Electromechanical product innovative design problem analysis method and system based on pre-training model

Abstract

The invention provides an electromechanical product innovative design problem analysis method and system based on a pre-training model. The method and system are used for solving the problem that the accuracy and efficiency of product problem analysis are low. The method comprises the following steps: acquiring an innovative design problem description text and a technique analysis tool type for a target mechanical and electrical product; searching a structured prompt template corresponding to the technical analysis tool type from a prompt template library, wherein a corresponding relationship between the technical analysis tool type and the structured prompt template is stored in the prompt template library; filling the innovative design problem description text into a structured prompt template to obtain a technique analysis prompt engineering text; reasoning the technical analysis prompt engineering text through the pre-training model to obtain a technical analysis result of the target mechanical and electrical product, the technical analysis prompt engineering text being used for guiding a thinking chain prompt reasoning process of the pre-training model.

Description

Technical Field

[0001] This application relates to the technical fields of artificial intelligence, natural language processing, industrial design and engineering problem analysis, and more specifically, to a method and system for analyzing innovative design problems of electromechanical products based on a pre-trained model. Background Technology

[0002] In the current product design field, traditional innovative analytical tools (such as causal analysis, fishbone analysis, fault tree analysis, and requirement identification) are widely used to solve engineering problems. These tools typically rely on designers' systematic logical reasoning, experiential judgment, and creative thinking, as the designer's knowledge and experience directly determine the depth and breadth of their analysis of complex engineering problems. This reliance on individual abilities not only reduces analytical efficiency but also makes it prone to deviations from optimal solutions due to subjective judgment biases, becoming a key bottleneck restricting the intelligent design of electromechanical products. Therefore, the accuracy and efficiency of current product problem analysis are relatively low. Summary of the Invention

[0003] The purpose of this application is to provide a method and system for analyzing innovative design problems of electromechanical products based on a pre-trained model, in order to improve the low accuracy and efficiency of product problem analysis.

[0004] This application provides a method for analyzing innovative design problems in electromechanical products based on a pre-trained model. The method includes: acquiring an innovative design problem description text and a type of technical analysis tool for the target electromechanical product; searching a structured prompt template corresponding to the technical analysis tool type from a prompt template library, which stores the correspondence between technical analysis tool types and structured prompt templates; filling the innovative design problem description text into the structured prompt template to obtain technical analysis prompt engineering text; and using a pre-trained model to infer the technical analysis prompt engineering text to obtain the technical analysis results for the target electromechanical product. The technical analysis prompt engineering text guides the thought process of the pre-trained model. In the implementation of the above solution, by combining the structured prompt template with the strong generalization capability of the pre-trained model, the structured template precisely binds the problem description with the logical framework of technical analysis tools (such as causal analysis and fault tree analysis), forcing the model to generate results according to a preset analysis path. This avoids deviations caused by ambiguous input information or insufficient designer experience, thereby improving the accuracy of product problem analysis. Furthermore, because pre-trained models possess cross-domain knowledge integration capabilities, they can quickly connect implicit factors in problem descriptions and automatically generate analysis results that conform to industry standards, thereby improving the efficiency of product problem analysis. Therefore, this solution simultaneously reduces the model's reliance on subjective instructions through template standardization and, combined with the strong generalization ability of pre-trained models, ensures analytical depth while significantly shortening response time, ultimately improving both the accuracy and efficiency of product problem analysis.

[0005] Optionally, in this embodiment, before reasoning about the technique analysis prompts in the engineering text using the pre-trained model, the method further includes: obtaining a fine-tuning training dataset, which includes problem description samples and analysis result samples; and using the problem description samples and analysis result samples to fine-tune the pre-trained model. In the implementation of the above scheme, by introducing fine-tuning training before the pre-trained model performs reasoning, the model learns the association patterns between the problem description samples and the analysis result samples, thereby enhancing its semantic understanding of the input text and its ability to generate structured outputs. This allows for more accurate capture of the core features of the problem and output of high-quality results that conform to the logic of the technique analysis tool during subsequent reasoning. Simultaneously, it reduces reasoning bias caused by differences in domain terminology or insufficient task constraints, significantly improving the model's adaptability to specific domains and its task matching ability.

[0006] Optionally, in this embodiment, obtaining the fine-tuning training dataset includes: obtaining a problem description sample and a structured prompt template corresponding to the problem description sample; filling the problem description sample into the structured prompt template corresponding to the problem description sample to obtain problem analysis prompt engineering text; inputting the problem analysis prompt engineering text into a general language model so that the general language model can reason about the problem description sample to obtain the analysis result text corresponding to the problem description sample, wherein the number of model parameters of the general language model is greater than the number of model parameters of the pre-trained model; correcting and annotating the analysis result text corresponding to the problem description sample to obtain analysis result samples; and adding the problem description sample and the analysis result sample to the fine-tuning training dataset. In the implementation of the above scheme, by combining the problem description sample with the structured prompt template and using the general language model with a larger number of parameters to generate preliminary analysis results, and then forming high-quality analysis result samples through manual correction and annotation, a fine-tuning training dataset that is highly matched with the domain task can be constructed. On the one hand, the structured prompt template constrains the reasoning logic of the general language model, making its output more in line with the specification requirements of the technique analysis tool; on the other hand, the correction and annotation step further optimizes the accuracy and completeness of the analysis results, thereby ensuring the quality and consistency of the fine-tuning training dataset. Therefore, fine-tuning the pre-trained model using this dataset can significantly improve its semantic understanding, output structure adaptability, and result reliability in the analysis of specific domain problems.

[0007] Optionally, in this embodiment, before fine-tuning the pre-trained model using problem description samples and analysis result samples, the method further includes: obtaining a target language model and inserting a preset matrix into the target language model to obtain a pre-trained model; fine-tuning the pre-trained model using problem description samples and analysis result samples includes: freezing the model parameters of the target language model in the pre-trained model, using problem description samples as training data and analysis result samples as training labels, and updating the preset matrix in the pre-trained model. In the implementation of the above scheme, fine-tuning the model by inserting a preset matrix into the target language model and updating only that matrix can achieve efficient adaptation and optimization of the pre-trained model without changing the original parameters of the target language model. On the one hand, freezing the parameters of the target language model can significantly reduce computational resource consumption and storage costs while retaining its original knowledge expression capabilities. On the other hand, by selectively updating the preset matrix, the model can quickly learn task-related feature mapping relationships, thereby improving the model's adaptability and output accuracy for specific domain problems with limited training data, while avoiding the risk of overfitting due to full parameter training, achieving a balance between model performance and training efficiency.

[0008] Optionally, in this embodiment, reasoning on the technical analysis prompt engineering text using a pre-trained model includes: obtaining the model calling program of the pre-trained model; inputting the technical analysis prompt engineering text into the pre-trained model through the model calling program, so that the pre-trained model can perform reasoning analysis on the technical analysis prompt engineering text and output the technical analysis results of the target electromechanical product. In the implementation of the above scheme, by introducing the model calling program as an interactive bridge between the pre-trained model and the technical analysis prompt engineering text, automated reasoning and structured output of the target electromechanical product problem can be achieved. This allows the filled prompt engineering text to be efficiently input into the pre-trained model, ensuring that the model strictly follows the logical framework of the technical analysis tool during the reasoning process. At the same time, through the semantic parsing and generation capabilities of the model, unstructured problem descriptions are transformed into analysis results that conform to engineering specifications, thereby effectively improving the execution efficiency and stability of the reasoning task.

[0009] Optionally, in this embodiment, the types of technical analysis tools include: requirement identification tools, requirement transformation tools, causal analysis tools, fault analysis tools, component analysis tools, fishbone analysis tools, product verification tools, technology prediction tools, ideal system tools, creative template tools, target mining tools, and / or resource analysis tools. In the implementation of the above solution, by integrating multiple types of technical analysis tools, a multi-dimensional and systematic analysis and solution capability for target electromechanical product problems can be achieved. Because different tool types provide differentiated methodological support for problem diagnosis, requirement transformation, innovative design, and resource optimization, the system can dynamically match the analysis needs of complex problems. Simultaneously, the logical correlation and structured output requirements between the tools ensure that the analysis process follows engineering specifications and generates operable results, thereby improving the comprehensiveness, accuracy, and feasibility of the problem analysis and solutions.

[0010] This application embodiment also provides a system for analyzing innovative design problems of electromechanical products based on a pre-trained model, including: a terminal device and a server; the terminal device is used to send an innovative design problem description text and a type of technique analysis tool for the target electromechanical product to the server; the server is used to receive the innovative design problem description text and the type of technique analysis tool for the target electromechanical product sent by the terminal device; search for a structured prompt template corresponding to the type of technique analysis tool from a prompt template library, the prompt template library storing the correspondence between the type of technique analysis tool and the structured prompt template; fill the innovative design problem description text into the structured prompt template to obtain the technique analysis prompt engineering text; and use the pre-trained model to reason about the technique analysis prompt engineering text to obtain the technique analysis result of the target electromechanical product, the technique analysis prompt engineering text being used to guide the thought chain prompt reasoning process of the pre-trained model.

[0011] This application also provides a device for analyzing innovative design problems of electromechanical products based on a pre-trained model, comprising: a tool type acquisition module for acquiring the description text of the innovative design problem and the type of technique analysis tool for the target electromechanical product; a prompt template search module for searching the structured prompt template corresponding to the type of technique analysis tool from a prompt template library, wherein the prompt template library stores the correspondence between the type of technique analysis tool and the structured prompt template; an engineering text acquisition module for filling the description text of the innovative design problem into the structured prompt template to obtain the technique analysis prompt engineering text; and an analysis result acquisition module for reasoning on the technique analysis prompt engineering text through the pre-trained model to obtain the technique analysis result of the target electromechanical product, wherein the technique analysis prompt engineering text is used to guide the thought chain prompt reasoning process of the pre-trained model.

[0012] Optionally, in this embodiment of the application, the electromechanical product innovation design problem analysis device based on the pre-trained model further includes: a training dataset acquisition module, used to acquire a fine-tuning training dataset, the fine-tuning training dataset including: problem description samples and analysis result samples; and a model problem training module, used to fine-tune the pre-trained model using the problem description samples and analysis result samples.

[0013] Optionally, in this embodiment, the training dataset acquisition module includes: a sample data acquisition submodule, used to acquire a problem description sample and a structured prompt template corresponding to the problem description sample; a prompt template filling submodule, used to fill the problem description sample into the structured prompt template corresponding to the problem description sample to obtain problem analysis prompt engineering text; a language model inference submodule, used to input the problem analysis prompt engineering text into a general language model so that the general language model can infer the problem description sample to obtain the analysis result text corresponding to the problem description sample, wherein the number of model parameters of the general language model is greater than the number of model parameters of the pre-trained model; a text correction and annotation submodule, used to correct and annotate the analysis result text corresponding to the problem description sample to obtain analysis result sample; and a sample data addition submodule, used to add the problem description sample and the analysis result sample to the fine-tuning training dataset.

[0014] Optionally, in this embodiment of the application, the electromechanical product innovation design problem analysis device based on the pre-trained model further includes: a preset matrix insertion module, used to obtain a target language model and insert a preset matrix into the target language model to obtain a pre-trained model; and a model problem training module, including: a preset matrix update module, used to update the preset matrix in the pre-trained model with problem description samples as training data and analysis result samples as training labels, while freezing the model parameters of the target language model in the pre-trained model.

[0015] Optionally, in this embodiment of the application, the analysis result acquisition module includes: a calling program acquisition submodule, used to acquire the model calling program of the pre-trained model; and a text reasoning analysis submodule, used to input the technique analysis prompt engineering text into the pre-trained model through the model calling program, so that the pre-trained model can perform reasoning analysis on the technique analysis prompt engineering text and output the technique analysis result of the target electromechanical product.

[0016] Optionally, in the embodiments of this application, the types of technical analysis tools include: demand identification tools, demand transformation tools, causal analysis tools, fault analysis tools, component analysis tools, fishbone analysis tools, product inspection tools, technology forecasting tools, ideal system tools, creative template tools, target mining tools, and / or resource analysis tools.

[0017] This application also provides an electronic device, including a processor and a memory, wherein the memory stores machine-readable instructions executable by the processor, and the machine-readable instructions are executed by the processor to perform the methods described above.

[0018] This application also provides a computer-readable storage medium storing a computer program, which is executed by a processor to perform the methods described above.

[0019] This application also provides a computer program product, including: a computer program or computer instructions, which are executed by a processor to perform the method described above. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 The diagram shown is a flowchart illustrating the method for analyzing innovative design problems of electromechanical products based on a pre-trained model, as provided in an embodiment of this application. Figure 2 The diagram shown is a schematic of an electromechanical product innovative design problem analysis system based on a pre-trained model provided in an embodiment of this application; Figure 3 The diagram shown is a structural schematic of the electromechanical product innovative design problem analysis device based on a pre-trained model provided in an embodiment of this application. Figure 4 The diagram shows a structural schematic of an electronic device provided in an embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in the embodiments of this application are for illustrative and descriptive purposes only and are not intended to limit the protection scope of the embodiments of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in the embodiments of this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of the embodiments of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0023] Furthermore, the described embodiments are merely a part of the embodiments of this application, and not all of them. The components of the embodiments of this application described and illustrated herein can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed embodiments of this application, but merely to illustrate selected embodiments of this application.

[0024] It is understood that the terms "first" and "second" in the embodiments of this application are used to distinguish similar objects. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that "first" and "second" do not necessarily imply that they are different. In the description of the embodiments of this application, the term "and / or" is merely a description of the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship. The term "multiple" refers to two or more (including two), and similarly, "multiple groups" refers to two or more groups (including two groups).

[0025] It should be noted that the electromechanical product innovation design problem analysis method based on pre-trained models provided in this application can be executed by electronic devices. Here, electronic devices refer to device terminals or servers with the function of executing computer programs. Device terminals include, for example, smartphones, personal computers, tablets, personal digital assistants, or mobile internet devices. Servers refer to devices that provide computing services through a network. Servers include, for example, x86 servers and non-x86 servers. Non-x86 servers include, for example, mainframes, minicomputers, and UNIX servers.

[0026] In the current field of electromechanical product design, traditional innovative analytical tools (such as causal analysis, fishbone analysis, fault tree analysis, and requirement identification) are widely used to solve engineering problems. These tools typically rely on designers' systematic logical reasoning, experiential judgment, and creative thinking, as the designer's knowledge and experience directly determine the depth and breadth of their analysis of complex engineering problems. For example, causal analysis requires designers to accurately identify the root cause chain of a problem. A lack of in-depth understanding of the multidisciplinary coupling mechanisms of electromechanical systems (such as mechanical-electronic-control interaction) can lead to misjudgments of causal relationships or one-sided analysis of key factors. This designer-dependent analytical approach, heavily reliant on individual capabilities, not only reduces analytical efficiency (due to the need for repeated hypothesis verification) but also easily leads to deviations from the optimal solution due to subjective judgment biases, becoming a key bottleneck restricting the intelligent design of electromechanical products. Therefore, the accuracy and efficiency of current product problem analysis are relatively low.

[0027] For the above questions, please refer to [link / reference]. Figure 1 The illustrated flowchart represents a method for analyzing innovative design problems in electromechanical products based on a pre-trained model, as provided in this application embodiment. The main idea of ​​this method is to combine structured prompt templates with the strong generalization ability of the pre-trained model. The standardization of the structured prompt templates reduces the model's reliance on subjective instructions, while the strong generalization ability of the pre-trained model ensures analytical depth while significantly shortening response time, ultimately improving both the accuracy and efficiency of product problem analysis. The implementation methods of the aforementioned method for analyzing innovative design problems in electromechanical products based on a pre-trained model may include: Step S110: Obtain the description text of the innovative design problem for the target electromechanical product and the type of technique analysis tool.

[0028] The target electromechanical product refers to the electromechanical product for which problem analysis is required. In specific practice, it can also be applied to other types of products, such as network products or communication equipment.

[0029] Innovative design problem description text is natural language text that describes specific problems or defects of target electromechanical products. Taking the problem description of electromechanical products as an example, "the motor overheats and shuts down when running in a high-temperature environment" or "the hydraulic system has insufficient pressure" are examples of texts that directly describe the product's problem performance.

[0030] Technique analysis tools are standardized categories of methods used for engineering problem analysis, such as requirement identification tools, requirement transformation tools, cause-and-effect analysis tools, failure analysis tools, component analysis tools, fishbone analysis tools, product verification tools, technology forecasting tools, ideal system tools, creative template tools, target mining tools, and / or resource analysis tools. Each type corresponds to a specific logical framework and output format.

[0031] Step S120: Find the structured prompt template corresponding to the technique analysis tool type from the prompt template library. The prompt template library stores the correspondence between technique analysis tool types and structured prompt templates.

[0032] The hint template library is a database that stores the correspondence between the types of technique analysis tools and structured hint templates. For example, the template library stores the correspondence between "cause and effect analysis → template A" and "fault tree analysis → template B". Template A can include problem description, possible cause classification, solution fields, etc.

[0033] Structured prompt templates are pre-defined text frameworks designed for specific techniques and analysis tools. They guide the model to generate logical analysis results through field placeholders (such as {problem description}, {cause classification}).

[0034] Step S130: Fill the innovative design problem description text into the structured prompt template to obtain the technique analysis prompt engineering text.

[0035] The technical analysis prompt engineering text is a complete input text generated by filling the innovative design problem description text into the structured prompt template, which serves as the input content for the pre-trained model's inference.

[0036] Step S140: The pre-trained model is used to reason about the technical analysis prompting engineering text to obtain the technical analysis results of the target electromechanical product. The technical analysis prompting engineering text is used to guide the thought chain prompting reasoning process of the pre-trained model.

[0037] Pre-trained models are large language models (LLMs) trained on large-scale data. They possess the ability to understand natural language and generate structured content, and are used to infer from prompt text and output analysis results. The aforementioned pre-trained models can employ Deepseek series models (such as Deepseek R1-7B), Qwen series models, Llama series models (such as Llama 3-8B), etc. These models can all process complex text and generate analysis results that conform to template logic.

[0038] The results of the technique analysis are obtained by using the structured logic in the structured prompt template corresponding to the technique analysis tool type to analyze and reason about the description text of the innovative design problem, resulting in a structured analysis logic.

[0039] In implementing the above solution, a structured prompt template is combined with the strong generalization capability of the pre-trained model. This structured template precisely binds the problem description to the logical framework of analytical techniques (such as causal analysis and fault tree analysis), forcing the model to generate results according to a preset analysis path. This avoids deviations caused by ambiguous input information or insufficient designer experience, thereby improving the accuracy of product problem analysis. Furthermore, because the pre-trained model possesses cross-domain knowledge integration capabilities, it can quickly associate implicit factors in the problem description and automatically generate analysis results that conform to industry standards, thus improving the efficiency of product problem analysis. Therefore, this solution simultaneously reduces the model's dependence on subjective instructions through template standardization and ensures analytical depth by combining the strong generalization capability of the pre-trained model, while also significantly shortening response time, ultimately improving both the accuracy and efficiency of product problem analysis.

[0040] As an optional implementation of step S110 above, for example: a user can input a description of the innovative design problem of the target electromechanical product (e.g., "the motor experiences overheating and shutdown when operating in a high-temperature environment") through the human-computer interaction interface or application programming interface (API) of the terminal device, and select the corresponding technique analysis tool type (e.g., the technique analysis tool type could be "cause-effect analysis tool") on the terminal device. Then, the user submits the description of the innovative design problem and the technique analysis tool type of the target electromechanical product to an electronic device (e.g., a server). The electronic device (e.g., a server) can receive the description of the innovative design problem and the technique analysis tool type of the target electromechanical product from the terminal device, or it can obtain the description of the innovative design problem and the technique analysis tool type of the target electromechanical product from the terminal device through the API interface.

[0041] As an optional implementation of step S120 above, for example, the prompt template library can pre-store the correspondence between various technique analysis tool types and structured prompt templates. If the technique analysis tool type selected by the user is "causal analysis tool", then the structured prompt template corresponding to "causal analysis tool" can be searched and retrieved from the prompt template library. The structured prompt template corresponding to the causal analysis tool can be expressed as: "You are a causal analysis expert in the electromechanical field. Please provide your answer strictly according to the following format: Question: {question}; 1. Possible causes: 2. In-depth analysis: 3. Suggested solutions: Please add the corresponding content after each section, and do not omit any headings. Analysis:".

[0042] As an optional implementation of step S130 above, for example: assuming the innovative design problem description text received by the electronic device (such as a server) is "the motor experiences overheating and shutdown when operating in a high-temperature environment", and the technique analysis tool type is "cause-and-effect analysis tool", then the structured prompt template corresponding to the above cause-and-effect analysis tool can be expressed as: "You are a cause-and-effect analysis expert in the field of electromechanical engineering. Please provide your answer strictly according to the following format: Problem: The motor experiences overheating and shutdown when operating in a high-temperature environment; 1. Possible causes: 2. In-depth analysis: 3. Suggested solutions: Please add the corresponding content after each subsection, and do not omit any headings. Analysis:".

[0043] As an optional implementation of the above-mentioned method for analyzing innovative design problems of electromechanical products based on pre-trained models, before reasoning about the technical analysis prompts in the engineering text through the pre-trained model, it may further include: Step S141: Obtain the fine-tuning training dataset, which includes: problem description samples and analysis result samples.

[0044] It is understood that the implementation of step S141 is relatively complex, and therefore the implementation of step S141 will be described in detail below.

[0045] Step S142: Fine-tune the pre-trained model using the problem description samples and analysis result samples.

[0046] For example, the above step S142 can be implemented by using problem description samples and analysis result samples to fine-tune pre-trained models such as Deepseek series models (e.g., DeepseekR1-7B), Qwen series models (e.g., Qwen-Max), and Llama series models (e.g., Llama3-8B) to obtain a pre-trained model for the vertical field of innovative design of target electromechanical products. This pre-trained model can provide professional technical analysis results for problems in the vertical field of innovative design of target electromechanical products.

[0047] As an optional implementation of step S141 above, the implementation of obtaining the fine-tuning training dataset may include: Step S141a: Obtain a problem description sample and the corresponding structured prompt template.

[0048] For example, in the implementation of step S141a above, in the scenario of electromechanical products, it is first necessary to extract typical problem description samples from historical maintenance records, user feedback, or test reports. The problem description samples vary depending on the specific data. Here, the problem description samples may include: "Abnormal vibration occurs when the industrial motor is running under load," "High-frequency noise is generated when the gearbox is running at high speed," etc. After obtaining the problem description samples, if the user selects "Causal Analysis Tool" as the technical analysis tool type corresponding to the innovative design problem description text on the terminal device, then the structured prompt template corresponding to the problem description sample can be the structured prompt template corresponding to the causal analysis tool. The structured prompt template can be expressed as: "You are a causal analysis expert in the electromechanical field. Please answer strictly according to the following format: Question: {question}; 1. Possible causes: 2. In-depth analysis: 3. Suggested solutions: Please add the corresponding content after each subsection, and do not omit any headings. Analysis:."

[0049] Step S141b: Fill the problem description sample into the structured prompt template corresponding to the problem description sample to obtain the problem analysis prompt engineering text.

[0050] For example, the implementation of step S141b above is as follows: Suppose the problem description sample is "An industrial motor experiences abnormal vibration when running under load," and the structured prompt template corresponding to this problem description sample is "You are an expert in cause-and-effect analysis in the field of electromechanical engineering. Please provide an answer strictly according to the following format: Question: {question}; 1. Possible causes: 2. In-depth analysis: 3. Suggested solutions: Please add the corresponding content after each subsection, and do not omit any headings. Analysis:", then fill the problem description sample into the structured prompt template corresponding to this problem description sample, and obtain the problem analysis prompt engineering text as "You are an expert in cause-and-effect analysis in the field of electromechanical engineering. Please provide an answer strictly according to the following format: Question: An industrial motor experiences abnormal vibration when running under load; 1. Possible causes: 2. In-depth analysis: 3. Suggested solutions: Please add the corresponding content after each subsection, and do not omit any headings. Analysis:".

[0051] Step S141c: Input the problem analysis prompt engineering text into the general language model so that the general language model can reason about the problem description sample and obtain the analysis result text corresponding to the problem description sample. The number of model parameters of the general language model is greater than the number of model parameters of the pre-trained model.

[0052] A general language model is a language model with more parameters than a pre-trained model and trained using a large general dataset.

[0053] For example, the implementation of step S141c above involves inputting the problem analysis prompt engineering text into a general language model such as Qwen-Max, so that the general language model can reason about the problem description sample and obtain the analysis result text corresponding to the problem description sample.

[0054] Step S141d: Correct and annotate the analysis result text corresponding to the problem description sample to obtain the analysis result sample.

[0055] An example implementation of step S141d above is as follows: During the process of correcting and annotating the analysis result text corresponding to the problem description sample, a thought chain hint method can be used to manually correct the analysis result text, obtaining a corrected result text that reflects the thought chain hint structure. Then, the corrected result text is annotated to mark each part of the analysis result sample, such as the thought chain hint part and the causal chain analysis part. Here, the thought chain hint is a technique that guides a large language model to gradually demonstrate the reasoning process, aiming to improve its ability to solve complex problems. The core idea of ​​the thought chain hint technique is to imitate the human "step-by-step thinking" mode, breaking down complex tasks into multiple intermediate steps, thereby enhancing the model's logic and accuracy.

[0056] Step S141e: Add the problem description sample and the analysis result sample to the fine-tuning training dataset.

[0057] For example, the implementation of step S141e above is as follows: After adding the problem description sample and the analysis result sample to the fine-tuning training dataset, a fine-tuning training dataset containing partial sample data can be obtained. The partial sample data in the fine-tuning training dataset can be represented using JavaScript Object Notation (JSON) format as follows: [{"input":"Insufficient hydraulic system pressure","output":"Causal chain analysis:\n1. Direct cause:\n-Reduced hydraulic pump output flow\n-Relief valve set pressure too low;\n2. Deep-seated cause:\n-Pump body wear leading to decreased volumetric efficiency (material fatigue)\n-Valve core jamming leading to pressure regulation failure (contaminant intrusion)\n3. Root cause:\n-Failure to perform regular maintenance (management system defects)"} \n-Improper filtration system selection (design flaw)\nSuggested solutions:\n1. Replace the hydraulic pump rotor assembly\n2. Clean or replace the overflow valve\n3. Add an oil contamination detection device"},{"input":"Machine arm positioning accuracy deviation","output":"Cause chain analysis:\n1. Direct cause:\n-Servo motor encoder signal interference\n-Excessive backlash in the reducer\n2. Deep-seated cause:\n-Shielded cable not properly grounded (installation process problem)\n-Harmonic reducer wear exceeds limits (poor lubrication)\n3. Root cause:\n-Electromagnetic compatibility design does not meet requirements (lax standard implementation)\n-Unreasonable preventive maintenance cycle setting\nSuggested solutions:\n1. Install a signal isolator\n2. Replace the reducer and improve the lubrication scheme\n3. Revise the equipment inspection standards"}).

[0058] As an optional implementation of the above-mentioned method for analyzing innovative design problems of electromechanical products based on pre-trained models, before fine-tuning the pre-trained model using problem description samples and analysis result samples, the method further includes: Step S142a: Obtain the target language model and insert a preset matrix into the target language model to obtain a pre-trained model.

[0059] An example implementation of step S142a above is as follows: It is understood that Low-Rank Adaptation (LoRA) technology can be used to fine-tune the pre-trained model. Specifically, a target language model, such as Deepseek R1-7B, can be obtained first. Then, a preset matrix (e.g., a pre-set low-rank matrix) can be inserted into the preset neural network layers of the target language model (e.g., Deepseek R1-7B) to obtain the pre-trained model. This Low-Rank Adaptation (LoRA) technology enables the model to quickly complete the training process of the dataset in the innovative design vertical domain when performing tasks, while avoiding retraining all parameters of the target language model (e.g., Deepseek R1-7B), thus effectively saving computational and memory resources.

[0060] As an optional implementation of step S142 above, the implementation of fine-tuning the pre-trained model using problem description samples and analysis result samples may include: Step S142b: With the model parameters of the target language model in the pre-trained model frozen, update the preset matrix in the pre-trained model using the problem description samples as training data and the analysis result samples as training labels.

[0061] For example, the implementation of step S142b above involves freezing the model parameters of the target language model (such as DeepseekR1-7B) in the pre-trained model, using problem description samples as training data and analysis result samples as training labels, and updating the preset matrix (such as a low-rank matrix) in the pre-trained model. This allows the pre-trained model to retain the general language understanding ability of DeepseekR1-7B while also improving its understanding of the vertical field of innovative design of the target electromechanical products.

[0062] As an optional implementation of step S140 above, the implementation of reasoning about the technique analysis prompt engineering text using a pre-trained model may include: Step S143: Obtain the model calling program for the pre-trained model.

[0063] Step S144: Input the technical analysis prompt engineering text into the pre-trained model through the model calling program, so that the pre-trained model can perform inference analysis on the technical analysis prompt engineering text and output the technical analysis results of the target electromechanical product.

[0064] For example, the implementation of steps S143 to S144 above can be as follows: Suppose the user inputs an innovative design problem description text for the target electromechanical product as "What are the faults of train floating slab vibration isolators?", then steps S110 to S130 above can be used to process this problem description text to obtain technical analysis prompt engineering text. Then, after obtaining the model calling program of the pre-trained model, the server can input the technical analysis prompt engineering text into the pre-trained model through the model calling program, so that the pre-trained model can perform inference analysis on the technical analysis prompt engineering text and output the technical analysis results of the target electromechanical product. The technical analysis results of the above-mentioned target electromechanical products can be summarized as follows: 1. Top Event: - Failure of the train's floating slab vibration isolator. 2. Intermediate Events: - Abnormal vibration: This may damage the floating slab vibration isolator, leading to the top event. - Damper failure: As a key component of the vibration isolation device, the failure of the damper will reduce the vibration isolation effect, increase the system vibration frequency, and may further lead to the top event. - Sensor malfunction: The sensor cannot properly monitor the vibration isolation status, causing the system to be unable to adjust in time, increasing system vibration and increasing the risk of the top event. - Control system failure: The control system cannot effectively adjust the vibration isolation device, which may lead to unstable vibration isolation parameters, affecting the vibration isolation effect and thus leading to the top event. 3. Basic Events: - The cause of abnormal vibration may be mechanical wear, temperature changes, external impact, or operational errors, leading to abnormal operation of the vibration isolator. These factors may cause the operating temperature of the damper to rise, thereby accelerating its failure; they may also cause abnormal sensor readings. When the threshold is exceeded, the protection mechanism is triggered, but the feedback from the control system fails to process it in time, resulting in the vibration isolation device being unable to effectively monitor the vibration. Responding to vibrations, which in turn triggers even greater vibrations transmitted to the vibration isolator, exacerbating its damage. - Damper failure is caused by factors such as material aging, installation misalignment, or fatigue damage due to overuse. Material aging reduces the damping coefficient and weakens the damping capacity, leading to increased system vibration. Installation misalignment causes inaccurate positioning of the vibration isolator, affecting its isolation effect and also increasing system vibration. Overuse causes severe wear on the internal structure of the damper, breakage of elastic elements, and ultimately, vibration isolator failure. - Sensor malfunctions include insufficient power supply, short circuits, environmental interference, and interference from other electronic devices. If the sensor itself malfunctions when detecting the vibration isolation status, it will not be able to provide accurate data to the control system. This may cause the control system to make adjustment decisions based on incorrect information, or fail to detect potential problems in time, thus leaving the vibration isolation device in an inactive state, increasing system vibration, and indirectly causing damage to the vibration isolator. - Control system failures may involve software malfunctions, hardware damage, or communication interruptions. The control system is responsible for adjusting the vibration isolation parameters based on real-time data.If the control system itself has error codes, fails to operate normally due to hardware problems, or experiences communication interruptions with other systems, the vibration isolation device will not be properly adjusted, the vibration isolation effect will not meet expectations, leading to increased system vibration, increased burden on the isolators, and ultimately, failure. 4. Solutions and Optimization Plans: - Regularly inspect the mechanical condition of the vibration isolators, checking for wear, looseness, or deformation, and replace or correct them promptly. - Strengthen the quality control of damper materials, using more durable materials or improving manufacturing processes to reduce the impact of material aging. - Install reliable sensors, ensuring stable power supply and anti-interference capabilities, and regularly calibrate sensors to prevent malfunction. - Regularly maintain the control system, update the software to the latest version, replace hardware components, ensure stable communication networks, and avoid control system failures. - Optimize the layout and connection method of the vibration isolators during the design phase to improve their installation accuracy and reduce the impact of installation deviations on the isolation effect. - Add redundancy design, such as using multiple sets of vibration isolators in parallel, so that if a single isolator fails, the remaining sets can still operate normally, improving the overall system reliability. - Introduce an intelligent monitoring system to monitor the status of vibration isolators in real time, provide early warnings of potential problems, and take proactive preventative measures to reduce issues caused by operational errors.

[0065] As an optional implementation of the aforementioned method for analyzing innovative design problems in electromechanical products based on pre-trained models, the types of technical analysis tools mentioned above may include: requirement identification tools, requirement transformation tools, causal analysis tools, failure analysis tools, component analysis tools, fishbone analysis tools, product verification tools, technology forecasting tools, ideal system tools, creative template tools, target mining tools, and / or resource analysis tools. Of course, in specific practice, the aforementioned types of technical analysis tools also include other tool types, such as failure mode and effects analysis tools, scenario analysis tools, and modular design tools.

[0066] The structured prompt template for the aforementioned requirement identification tool can be represented as follows: "You are an expert in product design and user research, skilled in using requirement identification tools to systematically analyze user needs. Please conduct a comprehensive user requirement identification and classification based on the following structure: Question Description: {question}; 1. User Profile (including typical user types, usage scenarios, and core behaviors); 2. Explicit Needs (needs directly expressed by users); 3. Implicit Needs (needs not explicitly expressed by users but which can be uncovered); 4. Contextual Drivers (scenarios, environments, or triggering conditions that generate the need); 5. Classification of Needs by Emotional and Functional Dimensions (e.g., security, efficiency, usability, emotional connection, etc.); 6. Analysis of User Pain Points and Unmet Needs; 7. Prioritization of Needs (high / medium / low) and Explanation of Reasons; 8. Suggestions for Potential Design Directions (based on the identified core needs); Please answer each part completely, ensuring the content is insightful and systematic. Analysis:".

[0067] The structured prompt template for the aforementioned requirement transformation tool can be represented as: "You are an expert in product design and development, skilled in requirement transformation analysis using the QFD (House of Quality Determination) method. Please complete the task according to the following format: Question Description: {question}; 1. User Requirements List (WHATs); 2. Design Engineering Features List (HOWs); 3. Requirement-Engineering Feature Relationship Matrix Analysis (describe the main correspondences and conflicts); 4. Technical Difficulties and Challenges; 5. Possible Innovation Opportunities; Please fill in each part completely, without omitting any subsections. Analysis:"

[0068] The structured prompt template corresponding to the aforementioned causal analysis tool can be represented as: "You are a causal analysis expert in the field of electromechanical engineering. Please provide your answer strictly according to the following format: Question: {question}; 1. Possible causes: 2. In-depth analysis: 3. Suggested solutions: Please add the corresponding content after each section, and do not omit any headings. Analysis:".

[0069] The structured prompt template for the aforementioned fault analysis tool can be represented as: "You are an expert in systems engineering and reliability analysis, skilled in using Fault Tree Analysis (FTA) to analyze system faults. Please perform fault analysis according to the following format: Problem Description: {question}; 1. Top Event: - What is the final failure event of this problem? 2. Intermediate Events: - What are the main intermediate events that lead to the top event? Please list the causal relationships between them. 3. Root Events: - What is the root cause or failure behind each intermediate event? List them and describe their impact. 4. Mitigation Measures and Optimization Schemes: - Propose optimization schemes or measures to reduce the probability of system failure. Please complete each section completely and do not omit any subsections. Analysis:".

[0070] The structured prompt template for the component analysis tool described above can be expressed as follows: "You are an expert in system design and optimization, skilled in functional analysis using component analysis. Please complete the task according to the following format: Problem Description: {question}; 1. System Component Decomposition: - List all components in the system and briefly describe the function of each component. 2. Functional Relationship Analysis Between Components: - Standard Functions: Which components have standard and effective functions? Please list their roles and contributions. - Redundant Functions: Which components have redundant or unnecessary functions? Please describe them and provide possible optimization solutions. - Insufficient Functions: Which components have insufficient or missing functions? Please analyze their deficiencies and propose improvement measures. - Harmful Functions: Which components' functions may have a negative impact on the system or waste resources? Please identify them and provide solutions. 3. Component Performance: - Analyze the difference between the performance of each component and the expected goals, and describe their advantages and disadvantages. 4. Optimization Directions: - Based on the functional analysis results, propose optimization and improvement directions to help the system reduce resource consumption and improve overall efficiency. Please fill in each part completely and do not omit any subsections. Analysis:"

[0071] The structured prompt template corresponding to the fishbone analysis tool mentioned above can be expressed as: "You are an expert in product design and development, skilled in using fishbone analysis to find the root causes of product problems. Please analyze according to the following format: Problem description: {question}; 1. Possible causes related to Man: 2. Possible causes related to Machine: 3. Possible causes related to Material: 4. Possible causes related to Method: 5. Possible causes related to Environment: 6. Possible causes related to Measurement: Please list the possible causes in detail for each aspect, identify the key causes, and finally propose a solution. Analysis:".

[0072] The structured prompt template corresponding to the above product inspection tool can be expressed as: "You are an expert in product innovation design and are skilled at using product inspection tools to identify innovation opportunities. Please analyze and explore potential innovation directions based on the following nine thinking dimensions and the given questions: Question Description: {question}; 1. Are there any other problems? (Are there any unnoticed defects, bottlenecks, or hidden dangers in the current product / system?): 2. Can it be borrowed? (Can existing solutions, resources, or technologies from other fields be borrowed?): 3. Can it be substituted? (Can other materials, processes, or methods be used to replace the current solution?): 4. Can it be combined? (Can functions, processes, or methods be combined?)" 5. Can it be changed? (Can the usage method, structural form, process logic, etc. be changed?) 6. Can it be expanded? (Can the structure, function, scope of use, or market size be expanded?) 7. Can it be reduced? (Can the volume, cost, complexity, or functional boundaries be reduced?) 8. Can it be reversed? (Can the usage order, logical relationship, energy flow, etc. be reversed?) 9. Can it be readjusted? (Can the structural layout, functional position, or resource allocation be reconfigured?) 10. Based on the above analysis, summarize feasible innovation opportunities: Please fill in each part completely, and the content should be logical and feasible. Analysis:

[0073] The structured prompt template for the aforementioned technology forecasting tool can be represented as: "You are a product technology development trend analysis expert, skilled in using technology system theory and the eight laws of evolution to predict the future development direction of products. Please conduct a comprehensive analysis around the question posed, based on the following structure: Question Description: {question}; 1. Current technology system description (including target functions and main structure): 2. Existing technical contradictions or limiting factors: 3. Analysis of applicable evolutionary laws: - Completeness Law: - Energy Transfer Law: - Dynamic Evolution Law: - Law for Improving Ideality: - Subsystem Imbalance Evolution Law: - Evolution Law Towards a Supersystem: - Evolution Law Towards a Microscopic Level: - Coordination Evolution Law: 4. Possible evolutionary directions or scheme predictions (based on the laws): 5. Inspiration for future design and innovation: Please fill in each part completely, using clear and organized language to explain the applicability and development possibilities of each law in this system. Analysis:".

[0074] The structured prompt template for the aforementioned ideal system tool can be represented as: "You are an expert in product design and system innovation, skilled at analyzing problems and proposing improvement directions using the 'ideal system tool.' Please conduct a system ideality analysis around the problem and complete the task according to the following structure: Problem Description: {question}; 1. Description of the current system state: - What is the goal of the current system? - How does the current system achieve this goal? - What are the key contradictions or limitations in the current solution? 2. Definition of Ideal Final Result (IFR): - If the system reaches a completely ideal state, how should it achieve its goal? - Is it possible to achieve this without introducing new..." 3. Gap Analysis Between the Current System and the Ideal System: - What are the main gaps between the current solution and the IFR? - What factors limit the system's evolution towards IFR? 4. Possible Ways to Improve the System's Ideality: - Are there alternative resources that can be utilized? - Can we approach IFR through functional separation, functional transfer, system simplification, etc.? - Are there any cross-domain solution paths for reference? 5. Suggested Optimization Scheme (Innovative): - Based on the above analysis, please propose at least one improvement suggestion and explain its feasibility, innovativeness, and effectiveness. Please fill in each section completely and do not omit any subsections. Analysis:

[0075] The structured prompt template corresponding to the above creative template tool can be expressed as: "As a product innovation design expert, please complete the following structured analysis using attribute association analysis: Problem description: {question}; 1. Identification of the product's main and auxiliary functions; 2. Listing of internal attributes: - List only the main specific attributes, no more than 5; 3. Listing of external attributes: - List only the main specific attributes, no more than 5; 4. Correlation analysis of internal and external attributes (indicate significant correlations or potential innovative combinations); 5. Potential innovation points (based on the above correlation combinations, propose possible novel design schemes or functional combinations); 6. Feasibility and potential assessment of innovation points: Please fill in each part completely, do not omit any subsections. Analysis:".

[0076] The structured prompt template corresponding to the above-mentioned goal mining tool can be represented as: "As an innovative design expert, please use the Concept Fan System Framework for goal mining analysis. Expand through the following steps: Problem Background: {question}; 1. Goal Layer (Abstract Goal): - Please list one higher-dimensional ultimate goal (using a verb + noun structure, such as "optimize system performance"); 2. Direction Layer (Implementation Path): - For the ultimate goal, brainstorm at least three implementation directions (cross-domain, interdisciplinary perspectives); 3. Concept Layer (Innovation Solution): - Transform each direction into an executable solution; 4. Innovation Opportunity Points: - Identify solutions with cross-disciplinary integration potential. Please fill in each part completely, without omitting any subsections. Analysis:".

[0077] The structured prompt template for the resource analysis tool mentioned above can be represented as: "You are an expert in product design and development, skilled in using resource analysis tools for product innovation and optimization. Please conduct a resource analysis of the problem based on the following structure, uncover potential resources, and propose innovative ideas: Problem Description: {question}; 1. Overview of the basic functions and structure of the current system or product: 2. Resource identification and analysis (expand from the following six categories of resources, listing resources and their potential value): - Material resources (components, materials, structures, etc.): - Energy resources (electrical energy, thermal energy, kinetic energy, environmental energy, etc.): - Information resources (sensor data, user data, environmental information, etc.): - Functional resources (existing functions, additional functions, backup functions, etc.): - Time resources (idle time, cycle time, response time, etc.): - Spatial resources (internal space, external space, usage environment, etc.): 3. Suggestions for uncovering and utilizing potential unused or inefficiently used resources: 4. Analysis of the possibility of resource reorganization or integration (such as cross-resource type collaboration): 5. Innovative design ideas or improvement suggestions based on resource optimization: Please answer each part completely, ensuring that the content is logically clear and the viewpoints are innovative. Analysis:".

[0078] Please see Figure 2 The diagram shown is a schematic of a pre-trained model-based electromechanical product innovation design problem analysis system provided in an embodiment of this application; the embodiment of this application also provides a pre-trained model-based electromechanical product innovation design problem analysis system 200, including: a terminal device 210 and a server 220; Terminal device 210 is used to send a description of the innovative design problem of the target electromechanical product and a type of technical analysis tool to the server; Server 220 receives the innovative design problem description text and technique analysis tool type sent by the terminal device for the target electromechanical product; it searches for the structured prompt template corresponding to the technique analysis tool type in the prompt template library, which stores the correspondence between technique analysis tool types and structured prompt templates; it fills the innovative design problem description text into the structured prompt template to obtain the technique analysis prompt engineering text; and it uses a pre-trained model to reason about the technique analysis prompt engineering text to obtain the technique analysis result of the target electromechanical product. The technique analysis prompt engineering text is used to guide the thought chain prompt reasoning process of the pre-trained model.

[0079] It should be understood that this system corresponds to the above-described embodiment of the method for analyzing innovative design problems of electromechanical products based on pre-trained models, and is capable of executing the various steps involved in the above method embodiment. The specific functions of this system can be found in the description above, and detailed descriptions are appropriately omitted here. The system includes at least one software functional module that can be stored in memory or embedded in the operating system (OS) in software form.

[0080] Please see Figure 3 The diagram shown is a structural schematic of the electromechanical product innovative design problem analysis device based on a pre-trained model provided in this application embodiment; this application embodiment also provides an electromechanical product innovative design problem analysis device 300 based on a pre-trained model, including: Tool type acquisition module 310 is used to acquire the description text of innovative design problems and the tool type of technique analysis for the target electromechanical product.

[0081] The prompt template lookup module 320 is used to search for the structured prompt template corresponding to the technique analysis tool type from the prompt template library. The prompt template library stores the correspondence between the technique analysis tool type and the structured prompt template.

[0082] The engineering text acquisition module 330 is used to fill the innovative design problem description text into the structured prompt template to obtain the technical analysis prompt engineering text.

[0083] The analysis result acquisition module 340 is used to reason about the technical analysis prompt engineering text through the pre-trained model to obtain the technical analysis result of the target electromechanical product. The technical analysis prompt engineering text is used to guide the thought chain prompt reasoning process of the pre-trained model.

[0084] As an optional implementation of the above-mentioned device, the electromechanical product innovation design problem analysis device based on a pre-trained model further includes: The training dataset acquisition module is used to acquire the fine-tuning training dataset, which includes: problem description samples and analysis result samples.

[0085] The model problem training module is used to fine-tune the pre-trained model using problem description samples and analysis result samples.

[0086] As an optional implementation of the above-mentioned device, the training dataset acquisition module includes: The sample data acquisition submodule is used to acquire a problem description sample and the corresponding structured prompt template.

[0087] The prompt template filling submodule is used to fill the problem description sample into the structured prompt template corresponding to the problem description sample, so as to obtain the problem analysis prompt engineering text.

[0088] The language model inference submodule is used to input the problem analysis prompt engineering text into the general language model, so that the general language model can infer the problem description sample and obtain the analysis result text corresponding to the problem description sample. The number of model parameters of the general language model is greater than the number of model parameters of the pre-trained model.

[0089] The text correction and annotation submodule is used to correct and annotate the analysis result text corresponding to the problem description sample to obtain the analysis result sample.

[0090] The sample data addition submodule is used to add problem description samples and analysis result samples to the fine-tuning training dataset.

[0091] As an optional implementation of the above-mentioned device, the electromechanical product innovation design problem analysis device based on a pre-trained model further includes: The preset matrix insertion module is used to obtain the target language model and insert a preset matrix into the target language model to obtain a pre-trained model.

[0092] The aforementioned model training module also includes: The preset matrix update submodule is used to update the preset matrix in the pre-trained model by using problem description samples as training data and analysis result samples as training labels, while freezing the model parameters of the target language model in the pre-trained model.

[0093] As an optional implementation of the above-mentioned device, the analysis result acquisition module includes: The calling program retrieves a submodule used to obtain the model calling program for the pre-trained model.

[0094] The text reasoning and analysis submodule is used to input the technical analysis prompt engineering text into the pre-trained model through the model calling program, so that the pre-trained model can perform reasoning analysis on the technical analysis prompt engineering text and output the technical analysis results of the target electromechanical product.

[0095] As an optional implementation of the above-mentioned device, the types of technology analysis tools include: demand identification tools, demand transformation tools, cause-and-effect analysis tools, fault analysis tools, component analysis tools, fishbone analysis tools, product inspection tools, technology forecasting tools, ideal system tools, creative template tools, target mining tools, and / or resource analysis tools.

[0096] It should be understood that this device corresponds to the above-described embodiment of the method for analyzing innovative design problems of electromechanical products based on pre-trained models, and is capable of performing the various steps involved in the above method embodiments. The specific functions of this device can be found in the description above, and detailed descriptions are appropriately omitted here. The device includes at least one software functional module that can be stored in memory or embedded in the device's operating system (OS) in the form of software or firmware.

[0097] Please see Figure 4The diagram shows a structural schematic of an electronic device provided in an embodiment of this application. An electronic device 400 provided in this application includes a processor 410 and a memory 420. The memory 420 stores machine-readable instructions executable by the processor 410. When the machine-readable instructions are executed by the processor 410, the method described above is performed.

[0098] This application embodiment also provides a computer-readable storage medium 430, on which a computer program is stored. This computer program, when executed by the processor 10, performs the method described above. The computer-readable storage medium 430 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0099] This application also provides a computer program product, including: a computer program or computer instructions, which are executed by a processor to perform the method described above.

[0100] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0101] It should be understood that the disclosed apparatus and methods can also be implemented in other ways, as provided in the embodiments of this application. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in the flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending primarily on the functions involved.

[0102] Furthermore, the functional modules of each embodiment in this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part. In addition, in the description of this specification, the reference to terms such as "one embodiment," "some embodiments," "example," "specific example," "some examples," etc., means that the specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0103] The above description is only an optional implementation of the embodiments of this application, but the protection scope of the embodiments of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the embodiments of this application should be covered within the protection scope of the embodiments of this application.

Claims

1. A method for analyzing innovative design problems of electromechanical products based on pre-trained models, characterized in that, include: Obtain descriptions of innovative design problems and types of technical analysis tools for the target electromechanical products; The structured prompt template corresponding to the technique analysis tool type is searched from the prompt template library, which stores the correspondence between technique analysis tool types and structured prompt templates; Fill the structured prompt template with the innovative design problem description text to obtain the technique analysis prompt engineering text; The technical analysis prompts of the target electromechanical product are obtained by reasoning through the technical analysis prompts of the engineering text using a pre-trained model. The technical analysis prompts of the engineering text are used to guide the thought chain prompts of the pre-trained model in its reasoning process.

2. The method according to claim 1, characterized in that, Before inferring the technique analysis prompt engineering text using the pre-trained model, the method further includes: Obtain a fine-tuning training dataset, which includes: problem description samples and analysis result samples; The pre-trained model is fine-tuned using the problem description samples and the analysis result samples.

3. The method according to claim 2, characterized in that, The process of obtaining the fine-tuning training dataset includes: Obtain a problem description sample and the corresponding structured prompt template; The problem description sample is filled into the structured prompt template corresponding to the problem description sample to obtain the problem analysis prompt engineering text; The problem analysis prompt engineering text is input into a general language model so that the general language model can reason about the problem description sample and obtain the analysis result text corresponding to the problem description sample. The number of model parameters of the general language model is greater than the number of model parameters of the pre-trained model. The analysis result text corresponding to the problem description sample is corrected and annotated to obtain the analysis result sample; The problem description sample and the analysis result sample are added to the fine-tuning training dataset.

4. The method according to claim 2, characterized in that, Before fine-tuning the pre-trained model using the problem description samples and the analysis result samples, the method further includes: Obtain the target language model, and insert a preset matrix into the target language model to obtain the pre-trained model; The step of fine-tuning the pre-trained model using the problem description samples and the analysis result samples includes: With the model parameters of the target language model in the pre-trained model frozen, the preset matrix in the pre-trained model is updated using the problem description samples as training data and the analysis result samples as training labels.

5. The method according to claim 1, characterized in that, The step of reasoning about the technique analysis prompts in the engineering text using a pre-trained model includes: Obtain the model calling program for the pre-trained model; The technique analysis prompt engineering text is input into the pre-trained model through the model calling program, so that the pre-trained model can perform inference analysis on the technique analysis prompt engineering text and output the technique analysis result of the target electromechanical product.

6. The method according to any one of claims 1-5, characterized in that, The types of analytical tools include: demand identification tools, demand transformation tools, cause-and-effect analysis tools, fault analysis tools, component analysis tools, fishbone analysis tools, product inspection tools, technology forecasting tools, ideal system tools, creative template tools, target mining tools, and / or resource analysis tools.

7. A system for analyzing innovative design problems of electromechanical products based on a pre-trained model, characterized in that, include: Terminal devices and servers; The terminal device is used to send the innovative design problem description text and technique analysis tool type of the target electromechanical product to the server; The server is used to receive the innovative design problem description text and technique analysis tool type sent by the terminal device for the target electromechanical product; and to search for the structured prompt template corresponding to the technique analysis tool type from the prompt template library, which stores the correspondence between technique analysis tool types and structured prompt templates. The innovative design problem description text is filled into the structured prompt template to obtain the technique analysis prompt engineering text; the technique analysis prompt engineering text is reasoned through the pre-trained model to obtain the technique analysis result of the target electromechanical product, and the technique analysis prompt engineering text is used to guide the thought chain prompt reasoning process of the pre-trained model.

8. An electronic device, characterized in that, include: A processor and a memory, the memory storing machine-readable instructions executable by the processor, the machine-readable instructions being executed by the processor to perform the method of any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that, include: A computer program or computer instructions that, when executed by a processor, perform the method according to any one of claims 1 to 6.

Images