of the texturing machine mechanism.

[0070] Embodiment 2

[0071] The embodiment of the present invention provides a GPT-based digital twin mechanism model construction system for a texturing machine, comprising:

[0072] A forward mapping module is used to obtain a number of historical texturing machine digital twin mechanism models and convert them into natural language texts through a "model-program-text" forward multimodal mapping mechanism, wherein the forward multimodal mapping mechanism is used to first convert the historical texturing machine digital twin mechanism model into a computer program, and then convert the computer program into a natural language text in a preset format;

[0073] A deep learning model acquisition module, used to construct a GPT deep learning model, and use the natural language text and the corresponding computer program generated in step S1 as samples to train the GPT deep learning model, wherein the GPT deep learning model is used to generate a corresponding computer program according to the natural language text of the digital twin mechanism model of the texturing machine;

[0074] The target model construction module is used to obtain the requirements of the digital twin mechanism model of the target texturing machine, convert the requirements into natural language text in a preset format, input it into the trained GPT deep learning model, and convert the output of the GPT deep learning model into a model to obtain the digital twin mechanism model of the target texturing machine.

[0075] Wherein, the forward mapping module specifically includes:

[0076] A model conversion unit is used to obtain several digital twin mechanism models of historical texturing machines, convert the digital twin mechanism models of historical texturing machines into computer programs in a preset programming language using a model conversion tool, and perform tests, and delete or modify computer programs that fail the tests;

[0077] A text mapping relationship establishment unit is used to extract key text knowledge of the texturing machine according to the digital twin mechanism model of the historical texturing machine, build a text knowledge database of the texturing machine, and map the text knowledge in the text knowledge database of the texturing machine with the key information of the texturing machine in the computer program to form a text knowledge-computer program mapping relationship;

[0078] A text conversion unit is used to parse the computer program obtained by the model conversion unit, extract the key information of the texturing machine therein, and convert the key information of the texturing machine in the computer program into a natural language text in a preset format by using natural language generation technology and combining the text knowledge-computer program mapping relationship;

[0079] The verification unit is used to verify the natural language text obtained by the text conversion unit, and to modify and adjust the inconsistencies or errors therein.

[0080] Wherein, the text conversion unit specifically includes:

[0081] A program parsing subunit is used to parse the computer program acquired by the model conversion unit and extract key information of the texturing machine therein;

[0082] The text generation subunit is used to match and fill in the key information of the texturing machine in the computer program with the text template through a pre-written text template, using rules or machine learning algorithms, according to the text knowledge-computer program mapping relationship, to generate a natural language text in a preset format, and check it after filling in; wherein the text template is a template that describes the different components and working states of the texturing machine.

[0083] Among them, the deep learning model acquisition module specifically includes:

[0084] A sample set generation unit, used for taking the natural language text and computer program of each historical texturing machine digital twin mechanism model as sample pairs, forming a sample set, and dividing it into a training set and a test set;

[0085] A deep learning model construction unit, used to construct a GPT deep learning model, wherein the GPT deep learning model includes a text encoder and a program parser connected to each other, wherein the text encoder is used to encode the input natural language text of the digital twin mechanism model of the texturing machine into a series of vector representations, wherein the vector representation captures the semantic and syntactic information in the natural language text; and the program parser is used to generate a corresponding computer program according to the vector representation;

[0086] The model training unit is used to pre-train the GPT deep learning model using the training set generated by the sample set generation unit, and after the pre-training is completed, use the test set to test the GPT deep learning model, and finally fine-tune the GPT deep learning model according to the test results until the GPT deep learning model meets the preset requirements.

[0087] The target model building module specifically includes:

[0088] A target model text acquisition unit is used to acquire the requirements of the digital twin mechanism model of the target texturing machine, fill the requirements of the digital twin mechanism model of the target texturing machine in a pre-written text template, and obtain a natural language text in a preset format;

[0089] A target program acquisition unit is used to input the natural language text acquired by the target model text acquisition unit into the trained GPT deep learning model, and the output is a computer program of the digital twin mechanism model of the target texturing machine;

[0090] The target model acquisition unit is used to convert the computer program of the digital twin mechanism model of the target texturing machine into the digital twin mechanism model of the target texturing machine by using a model conversion tool.

[0091] The system provided in the embodiment of the present invention can be used to execute the method provided in the first embodiment of the present invention, and has the corresponding functions and beneficial effects of the execution method.

[0092] It is worth noting that in the embodiment of the above system, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0093] The embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, or of course, it can be implemented only by hardware, as long as the function or effect can be achieved.

Claims

1. A method for constructing a digital twin mechanism model of a texturing machine based on GPT, characterized in that: The steps include: S1. Obtain several digital twin mechanism models of historical texturing machines, and convert them into natural language texts through a "model-program-text" forward multimodal mapping mechanism, wherein the forward multimodal mapping mechanism is used to first convert the digital twin mechanism model of the historical texturing machine into a computer program, and then convert the computer program into a natural language text in a preset format; S2. Construct a GPT deep learning model, and use the natural language text and the corresponding computer program generated in step S1 as samples to train the GPT deep learning model, wherein the GPT deep learning model is used to generate a corresponding computer program according to the natural language text of the digital twin mechanism model of the texturing machine; S3. Obtain the requirements of the digital twin mechanism model of the target texturing machine, convert the requirements into natural language text in a preset format, input it into the trained GPT deep learning model, and convert the output of the GPT deep learning model into a model to obtain the digital twin mechanism model of the target texturing machine.

2. The method for constructing a digital twin mechanism model of a texturizing machine based on GPT according to claim 1, characterized in that: Step S1 specifically includes the following steps: S11. Obtain several digital twin mechanism models of historical texturing machines, convert the digital twin mechanism models of historical texturing machines into computer programs in a preset programming language using a model conversion tool, conduct tests, and delete or modify computer programs that fail the tests; S12. According to the digital twin mechanism model of the historical texturing machine, extract the key textual knowledge of the texturing machine, build a textual knowledge database of the texturing machine, and map the textual knowledge in the texturing machine textual knowledge database with the key information of the texturing machine in the computer program to form a textual knowledge-computer program mapping relationship; S13, parsing the computer program obtained in step S11, extracting key information of the texturing machine therein, and using natural language generation technology, combined with a text knowledge-computer program mapping relationship, converting the key information of the texturing machine in the computer program into a natural language text in a preset format; S14: Verify the natural language text obtained in step S13, and modify and adjust the inconsistencies or errors therein.

3. The method for constructing a digital twin mechanism model of a texturizing machine based on GPT according to claim 2, characterized in that: The method of using natural language generation technology and combining the text knowledge-computer program mapping relationship to convert the key information of the texturing machine in the computer program into a natural language text in a preset format specifically includes: Through a pre-written text template, using rules or machine learning algorithms, according to the text knowledge-computer program mapping relationship, the key information of the texturing machine in the computer program is matched and filled with the text template, and a natural language text in a preset format is generated, and the text is checked after filling in; wherein, the text template is a template that describes the different components and working states of the texturing machine.

4. The method for constructing a digital twin mechanism model of a texturizing machine based on GPT according to claim 1, characterized in that: Step S2 specifically includes the following steps: S21, taking the natural language text and computer program of each historical texturing machine digital twin mechanism model as sample pairs, forming a sample set, and dividing it into a training set and a test set; S22. Construct a GPT deep learning model, wherein the GPT deep learning model includes a text encoder and a program parser connected to each other, wherein the text encoder is used to encode the input natural language text of the digital twin mechanism model of the texturing machine into a series of vector representations, wherein the vector representation captures the semantic and syntactic information in the natural language text; and the program parser is used to generate a corresponding computer program according to the vector representation; S23. Use the training set of step S21 to pre-train the GPT deep learning model, and after the pre-training is completed, use the test set to test the GPT deep learning model, and finally fine-tune the GPT deep learning model according to the test results until the GPT deep learning model meets the preset requirements.

5. The method for constructing a digital twin mechanism model of a texturizing machine based on GPT according to claim 1, characterized in that: Step S3 specifically includes the following steps: S31. Obtaining the requirements of the digital twin mechanism model of the target texturing machine, filling the requirements of the digital twin mechanism model of the target texturing machine into a pre-written text template to obtain a natural language text in a preset format; S32, inputting the natural language text obtained in step S31 into the trained GPT deep learning model, and outputting a computer program of the digital twin mechanism model of the target texturing machine; S33. Convert the computer program of the digital twin mechanism model of the target texturing machine into the digital twin mechanism model of the target texturing machine by using a model conversion tool.

6. A digital twin mechanism model construction system for a texturing machine based on GPT, characterized in that: include: A forward mapping module is used to obtain several historical texturing machine digital twin mechanism models and convert them into natural language texts through a "model-program-text" forward multimodal mapping mechanism, wherein the forward multimodal mapping mechanism is used to first convert the historical texturing machine digital twin mechanism model into a computer program, and then convert the computer program into a natural language text in a preset format; A deep learning model acquisition module, used to construct a GPT deep learning model, and use the natural language text and the corresponding computer program generated in step S1 as samples to train the GPT deep learning model, wherein the GPT deep learning model is used to generate a corresponding computer program according to the natural language text of the digital twin mechanism model of the texturing machine; The target model construction module is used to obtain the requirements of the digital twin mechanism model of the target texturing machine, convert the requirements into natural language text in a preset format, input it into the trained GPT deep learning model, and convert the output of the GPT deep learning model into a model to obtain the digital twin mechanism model of the target texturing machine.

7. The GPT-based digital twin mechanism model construction system for texturizing machines according to claim 6 is characterized in that: The forward mapping module specifically includes: A model conversion unit is used to obtain several digital twin mechanism models of historical texturing machines, convert the digital twin mechanism models of historical texturing machines into computer programs in a preset programming language using a model conversion tool, and perform tests, and delete or modify computer programs that fail the tests; A text mapping relationship establishment unit is used to extract key text knowledge of the texturing machine according to the digital twin mechanism model of the historical texturing machine, build a text knowledge database of the texturing machine, and map the text knowledge in the text knowledge database of the texturing machine with the key information of the texturing machine in the computer program to form a text knowledge-computer program mapping relationship; A text conversion unit is used to parse the computer program obtained by the model conversion unit, extract the key information of the texturing machine therein, and convert the key information of the texturing machine in the computer program into a natural language text in a preset format by using natural language generation technology and combining the text knowledge-computer program mapping relationship; The verification unit is used to verify the natural language text obtained by the text conversion unit, and to modify and adjust the inconsistencies or errors therein.

8. The GPT-based digital twin mechanism model construction system for texturizing machines according to claim 7 is characterized in that: The text conversion unit specifically includes: A program parsing subunit is used to parse the computer program acquired by the model conversion unit and extract key information of the texturing machine therein; The text generation subunit is used to match and fill in the key information of the texturing machine in the computer program with the text template through a pre-written text template, using rules or machine learning algorithms, according to the text knowledge-computer program mapping relationship, to generate a natural language text in a preset format, and check it after filling in; wherein the text template is a template that describes the different components and working states of the texturing machine.

9. The GPT-based digital twin mechanism model construction system for texturizing machines according to claim 6 is characterized in that: The deep learning model acquisition module specifically includes: A sample set generation unit, used for taking the natural language text and computer program of each historical texturing machine digital twin mechanism model as sample pairs, forming a sample set, and dividing it into a training set and a test set; A deep learning model construction unit, used to construct a GPT deep learning model, wherein the GPT deep learning model includes a text encoder and a program parser connected to each other, wherein the text encoder is used to encode the input natural language text of the digital twin mechanism model of the texturing machine into a series of vector representations, wherein the vector representation captures the semantic and syntactic information in the natural language text; and the program parser is used to generate a corresponding computer program according to the vector representation; The model training unit is used to pre-train the GPT deep learning model using the training set generated by the sample set generation unit, and after the pre-training is completed, use the test set to test the GPT deep learning model, and finally fine-tune the GPT deep learning model according to the test results until the GPT deep learning model meets the preset requirements.

10. The GPT-based digital twin mechanism model construction system for texturizing machines according to claim 6, characterized in that: The target model building module specifically includes: A target model text acquisition unit is used to acquire the requirements of the digital twin mechanism model of the target texturing machine, fill the requirements of the digital twin mechanism model of the target texturing machine in a pre-written text template, and obtain a natural language text in a preset format; A target program acquisition unit is used to input the natural language text acquired by the target model text acquisition unit into the trained GPT deep learning model, and the output is a computer program of the digital twin mechanism model of the target texturing machine; The target model acquisition unit is used to convert the computer program of the digital twin mechanism model of the target texturing machine into the digital twin mechanism model of the target texturing machine by using a model conversion tool.

Citation Information

Patent Citations

  • Digital twin three-dimensional scene construction method, device, equipment and medium

    CN116310148A

  • Image generation method and device, electronic equipment, storage medium and program product

    CN117315070A

  • Navigation broadcast data processing method and device, computer equipment and storage medium

    CN117933195A

  • Method and system for constructing language transformation assistant based on generative model

    CN118332000A

  • Method and device for converting natural language statements into SQL statements and medium

    CN118503272A