Design method, system, medium and equipment based on large model assistance

Abstract

The invention relates to a design method and system based on large model assistance, a medium and equipment. The method comprises the steps that design demand information is acquired; analyzing the design demand information based on a natural language processing technology to obtain analysis data, and converting the analysis data to obtain initial structure data; analyzing, extracting and converting the initial structure data based on the large model to obtain structured initial design data; performing visual display on the initial design data; receiving feedback suggestions and performing iterative optimization processing according to the feedback suggestions to obtain target design data; and performing standardized conversion on the target design data to obtain standard design data. According to the method, a traditional document-driven design mode is converted into a driving mode with the structural data and the design model as the core by utilizing a large model, so that the efficiency and safety problems of the existing design mode are fundamentally solved, and the design safety, efficiency and quality are remarkably improved; and a solid technical support is provided for digital engineering reconstruction in key fields such as nuclear power.

Description

Technical Field

[0001] This invention relates to the technical field of nuclear power plant engineering modification, and more specifically, to a design method, system, medium, and equipment based on large model-assisted design. Background Technology

[0002] For a long time, engineering modifications during the operation phase of nuclear power units have followed the traditional document-based design model. The core of this model is that all modification work revolves around paper or electronic documents and is completed through a rigorous approval process, from requirement identification, detailed design, strict review, to final construction implementation and verification.

[0003] In this document-driven model, the renovation project first proposes and initiates a project based on operational experience feedback, equipment aging, or safety performance improvements. This is followed by the crucial detailed design and documentation phase. The design team, adhering to established standards, prepares comprehensive professional documents, including design reports, system drawings, equipment specifications, construction plans, and safety analysis reports, and implements strict version control and archiving management. After design completion, the plan undergoes multi-disciplinary review and hierarchical approval, involving expert review and approval at each level. Subsequently, the renovation enters the construction and on-site execution phase, strictly following the approved drawings and plans. Any design changes must adhere to strict procedures and update the documentation. Finally, after the renovation is completed, commissioning, acceptance, and operation are conducted, and as-built documentation is updated based on actual conditions. Post-project evaluation and experience feedback are also performed after commissioning.

[0004] While traditional document design models offer advantages in terms of rigor and traceability in ensuring the safety and quality of nuclear power plant retrofits, they have become increasingly inadequate to meet current needs and have exposed numerous problems with the deepening of modernization and digital transformation. This model leads to information silos and data redundancy, with design information scattered across numerous unstructured documents, hindering efficient sharing, reuse, and traceability, and resulting in significant duplication of information. Version control difficulties and low collaboration efficiency are common problems; version confusion easily arises during multi-party collaboration, severely hindering efficiency and extending the cycle. Furthermore, the lack of dynamic tracking management makes it impossible to grasp the real-time status of interfaces and technologies, resulting in delayed analysis of the impact of changes. Design solutions are difficult to reuse effectively; due to the scattered and poorly structured information, past experience is difficult to utilize systematically. Most importantly, this model lacks a unified underlying platform and fails to support automated analysis and artificial intelligence (AI)-assisted design, severely restricting the potential for nuclear power plant retrofits to develop towards higher efficiency, greater precision, and greater intelligence. These deep-seated problems seriously challenge the efficiency, accuracy, and safety of nuclear power plant retrofits. For example, manual processing of design requirements is time-consuming, and data inconsistencies exist between multiple systems. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a design method, system, medium and device based on large model-assisted design, which addresses the problems existing in the prior art.

[0006] The technical solution adopted by this invention to solve its technical problem is: to construct a design method based on large model assistance, including the following steps:

[0007] Obtain design requirements information;

[0008] The design requirement information is parsed using natural language processing technology to obtain parsed data, and the parsed data is then transformed to obtain initial structure data.

[0009] Based on the large model, the initial structural data is parsed, extracted, and transformed to obtain structured initial design data;

[0010] The initial design data is then visualized.

[0011] Receive feedback suggestions and perform iterative optimization based on the feedback suggestions to obtain target design data;

[0012] The target design data is standardized to obtain standard design data.

[0013] In the large model-assisted design method described in this invention, the method further includes:

[0014] The standard design data is then output to the downstream system.

[0015] In the large model-assisted design method described in this invention, the large model includes: a large language model or a multimodal large model;

[0016] The process of parsing, extracting, and transforming the initial structural data based on the large model to obtain structured initial design data includes:

[0017] Based on the large language model, the initial structural data is deeply understood and knowledge is extracted to obtain the initial extracted data.

[0018] The initial extracted data is subjected to structured processing based on preset design patterns and rules to obtain structured initial design data;

[0019] Alternatively, the step of parsing, extracting, and transforming the initial structural data based on a large model to obtain structured initial design data includes:

[0020] Based on the multimodal large model, the initial structural data is directly parsed and extracted, and multimodal information is fused to obtain the structured initial design data.

[0021] In the large model-assisted design method described in this invention, the step of performing deep understanding and knowledge extraction on the initial structural data based on the large language model to obtain the initial extracted data includes:

[0022] The large language model learns from pre-trained engineering design knowledge, industry standards, real-time learning capabilities, and historical engineering renovation design schemes to acquire training patterns and experience.

[0023] Based on the training mode and experience, the initial structure data is deeply understood and knowledge is extracted to obtain the initial extracted data.

[0024] In the large model-assisted design method described in this invention, the method further includes:

[0025] The large language model processes the initial structural data according to the preset design template directory requirements to obtain the document chapter content.

[0026] In the large model-assisted design method described in this invention, the step of receiving feedback suggestions and performing iterative optimization based on the feedback suggestions to obtain target design data includes:

[0027] Receive the feedback suggestions;

[0028] The large model analyzes and reasones the feedback suggestions to identify the data to be processed in the initial design data;

[0029] Based on the data to be processed, and combined with a rule engine, multi-objective optimization algorithm and automated compliance checks, the initial design data is optimized to obtain the target design data and output intelligent optimization suggestions.

[0030] In the large model-assisted design method described in this invention, the intelligent optimization suggestions include:

[0031] Recommendations for data structure optimization, model logic improvement schemes, adjustment suggestions to meet specified security standards or specifications, data comparison and risk assessment of different design schemes, and fine-tuning of key design parameters to meet specified security or performance indicators.

[0032] This invention also provides a design system based on large model-assisted design, comprising:

[0033] The human-computer interaction module is used to perform human-computer interaction functions;

[0034] The natural language processing module is used to parse the design requirement information based on natural language processing technology to obtain parsed data, and to transform the parsed data to obtain initial structure data;

[0035] The large model reasoning and generation module is used to parse, extract, and transform the initial structural data based on the large model to obtain structured initial design data, and to standardize and transform the target design data to obtain standard design data; and to generate intelligent optimization suggestions based on feedback suggestions.

[0036] The visualization and verification module is used to visualize the initial design data and allow users to review, verify, and receive feedback suggestions.

[0037] The feedback and optimization module is used to receive feedback suggestions and perform iterative optimization processing based on the feedback suggestions to obtain target design data.

[0038] The data interface module is used to transmit the standard design data to downstream systems;

[0039] The design knowledge base and historical data management module is used to manage design knowledge and historical data.

[0040] The present invention also provides a storage medium storing a computer program adapted for loading by a processor to perform the steps of the large model-aided design method as described above.

[0041] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the large model-aided design method described above by calling the computer program stored in the memory.

[0042] The design method, system, medium, and device based on large model-assisted design of the present invention have the following beneficial effects: including the following steps: acquiring design requirement information; parsing the design requirement information using natural language processing technology to obtain parsed data, and transforming the parsed data to obtain initial structure data; parsing, extracting, and transforming the initial structure data based on a large model to obtain structured initial design data; visualizing the initial design data; receiving feedback suggestions and performing iterative optimization processing based on the feedback suggestions to obtain target design data; and standardizing the target design data to obtain standard design data. This invention utilizes a large model to transform the traditional document-driven design pattern into a driven pattern centered on structured data and design models, fundamentally solving the efficiency and security issues of existing design patterns. It fully leverages the advantages of large models in integrating massive amounts of engineering knowledge, complex reasoning, and generating structured data, enabling rapid conversion from natural language requirements to precise design data and models. Furthermore, it allows for multiple rounds of intelligent iterative optimization and security verification based on designer feedback. The large model can deeply learn from historical engineering renovation design schemes and automatically generate high-quality document chapters based on design templates, significantly improving document writing efficiency and consistency while ensuring that the generated content remains synchronized with structured data and models. Compared to the traditional document-driven design pattern, the large model-assisted design method and system of this invention have the following advantages: A fundamental shift in design patterns: from a decentralized, unstructured document-driven design pattern to a unified, structured, and traceable data structure or model-driven design pattern, fundamentally solving problems such as information silos, data redundancy, and version control difficulties, laying the foundation for digital transformation. Significantly Improved Design Efficiency: Through natural language interaction, intelligent generation of large models, and automated document chapter generation, the design cycle is greatly shortened, reducing designers' investment in repetitive operations and information organization, and significantly improving the overall efficiency of engineering transformation design. Significantly Enhanced Design Safety and Quality: The large model can integrate historical data, industry standards, and safety regulations to conduct intelligent compliance checks and risk predictions, providing more accurate design suggestions and risk assessments, and assisting designers in rigorous verification and validation. This effectively reduces human error, anticipates potential risks, ensures the inherent safety and compliance of design solutions, and improves the accuracy, consistency, and safety of design solutions, which is particularly important in fields with high safety requirements such as nuclear power. Enhanced Reusability and Manageability of Design Solutions: Design results are stored in the form of structured data or models, facilitating management, retrieval, and reuse. This breaks through the limitations of traditional document-based design solutions, increasing the value of knowledge assets. Provides a Solid Foundation for AI-Assisted Design: A unified data and model platform has been established, providing a high-quality, structured data foundation for subsequent higher-level AI-assisted analysis, intelligent decision-making, and digital twin applications.Achieving deep human-machine collaboration: An efficient and natural collaborative working mode has been established between designers and intelligent systems. The large model becomes the designer's intelligent "co-pilot," freeing designers from tedious document work so that they can focus on creativity, verification, and key decisions, and jointly complete high-safety and high-quality design tasks. Attached Figure Description

[0043] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings:

[0044] Figure 1 This is a flowchart illustrating an embodiment of the large model-assisted design method provided by the present invention;

[0045] Figure 2 This is a flowchart illustrating Embodiment 2 of the design method based on large model assistance provided by the present invention;

[0046] Figure 3 This is a logical block diagram of the design system based on large model assistance provided by the present invention;

[0047] Figure 4 This is a schematic diagram of the overall architecture of the design system based on large model assistance provided by the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] To address the problems of inefficiency, information silos, data redundancy, version control difficulties, low collaboration efficiency, disconnect between security analysis and design, and lack of a foundational platform for AI-assisted design in traditional document-driven engineering modification and design models, this invention provides a design method based on large-model assistance. (Reference) Figure 1 , Figure 1 A preferred embodiment of the present invention is shown.

[0050] Specifically, such as Figure 1 As shown, this large model-assisted design method includes the following steps:

[0051] Step S101: Obtain design requirements information.

[0052] Specifically, design requirements information may include, but is not limited to: design requirements for the engineering modification, objectives of the engineering modification, constraints of the engineering modification, functional requirements of the engineering modification, safety specifications of the engineering modification, and technical parameters of the engineering modification. In this embodiment of the invention, design requirements information can be obtained by the designer through direct input of natural language (text, voice, etc.). For example, it can be directly input on the human-computer interaction module 100.

[0053] Step S102: Parse the design requirement information using natural language processing technology to obtain parsed data, and transform the parsed data to obtain initial structure data.

[0054] Specifically, in this embodiment of the invention, after the designer inputs design requirement information via natural language, the system's built-in Natural Language Processing (NLP) module can perform deep analysis on the input design requirement information to obtain analytical data (including but not limited to key design parameters, entities, relationships, and intents). Furthermore, after obtaining the analytical data, it is transformed into structured or semi-structured data that can be understood by a large model, i.e., initial structured data. For example, JSON or XML formats may be used, which contain initially identified design elements and their relationships.

[0055] Step S103: Based on the large model, the initial structural data is parsed, extracted, and transformed to obtain structured initial design data.

[0056] Optionally, in this embodiment of the invention, the large model includes a large language model (LLM) or a multimodal large model (LMM). This large language model or multimodal large model can be a model based on a general model architecture (such as GPT-4, Transformer, etc.) and pre-trained and fine-tuned for the nuclear power engineering design field. Specifically, the parsing of initial structural data can be driven by either a large language model or a multimodal large model to drive the generation of design data / models.

[0057] In some embodiments, parsing, extracting, and transforming initial structural data based on a large model to obtain structured initial design data includes: performing deep understanding and knowledge extraction on the initial structural data based on a large language model to obtain initial extracted data; and performing structured processing on the initial extracted data based on pre-design patterns and rules to obtain structured initial design data. Specifically, performing deep understanding and knowledge extraction on the initial structural data based on a large language model to obtain initial extracted data includes: the large language model learning based on pre-trained engineering design knowledge, industry standards, real-time learning capabilities, and historical engineering renovation design schemes to obtain training patterns and experience; and performing deep understanding and knowledge extraction on the initial structural data based on the training patterns and experience to obtain initial extracted data. Optionally, in embodiments of the present invention, the initial design data includes, but is not limited to: structured, machine-readable design data structures, design models (such as functional models, logical models, physical models, etc.), key parameter sets, or code snippets (such as data model definition scripts), etc.

[0058] Specifically, the initial structured data, parsed by the natural language processing module 200 and represented in a structured or semi-structured manner, is sent to the core large-scale model inference and generation module 300. Based on its pre-trained engineering design knowledge, industry standards, real-time learning capabilities, and patterns and experiences learned from a vast amount of historical engineering renovation design schemes, the large-scale model inference and generation module 300 performs deep understanding and knowledge extraction on the parsed natural language requirements, forming initial extracted data. Subsequently, the system performs structured processing on this initial extracted data according to preset design patterns, data model definitions, and rules, transforming it into structured, machine-readable design data structures, design models (such as functional models, logical models, physical models, etc.), key parameter sets, or code snippets (such as data model definition scripts). Further, in this embodiment of the invention, the large language model processes the initial structured data according to preset design template directory requirements to obtain document chapter content. That is, the large language model can automatically generate corresponding document chapter content based on preset design template directory requirements, such as design background, technical routes, and scheme descriptions; this content is generated based on structured data and models. For example, if the design requirement is "design a nuclear-grade cooling water pump with a flow rate of 100 m³ / h and a head of 50 m", the natural language processing module will parse out key information such as "flow rate", "head", "cooling water pump", and "nuclear-grade" and convert it into initial structured data, such as {'component':'pump','type':'cooling water','grade':'nuclear','flow_rate':'100 m³ / h','head':'50 m'}. The large model inference and generation module will generate structured initial design data based on this initial structural data, combined with nuclear power pump design specifications and historical data. This data may include recommended pump models, materials, interface dimensions, preliminary performance curves, etc., for example, {'pump_model':'CRF-TD022','material':'stainless steel 316L','inlet_diameter':'DN200','outlet_diameter':'DN150','power_rating':'55kW','NPSH_required':'5m'}.

[0059] Alternatively, in other embodiments, parsing, extracting, and transforming initial structural data based on a large model to obtain structured initial design data includes: directly parsing and extracting initial structural data and fusing multimodal information based on a multimodal large model to obtain structured initial design data. Specifically, for a multimodal large model, it can directly process initial structural data containing multimodal information such as text, images, and charts, and directly generate initial design data, i.e., a preliminary structured data representation or conceptual model diagram, through multimodal information fusion and reasoning.

[0060] Step S104: Visualize the initial design data.

[0061] In this embodiment of the invention, the generated initial design data (including design data structures or design models, etc.) can be visualized and displayed through the visualization and verification module 400, thereby presenting the generated preliminary design data structure or design model to the designer. The displayed initial design data includes, but is not limited to: data tables, model diagrams, flowcharts, or automatically generated document previews, etc.

[0062] The initial design data is displayed through visualization and verification models, and designers review, verify, modify, and annotate the generated initial design data, providing feedback and suggestions again through natural language or graphical interfaces. This step ensures the safety and quality of the design.

[0063] Step S105: Receive feedback suggestions and perform iterative optimization based on the feedback suggestions to obtain target design data.

[0064] In this embodiment of the invention, receiving feedback suggestions and performing iterative optimization based on those suggestions to obtain target design data includes: receiving feedback suggestions; a large model parsing and reasoning on the feedback suggestions to identify data to be processed in the initial design data; and optimizing the initial design data based on the data to be processed, combined with a rule engine, a multi-objective optimization algorithm, and automated compliance checks, to obtain target design data and output intelligent optimization suggestions. The intelligent optimization suggestions include:

[0065] Data structure optimization suggestions (such as improving data consistency and reducing redundancy);

[0066] Model logic improvement schemes (such as improving functional completeness and meeting performance indicators);

[0067] Adjustment recommendations that comply with specified safety standards or specifications;

[0068] Data comparison and risk assessment of different design schemes;

[0069] Fine-tune key design parameters to meet specified safety or performance requirements.

[0070] Specifically, the feedback suggestions from designers are further analyzed and input into the larger model along with the current design data / model. The larger model then uses this feedback to reason, comparing the design with best practices, industry standards, and safety regulations in the design knowledge base, or identifying potential risks and inconsistencies through internal simulation analysis. This process identifies design deficiencies, potential risks, or inconsistencies and provides intelligent optimization suggestions. These optimization suggestions aim to improve the safety, efficiency, and quality of the design, especially in the nuclear power sector, ensuring that the design meets the most stringent safety standards.

[0071] Step S105 can be repeated until the designer confirms the accuracy and completeness of the design data / model.

[0072] Step S106: Standardize and transform the target design data to obtain standard design data.

[0073] Specifically, the generated target design data is the finalized design data structure or design model, which can be further converted into standardized data formats or model description languages ​​to obtain standard design data. Standardized data formats can include SysML models, JSON, XML, OWL, etc. Model description languages ​​can include SysML, UML, etc. This conversion process can be completed directly by the large model, ensuring the output conforms to the preset standardized pattern, or it can be formatted by a dedicated data conversion module based on instructions generated by the large model. By converting to a standard data format or model description language, it can be easily imported into an enterprise's internal engineering management system, safety analysis platform, or digital twin platform.

[0074] like Figure 2 As shown, Figure 2 This invention provides another preferred embodiment of a design method based on a large model.

[0075] like Figure 2 As shown, this embodiment Figure 1 Building upon the example, the following steps are further included:

[0076] Step S107: Output the standard design data to the downstream system.

[0077] Specifically, in this embodiment of the invention, the system can provide an open data interface module 600, through which standard design data can be directly transmitted to relevant tools for security assessment, reliability analysis or performance simulation, and the analysis results can be fed back to the large model to further guide optimization and verification.

[0078] refer to Figure 3 , Figure 3This paper illustrates a preferred embodiment of the large model-assisted design system provided by the present invention. Figure 4 The overall architecture of the large model-assisted design system provided by this invention is shown.

[0079] like Figure 3 As shown and Figure 4 As shown, this large model-assisted design system includes:

[0080] The human-computer interaction module 100 is used to perform human-computer interaction functions.

[0081] Specifically, the human-computer interaction module 100 provides a user-friendly graphical interface that supports natural language input (text and voice), 2D / 3D visualization of design data / models, preview of automatically generated documents, and direct modification and feedback from designers on the design.

[0082] The natural language processing module 200 is used to parse the design requirement information based on natural language processing technology to obtain parsed data, and to transform the parsed data to obtain initial structure data.

[0083] Specifically, the Natural Language Processing module 200 parses the natural language instructions input by the design quantity and converts them into structured or semi-structured design intents and parameters.

[0084] The large model reasoning and generation module 300 is used to parse, extract, and transform the initial structural data based on the large model to obtain structured initial design data, and to standardize and transform the target design data to obtain standard design data; and to generate intelligent optimization suggestions based on feedback suggestions.

[0085] The large model reasoning and generation module 300 is the core of the large model-assisted design system of this invention, and it can be a large language model or a multimodal large model. This large model reasoning and generation module 300 is mainly used for: understanding the parsed design requirements; performing reasoning based on massive amounts of engineering design knowledge and historical data; generating preliminary design data structures, design models, or code; learning and extracting patterns from historical engineering renovation design schemes and automatically generating corresponding document chapters according to the design template's directory requirements; and iteratively optimizing based on designer feedback and providing intelligent optimization suggestions.

[0086] The visualization and verification module 400 is used to visualize the initial design data and allow users to review, verify, and receive feedback.

[0087] In this embodiment of the invention, the visualization and verification module 400 is mainly used to transform the abstract design data or model generated from the large model into intuitive charts, diagrams or document previews for designers to review and verify.

[0088] The feedback and optimization module 500 is used to receive feedback suggestions and perform iterative optimization based on the feedback suggestions to obtain target design data.

[0089] In this embodiment of the invention, the feedback and optimization module 500 is mainly used to receive feedback suggestions (modifications and evaluations) during the design process, transform them into inputs that the large model can understand, and coordinate the large model to generate optimization suggestions and perform security verification in the next round.

[0090] The data interface module 600 is used to transmit standard design data to downstream systems.

[0091] In this embodiment of the invention, the data interface module 600 is mainly used to manage and store the generated structured design data and design models, and to provide a standardized data interface to realize the interconnection between this system and downstream systems such as enterprise content engineering management system, security analysis platform, and digital twin platform.

[0092] The Design Knowledge Base and Historical Data Management Module 700 is used to manage design knowledge and historical data.

[0093] In this embodiment of the invention, the design knowledge base and historical data management module 700 is mainly used to store structured and unstructured knowledge related to engineering design, including industry standards, engineering specifications, safety regulations, material properties, design principles, past renovation cases, and historical engineering renovation design schemes. This design knowledge base and historical data management module 700 is an important data source for AI training and learning, supporting the extraction of patterns and experiences from large models.

[0094] Specifically, the specific coordination process between the units in the large model-assisted design system can be referred to the above-mentioned large model-assisted design method, and will not be repeated here.

[0095] This invention utilizes a large model to transform the traditional document-driven design pattern into a driving pattern centered on structured data and design models, fundamentally solving the efficiency and security issues of existing design patterns. It fully leverages the advantages of large models in integrating massive amounts of engineering knowledge, complex reasoning, and generating structured data, enabling rapid conversion from natural language requirements to precise design data and models. Furthermore, it can perform multiple rounds of intelligent iterative optimization and security checks based on designer feedback. The large model can deeply learn from historical engineering renovation design schemes and automatically generate high-quality document chapters based on design templates, significantly improving document writing efficiency and consistency while ensuring that the generated content remains synchronized with the structured data and models.

[0096] Compared with traditional document-driven design patterns, the design method and system based on large model assistance of the present invention have the following advantages:

[0097] Achieving a fundamental shift in design patterns: from a decentralized, unstructured document-driven design pattern to a unified, structured, and traceable data structure or model-driven design pattern, fundamentally solving problems such as information silos, data redundancy, and version control difficulties, laying the foundation for digital transformation.

[0098] Significantly improve design efficiency: Through natural language interaction, intelligent generation of large models, and automated document chapter generation, the design cycle is greatly shortened, reducing designers' investment in repetitive operations and information organization, and significantly improving the overall efficiency of engineering transformation design.

[0099] Significantly improve design safety and quality: The large model can integrate historical data, industry standards and safety regulations to conduct intelligent compliance checks and risk predictions, provide more accurate design suggestions and risk assessments, and assist designers in conducting rigorous checks and verifications. This effectively reduces human error, anticipates potential risks, ensures the inherent safety and compliance of design solutions, and improves the accuracy, consistency and safety of design solutions.

[0100] This is particularly significant in fields with high safety requirements, such as nuclear power. It enhances the reusability and manageability of design solutions: design deliverables are stored in the form of structured data or models, facilitating management, retrieval, and reuse. This breaks through the limitations of traditional document-based design solutions, which are difficult to reuse, and increases the value of knowledge assets.

[0101] Provides a solid foundation for AI-assisted design: A unified data and model platform has been established, providing a high-quality, structured data foundation for subsequent higher-level AI-assisted analysis, intelligent decision-making, and digital twin applications.

[0102] Achieving deep human-machine collaboration: An efficient and natural collaborative working mode has been established between designers and intelligent systems. The large model becomes the designer's intelligent "co-pilot," freeing designers from tedious document work so that they can focus on creativity, verification, and key decisions, and jointly complete high-safety and high-quality design tasks.

[0103] Furthermore, an electronic device according to the present invention includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program to implement the large model-aided design method as described above. Specifically, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, when the computer program is downloaded, installed, and executed by an electronic device, it performs the functions defined in the methods of the embodiments of the present invention. The electronic device in the present invention can be a terminal such as a laptop, desktop computer, tablet computer, or smartphone, or it can be a server.

[0104] Furthermore, one type of storage medium of the present invention stores a computer program thereon, which, when executed by a processor, implements any of the above-mentioned large model-aided design methods. Specifically, it should be noted that the storage medium described above in the present invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, wherein computer-readable program code is carried. The transmitted data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0105] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0106] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0107] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0108] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0109] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They do not limit the scope of protection of the present invention. All equivalent changes and modifications made within the scope of the claims of the present invention should fall within the scope of the claims of the present invention.

Claims

1. A design method based on large model-assisted design, characterized in that, Includes the following steps: Obtain design requirements information; The design requirement information is parsed using natural language processing technology to obtain parsed data, and the parsed data is then transformed to obtain initial structure data. Based on the large model, the initial structural data is parsed, extracted, and transformed to obtain structured initial design data; The initial design data is then visualized. Receive feedback suggestions and perform iterative optimization based on the feedback suggestions to obtain target design data; The target design data is standardized to obtain standard design data.

2. The design method based on large model-assisted design according to claim 1, characterized in that, The method further includes: The standard design data is then output to the downstream system.

3. The design method based on large model assistance according to claim 1, characterized in that, The large model includes: a large language model or a multimodal large model; The process of parsing, extracting, and transforming the initial structural data based on the large model to obtain structured initial design data includes: Based on the large language model, the initial structural data is deeply understood and knowledge is extracted to obtain the initial extracted data. The initial extracted data is subjected to structured processing based on pre-design patterns and rules to obtain structured initial design data; Alternatively, the step of parsing, extracting, and transforming the initial structural data based on a large model to obtain structured initial design data includes: Based on the multimodal large model, the initial structural data is directly parsed and extracted, and multimodal information is fused to obtain the structured initial design data.

4. The design method based on large model assistance according to claim 3, characterized in that, The process of performing deep understanding and knowledge extraction on the initial structural data based on the large language model to obtain the initial extracted data includes: The large language model learns from pre-trained engineering design knowledge, industry standards, real-time learning capabilities, and historical engineering renovation design schemes to acquire training patterns and experience. Based on the training mode and experience, the initial structure data is deeply understood and knowledge is extracted to obtain the initial extracted data.

5. The design method based on large model assistance according to claim 3, characterized in that, The method further includes: The large language model processes the initial structural data according to the preset design template directory requirements to obtain the document chapter content.

6. The design method based on large model-assisted design according to claim 1, characterized in that, The process of receiving feedback suggestions and performing iterative optimization based on those suggestions to obtain target design data includes: Receive the feedback suggestions; The large model analyzes and reasones the feedback suggestions to identify the data to be processed in the initial design data; Based on the data to be processed, and combined with a rule engine, multi-objective optimization algorithm and automated compliance checks, the initial design data is optimized to obtain the target design data and output intelligent optimization suggestions.

7. The design method based on large model-assisted design according to claim 6, characterized in that, The intelligent optimization suggestions include: Recommendations for data structure optimization, model logic improvement schemes, adjustment suggestions to meet specified security standards or specifications, data comparison and risk assessment of different design schemes, and fine-tuning of key design parameters to meet specified security or performance indicators.

8. A design system based on large model-assisted design, characterized in that, include: The human-computer interaction module is used to perform human-computer interaction functions; The natural language processing module is used to parse the design requirement information based on natural language processing technology to obtain parsed data, and to transform the parsed data to obtain initial structure data; The large model reasoning and generation module is used to parse, extract and transform the initial structural data based on the large model to obtain structured initial design data, and to standardize and transform the target design data to obtain standard design data. It also generates intelligent optimization suggestions based on feedback. The visualization and verification module is used to visualize the initial design data and allow users to review, verify, and receive feedback suggestions. The feedback and optimization module is used to receive feedback suggestions and perform iterative optimization processing based on the feedback suggestions to obtain target design data. The data interface module is used to transmit the standard design data to downstream systems; The design knowledge base and historical data management module is used to manage design knowledge and historical data.

9. A storage medium, characterized in that, The storage medium stores a computer program adapted for loading by a processor to perform the steps of the large model-aided design method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the large model-assisted design method as described in any one of claims 1 to 7 by calling the computer program stored in the memory.

Images