Client demand mining method and system in high polymer material modification research and development

Abstract

The invention discloses a customer demand mining method and system in modification research and development of a high polymer material. The mining method comprises the following steps: receiving a user demand, obtaining core elements, outline content and viewpoint content, generating a query keyword and returning external evidence, modifying and optimizing the viewpoint content, and filling or expanding the viewpoint content into a complete answer as output; querying material data from a material characteristic database according to the customer demand analysis report, performing material characteristic simulation, and generating a material characteristic analysis report; and based on the dynamic report, the customer demand analysis report and the material characteristic analysis report, generating at least one modification scheme by using the high polymer material modified large model as a base model, evaluating the feasibility of the scheme by using simulation software, predicting the performance and potential problems of different modification schemes by applying machine learning, and generating a modification scheme report. By realizing automation and intelligentization of the research and development process, the research and development efficiency of high polymer material modification research and development is improved, and the development period is shortened.

Description

Technical Field

[0001] The present invention belongs to the technical field of data mining, and in particular relates to a method and system for mining customer demands in the research and development of polymer material modification. Background Art

[0002] With the widespread application of polymer materials in various industries, the market demand for high-performance, high-reliability and cost-effective polymer materials is increasing. As an important means to improve material performance and expand application areas, polymer material modification technology has received widespread attention. For example, the Japanese invention patent application with publication number JP2016095218A discloses a simulation method for modified polymer materials, which uses a computer model to evaluate the dispersibility of modified polymer materials and fillers.

[0003] However, the traditional polymer material modification research and development process usually has the following problems: polymer materials have a wide range of applications, and user needs are diverse and complex. Traditional methods mainly rely on customer feedback and market research, but it is often difficult to fully and accurately obtain users' specific needs, resulting in unclear R&D directions and low R&D efficiency; the polymer material market changes rapidly, and timely and accurate acquisition and analysis of market information and trends are crucial to the adjustment of R&D directions and the development of new products. Traditional market analysis methods have problems such as information lag and incomplete data analysis, and are difficult to meet rapidly changing market demands; polymer materials have diverse properties, and different properties determine their performance in different applications. Traditional methods rely on expert experience and intuition. In some cases, R&D decisions may rely too much on the experience and intuition of individual experts, which may lead to deviations in the understanding of customer needs.

[0004] How to select materials that meet the requirements from a large amount of material data according to the specific needs of customers, and conduct property simulation and analysis is a major difficulty in the R&D process. In the existing technology, polymer material modification research mainly relies on expert experience, lacks systematic customer demand analysis and automated solution design, resulting in long R&D cycles and low efficiency. Summary of the invention

[0005] The present invention provides a method and system for exploring customer needs in the research and development of polymer material modification, aiming to solve the problem of long research and development cycle and low efficiency due to the lack of systematic customer demand analysis and automated solution design in the prior art in the research and development of polymer materials.

[0006] In order to solve the above technical problems, the present invention proposes a method for exploring customer needs in polymer material modification research and development, comprising the following steps:

[0007] Receive user needs, obtain the core elements of user needs, outline content, and viewpoints corresponding to the outline content;

[0008] Based on the generated opinion content, generate query keywords for the opinion content that needs external evidence support, use the query keywords to search and return external knowledge evidence;

[0009] Based on external knowledge evidence, modify and optimize the pre-generated opinion content, fill in or expand it into a complete answer as output, and generate a customer demand analysis report;

[0010] According to the customer demand analysis report, query material data from the material property database, perform material property simulation, and generate a material property analysis report;

[0011] Based on dynamic reports, customer demand analysis reports and material property analysis reports, use the large polymer material modification model as the base model to generate at least one modification scheme, use simulation software to evaluate the feasibility of the scheme, apply the prediction model to predict the performance and potential problems of different modification schemes, and generate a modification scheme report.

[0012] Preferably, the mining method also includes evaluating the generated modification scheme report, and the evaluation includes: evaluating the technical feasibility of the modification scheme based on experimental and simulation data; identifying potential compliance risks using a risk assessment model; and evaluating the economic feasibility of the modification scheme using a cost analysis tool.

[0013] Preferably, the risk assessment model is a pre-trained large language model that has been fine-tuned for compliance, and the training data includes mandatory standards, recommended standards, group standards and environmental regulations of different countries, organizations and institutions.

[0014] Preferably, the user demand is one or more of text, pictures, tables and physical objects, and the carrier can be one or more of electronic documents, chat records, pictures and physical objects. The physical samples are first subjected to laboratory tests and the test results are used as part of the user demand.

[0015] Preferably, the core elements include:

[0016] Material objects, including base materials and additives;

[0017] Limited scope or application scenarios, including set temperature conditions and drop resistance scenarios;

[0018] Performance requirements, including strength requirements and light aging resistance;

[0019] Other key requirements include cost requirements and supply requirements.

[0020] Preferably, the method for obtaining external knowledge evidence is to use the generated query keywords as a multi-path hybrid search, and obtain a preset number of entries from the search results as external knowledge evidence. The search paths include: full-network retrieval through a search engine API; paper search through a paper search API; patent search through a patent search API; and search of a self-built knowledge graph.

[0021] Preferably, the material property simulation is performed using computational chemistry software, and the simulation results are extracted and visualized.

[0022] Preferably, the prediction model includes: a machine learning model trained with a large amount of formula-performance data, a dedicated large model fine-tuned with a large amount of formula-performance data, a theoretical calculation method based on molecular dynamics, density functional theory, phase field simulation or finite element analysis, and a statistical mechanics method based on Monte Carlo simulation.

[0023] Preferably, the method for constructing the polymer material modified large model is:

[0024] Collect polymer material-related papers, polymer material-related patents, related professional books and forum discussion records, segment them into different modalities according to text, pictures and tables, further extract them into knowledge graphs with professional knowledge, and then use them as training data to supervise and fine-tune the open source big model to obtain a polymer material modification big model that specializes in polymer material modification.

[0025] Accordingly, another aspect of the present invention further provides a customer demand mining system in polymer material modification research and development, wherein the system executes the above-mentioned customer demand mining method, including:

[0026] Demand mining agent, the demand mining agent further includes an element extraction sub-agent, an outline extraction sub-agent, a content generation sub-agent, a query keyword generation sub-agent, a retrieval recall sub-agent, a verification feedback sub-agent and an optimization output sub-agent;

[0027] The element extraction sub-agent breaks down the core elements of the user's needs according to the set scenario requirements;

[0028] The outline extraction sub-agent supplements the outline content based on the questions in the user's needs, combined with the preset basic answer structure and extracted elements;

[0029] The content generation sub-agent generates the corresponding outline based on the questions in the user's needs, and outputs the viewpoint content corresponding to the outline;

[0030] The query keyword generation sub-agent generates query keywords for content or opinions that require external evidence support through pre-generated content;

[0031] The retrieval and recall sub-agent searches and returns relevant external knowledge evidence based on the query keywords provided by the query keyword generation sub-agent;

[0032] The verification feedback sub-intelligence checks the retrieval recall and pre-generated content and outputs the optimization direction;

[0033] The optimized output sub-agent modifies, optimizes, fills in or expands the pre-generated answer content based on external evidence to produce a complete answer as output;

[0034] Material Characterization Agent, which maintains a polymer material properties database, including existing material attributes and performance data, uses simulation and computational chemistry methods to predict material behavior, and provides material selection and formulation recommendations to meet specific customer needs;

[0035] The modification scheme design agent combines dynamic reports, customer demand analysis and material properties, uses the polymer material modification model as the base model to generate at least one modification scheme, and applies machine learning to predict the performance and potential problems of different modification schemes;

[0036] The scheme evaluation agent evaluates the feasibility and effectiveness of the modification scheme, integrates experimental simulation tools, risk assessment models and cost analysis tools, uses experimental simulation tools to evaluate the technical feasibility of the modification scheme, uses risk assessment models to identify potential compliance risks, and uses cost analysis tools to evaluate the economic feasibility of the scheme.

[0037] Compared with the prior art, the present invention has the following technical effects:

[0038] 1. The customer demand mining method proposed in the present invention utilizes artificial intelligence and big data technology, and generates at least one modification plan through a self-developed large model of polymer material modification, combined with dynamic reports, customer demand analysis and material properties, to realize the automation and intelligence of the R&D process, improve the R&D efficiency of polymer material modification R&D, and shorten the development cycle.

[0039] 2. The customer demand mining method proposed in the present invention identifies and avoids potential risks in advance through risk assessment and cost analysis, thereby ensuring the feasibility and economy of the R&D plan.

[0040] 3. The customer demand mining method proposed in the present invention is based on data-driven modification scheme design and evaluation, which improves the innovation capability and makes the developed materials more competitive in the market. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the process of the mining method of the present invention;

[0042] Figure 2It is a schematic diagram of the demand analysis report generation process described in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make the objectives, technical solutions and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in combination with specific embodiments of the present application and with reference to the accompanying drawings.

[0044] Embodiment 1

[0045] Methods for exploring customer needs in polymer material modification research and development, such as Figure 1 , 2 As shown, the following steps are included:

[0046] Receive user needs, obtain the core elements of user needs, outline content, and opinion content corresponding to the outline content.

[0047] The requirements submitted by users can be one or more of text, pictures, tables, and physical objects. The carrier can be one or more of electronic documents (such as Excel, word, pdf), chat records, pictures (screenshots, photos, scanned images, etc.), and can be delivered to enterprises through email, chat tools, physical mail, etc. Among them, physical samples need to be tested in the laboratory, and the test results will be used as part of the user's requirements. Other types can be left unprocessed or simply collected and sorted.

[0048] The different types of user requirement documents collected above are used as the input of the polymer material modification model, and its output is the following four elements:

[0049] 1. Determine the material object: including but not limited to base materials, additives, etc.;

[0050] 2. Determine the limited scope or application scenario: for example, under certain temperature conditions, drop-resistant scenarios, etc.

[0051] 3. Determine performance requirements: such as strength requirements, light aging resistance, etc.;

[0052] 4. Other key requirements: such as cost requirements, supply requirements, etc.

[0053] For the generation of the outline, this embodiment provides a set of demand mining outlines that are as comprehensive as possible in advance. The polymer material modification model uses the output results of the previous step as input to delete and modify the outline, retain valid chapters, delete irrelevant chapters, and add new chapters. Sometimes, the input will also include optimization suggestions for subsequent intelligent agents. The modified outline and user needs are used as input, and the polymer material modification model will generate a demand analysis report based on the outline.

[0054] Specifically, the input for outline generation includes extracted element information and predefined requirements mining outlines, and may also include optimization suggestions from subsequent agents (such as requirements verification feedback agents). The processing logic includes: outline matching, matching chapters of predefined outlines, associating core elements with chapter contents in the outline; using natural language processing (NLP) algorithms to automatically mark and delete outline chapters that are not related to user needs; content supplementation, combining specific scenarios or key requirements in user needs (such as new performance requirements or special application areas), automatically expanding or generating new chapters related to the needs; outline optimization, based on the optimization suggestions of subsequent step agents (such as requirements verification feedback agents), dynamically adjusting the outline structure to make it more logical and comprehensive for requirements analysis.

[0055] The input data for generating viewpoint content is the outline content and core elements. With user needs as the core, content is generated for each chapter of the optimized outline. For example: in the "Material Object" chapter, based on the user's needs description, clear material names and characteristics are listed; in the "Application Scenario" chapter, the description is supplemented in combination with the scenario requirements. Then, through the natural language generation capability of the big model, ensure that the content of each chapter is logically smooth and the context is consistent. Then use the self-built knowledge base or the knowledge capabilities of the big model to provide background support and industry references for the demand report. Finally, generate charts or data tables for some complex requirements (such as performance requirements and material modification indicators) to improve the readability and professionalism of the report.

[0056] Based on the generated opinion content, query keywords are generated for the opinion content that needs external evidence support, and the query keywords are used to search and return external knowledge evidence. The user demand analysis report generated in the previous step is used as input, and query keywords are generated for the content that needs external evidence support. Then use the generated query keywords as input to perform a multi-path hybrid search: perform a full network search through the search engine API; perform a paper search through the paper search API; perform a patent search through the patent search API; and search the self-built knowledge graph. The top 20 most relevant search results of the above different paths are taken as output.

[0057] The generation of query keywords first identifies the content that needs to be verified, and then scans the content in the demand analysis report to find the key points that need external evidence support (such as performance indicators, cost analysis, industry standards, etc.). Generate targeted query keywords based on each point that needs to be verified. For example: for "material performance at a specific temperature", generate query statements such as "specific material + specific temperature + performance test"; for "market demand trends", generate query statements such as "specific industry + trend analysis".

[0058] Based on external knowledge evidence, the pre-generated opinion content is modified, optimized, filled or expanded into a complete answer as output to generate a customer demand analysis report. The demand analysis report and search results generated by the polymer material modification model based on the outline are used as input to analyze whether the search results can support the demand analysis report. If the intelligence is insufficient, optimization suggestions are output and the suggestions are fed back to the outline mining step; if they can be well supported, they are passed to the next step. The next step is to optimize the content as a whole, fill in the blank content of the expression and opinion, and generate the final demand analysis report.

[0059] Compare the content to be verified in the demand analysis report with the search results. If the search results can directly support the content of the demand analysis report, mark it as "support". If the search results are not sufficient to support it, mark it as "need to be optimized" and record the specific problem points. Generate optimization suggestions for the content marked as "need to be optimized", including specific content to be modified or supplemented, new query keyword suggestions or analysis angles. Feedback the optimization suggestions to the outline content generation step, and adjust the outline to make up for the deficiencies.

[0060] According to the customer demand analysis report, query material data from the material property database, perform material property simulation, and generate a material property analysis report. The material property simulation is simulated using computational chemistry software, and the simulation results are extracted and visualized. In this embodiment, the task is completed by the material property analysis agent (Material Property Analysis Agent, MPA). The material property analysis agent maintains a polymer material property database, including existing material properties and performance data; uses simulation and computational chemistry methods to predict material behavior; and provides material selection and formulation recommendations to meet specific customer needs. It uses SQL or ElasticSearch to query the material database maintained by itself to obtain the property data of related materials. Use computational chemistry software (such as Gaussian, Materials Studio) and data visualization tools (such as Matplotlib, Plotly) to run computational chemistry software to simulate material properties, extract simulation results and visualize them. Finally, based on customer needs and material property data, recommend suitable polymer materials; and provide specific material formulation recommendations to optimize material performance.

[0061] Based on the dynamic report, customer demand analysis report and material property analysis report, use the polymer material modification large model as the base model to generate at least one modification scheme, use simulation software to evaluate the feasibility of the scheme, use the prediction model to predict the performance and potential problems of different modification schemes, and generate a modification scheme report. In this step, the implementer of the scheme is the modification design agent (MDA), which combines customer demand analysis and material properties, uses algorithms to generate modification schemes, uses simulation software to evaluate the feasibility of the scheme, and uses machine learning to predict the performance and potential problems of different modification schemes.

[0062] The method for constructing the macromolecular material modified model is as follows:

[0063] Collect polymer material-related papers, polymer material-related patents, related professional books and forum discussion records, segment them into different modalities according to text, pictures and tables, further extract them into knowledge graphs with professional knowledge, and then use them as training data to supervise and fine-tune the open source big model to obtain a polymer material modification big model that specializes in polymer material modification.

[0064] Specifically, the modification scheme design agent uses literature retrieval tools (such as Google Scholar, PubMed) and text parsing tools to retrieve and parse relevant literature based on dynamic reports, customer demand analysis reports and material property analysis reports, extract existing modification schemes and technical details, and compare and evaluate them. Based on the above data, the polymer material modification large model is used as the base model to generate at least one modification scheme, and the generated modification scheme is simulated using Gaussian, and the results are displayed using Matplotlib, experimental verification is arranged, and experimental data are recorded and analyzed to optimize the scheme.

[0065] In some embodiments of the present invention, after generating the modification plan report, the mining method further includes evaluating the generated modification plan report, wherein the evaluation includes: evaluating the technical feasibility of the modification plan based on experimental and simulation data; identifying potential compliance risks using a risk assessment model; and evaluating the economic feasibility of the modification plan using a cost analysis tool. This mainly involves a modification plan evaluation agent (MPEA), which integrates an experimental simulation tool, a risk assessment model, and a cost analysis tool, uses an experimental simulation tool to evaluate the technical feasibility of the modification plan, uses a risk assessment model to identify potential compliance risks, and uses a cost analysis tool to evaluate the economic feasibility of the plan.

[0066] In one embodiment of the present invention, the risk assessment model is a pre-trained large language model that has been fine-tuned for compliance, and the training data includes mandatory standards, recommended standards, group standards and environmental regulations of different countries, organizations and institutions.

[0067] In one embodiment of the present invention, the natural language processing tools include latent Dirichlet allocation, latent semantic analysis, neural topic model, topic evolution model, hierarchical Dirichlet process, BERTopic, WORD2vec, and Top2ve.

[0068] In one embodiment of the present invention, the prediction model includes: a machine learning model trained with a large amount of formula-performance data, a dedicated large model fine-tuned with a large amount of formula-performance data, a theoretical calculation method based on molecular dynamics, density functional theory, phase field simulation or finite element analysis, and a statistical mechanics method based on Monte Carlo simulation.

[0069] In this embodiment, a total of 2 million polymer material related papers, 2 million polymer material related patents, 1,000 related professional books, and 100,000 forum discussion records were collected, which were divided into different modes: text, pictures, tables, and further extracted into knowledge graphs in combination with professional knowledge. These data were used as training data to supervise and fine-tune the open source big model, thereby obtaining a big model specialized in polymer material modification. A part of the questions and answers collected from 0 forums was used as a test set, which did not participate in training. On this part of the data, it showed accuracy and understanding ability far exceeding the general big model, so it was determined that the big model was effective and usable in the field of polymer material modification.

[0070] Embodiment 2

[0071] This embodiment is a customer demand mining system in polymer material modification research and development. The system executes the customer demand mining method as described in the first embodiment, including:

[0072] Demand mining agent, the demand mining agent further includes an element extraction sub-agent, an outline extraction sub-agent, a content generation sub-agent, a query keyword generation sub-agent, a retrieval recall sub-agent, a verification feedback sub-agent and an optimization output sub-agent;

[0073] The element extraction sub-agent breaks down the core elements of the user's needs according to the set scenario requirements;

[0074] The outline extraction sub-agent supplements the outline content based on the questions in the user's needs, combined with the preset basic answer structure and extracted elements;

[0075] The content generation sub-agent generates the corresponding outline based on the questions in the user's needs, and outputs the viewpoint content corresponding to the outline;

[0076] The query keyword generation sub-agent generates query keywords for content or opinions that require external evidence support through pre-generated content;

[0077] The retrieval and recall sub-agent searches and returns relevant external knowledge evidence based on the query keywords provided by the query keyword generation sub-agent;

[0078] The verification feedback sub-intelligence checks the retrieval recall and pre-generated content and outputs the optimization direction;

[0079] The optimized output sub-agent modifies, optimizes, fills in or expands the pre-generated answer content based on external evidence to produce a complete answer as output;

[0080] Material Characterization Agent, which maintains a polymer material properties database, including existing material attributes and performance data, uses simulation and computational chemistry methods to predict material behavior, and provides material selection and formulation recommendations to meet specific customer needs;

[0081] The modification scheme design agent combines dynamic reports, customer demand analysis and material properties, uses the polymer material modification model as the base model to generate at least one modification scheme, and applies machine learning to predict the performance and potential problems of different modification schemes;

[0082] The scheme evaluation agent evaluates the feasibility and effectiveness of the modification scheme, integrates experimental simulation tools, risk assessment models and cost analysis tools, uses experimental simulation tools to evaluate the technical feasibility of the modification scheme, uses risk assessment models to identify potential compliance risks, and uses cost analysis tools to evaluate the economic feasibility of the scheme.

[0083] The above is only a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several modifications and improvements without departing from the inventive concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for exploring customer needs in the research and development of polymer material modification, characterized in that: The following steps are involved: Receive user needs, obtain the core elements of user needs, outline content, and viewpoints corresponding to the outline content; Based on the generated opinion content, generate query keywords for the opinion content that needs external evidence support, use the query keywords to search and return external knowledge evidence; Based on external knowledge evidence, modify and optimize the pre-generated opinion content, fill in or expand it into a complete answer as output, and generate a customer demand analysis report; According to the customer demand analysis report, query material data from the material property database, perform material property simulation, and generate a material property analysis report; Based on dynamic reports, customer demand analysis reports and material property analysis reports, use the large polymer material modification model as the base model to generate at least one modification scheme, use simulation software to evaluate the feasibility of the scheme, apply the prediction model to predict the performance and potential problems of different modification schemes, and generate a modification scheme report.

2. The method for mining customer needs in polymer material modification research and development according to claim 1, characterized in that: The mining method also includes evaluating the generated modification scheme report, and the evaluation includes: evaluating the technical feasibility of the modification scheme based on experimental and simulation data; identifying potential compliance risks using a risk assessment model; and evaluating the economic feasibility of the modification scheme using a cost analysis tool.

3. The method for mining customer needs in polymer material modification research and development according to claim 2, characterized in that: The risk assessment model is a pre-trained large language model that has been fine-tuned for compliance. The training data includes mandatory standards, recommended standards, group standards and environmental regulations of different countries, organizations and institutions.

4. The method for mining customer needs in polymer material modification research and development according to claim 1, characterized in that: The user demand is one or more of text, pictures, tables and physical objects. The carrier can be one or more of electronic documents, chat records, pictures and physical objects. The physical samples are first subjected to laboratory tests and the test results are used as part of the user demand.

5. The method for mining customer needs in polymer material modification research and development according to claim 1, characterized in that: The core elements include: Material objects, including base materials and additives; Limited scope or application scenarios, including set temperature conditions and drop resistance scenarios; Performance requirements, including strength requirements and light aging resistance; Other key requirements include cost requirements and supply requirements.

6. The method for mining customer needs in polymer material modification research and development according to claim 1, characterized in that: The method for obtaining external knowledge evidence is to use the generated query keywords as multi-path hybrid search, and obtain a preset number of entries from the search results as external knowledge evidence. The search paths include: full network retrieval through the search engine API; paper search through the paper search API; patent search through the patent search API; and search of the self-built knowledge graph.

7. The method for mining customer needs in polymer material modification research and development according to claim 1, characterized in that: The material property simulation is performed using computational chemistry software, and the simulation results are extracted and visualized.

8. The method for mining customer needs in polymer material modification research and development according to claim 1, characterized in that: The prediction models include: machine learning models trained with a large amount of formula-performance data, dedicated large models fine-tuned with a large amount of formula-performance data, theoretical calculation methods based on molecular dynamics, density functional theory, phase field simulation or finite element analysis, and statistical mechanics methods based on Monte Carlo simulation.

9. The method for mining customer needs in polymer material modification research and development according to claim 1, characterized in that: The method for constructing the macromolecular material modified model is as follows: Collect polymer material-related papers, polymer material-related patents, related professional books and forum discussion records, segment them into different modalities according to text, pictures and tables, further extract them into knowledge graphs with professional knowledge, and then use them as training data to supervise and fine-tune the open source big model to obtain a polymer material modification big model that specializes in polymer material modification.

10. The customer demand mining system in polymer material modification research and development is characterized by: The system executes the customer demand mining method according to any one of claims 1 to 9, comprising: Demand mining agent, the demand mining agent further includes an element extraction sub-agent, an outline extraction sub-agent, a content generation sub-agent, a query keyword generation sub-agent, a retrieval recall sub-agent, a verification feedback sub-agent and an optimization output sub-agent; The element extraction sub-agent breaks down the core elements of the user's needs according to the set scenario requirements; The outline extraction sub-agent supplements the outline content based on the questions in the user's needs, combined with the preset basic answer structure and extracted elements; The content generation sub-agent generates the corresponding outline based on the questions in the user's needs, and outputs the viewpoint content corresponding to the outline; The query keyword generation sub-agent generates query keywords for content or opinions that require external evidence support through pre-generated content; The retrieval and recall sub-agent searches and returns relevant external knowledge evidence based on the query keywords provided by the query keyword generation sub-agent; The verification feedback sub-intelligence checks the retrieval recall and pre-generated content and outputs the optimization direction; The optimized output sub-agent modifies, optimizes, fills in or expands the pre-generated answer content based on external evidence to produce a complete answer as output; Material Characterization Agent, which maintains a polymer material properties database, including existing material attributes and performance data, uses simulation and computational chemistry methods to predict material behavior, and provides material selection and formulation recommendations to meet specific customer needs; The modification scheme design agent combines dynamic reports, customer demand analysis and material properties, uses the polymer material modification model as the base model to generate at least one modification scheme, and applies machine learning to predict the performance and potential problems of different modification schemes; The scheme evaluation agent evaluates the feasibility and effectiveness of the modification scheme, integrates experimental simulation tools, risk assessment models and cost analysis tools, uses experimental simulation tools to evaluate the technical feasibility of the modification scheme, uses risk assessment models to identify potential compliance risks, and uses cost analysis tools to evaluate the economic feasibility of the scheme.

Images