Intelligent welding process evaluation optimization system based on artificial intelligence big language model

Abstract

The invention discloses an intelligent welding process evaluation optimization system based on an artificial intelligence big language model, and aims to improve the screening and optimization efficiency of a welding process evaluation report. According to the system, a natural language processing technology is used for analyzing welding requirements input by a user, relevant parameters are automatically extracted from a welding process evaluation report database, and evaluation reports are intelligently sorted and screened through a multi-dimensional data analysis module. The system generates optimized welding process suggestions according to multiple factors such as welding materials, process parameters and structural complexity. Meanwhile, the system has a self-adaptive learning function, process optimization adjustment can be carried out based on historical data and simulation results, and continuous improvement of the recommended process is ensured. The system can also be integrated with welding simulation tools (such as ANSYS and SYSWELD), the feasibility of a recommended process is verified, and the system is suitable for welding process selection and optimization of key structures such as pressure containers and pipelines in complex welding projects. According to the method, the efficiency, accuracy and reliability of welding process evaluation are remarkably improved, and the method has wide industrial application prospects.

Description

Technical Field

[0001] The present invention relates to the field of welding technology, and in particular to an intelligent welding process evaluation and optimization system based on an artificial intelligence large language model. The system helps welding factories or engineers screen and optimize welding process evaluation reports through natural language processing, intelligent analysis of large language models, and adaptive learning functions, and provides the best selection and optimization suggestions for welding processes, which is applicable to various welding scenarios and complex process requirements. Background Art

[0002] In the welding industry, welding process assessment is a key link to ensure welding quality, especially for key structures such as pressure vessels and pipelines. Welding process assessment ensures the feasibility and safety of welding processes in practical applications. At present, the welding process assessment management system commonly used in factories has added some artificial intelligence technologies, such as optimization algorithms based on artificial neural networks, which can help optimize or evaluate some parameters. However, such systems still have certain limitations in processing natural language input, cross-dimensional data analysis, multi-factor decision-making and automatic optimization.

[0003] In recent years, the development of large language models (such as Tongyi Qianwen, iFlytek Spark, Wenxin Yiyan, and GPT-4) has provided new intelligent solutions for welding process assessment. Large language models can process complex natural language descriptions, combine welding process parameters and historical assessment data, generate the best welding process assessment plan, and continuously optimize the recommendation strategy in adaptive learning. The present invention aims to improve the intelligence level of welding process assessment, reduce human subjective factors, and provide a more accurate and efficient assessment plan by adding large language models. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent welding process assessment optimization system based on a large language model. The system can automatically analyze the welding requirements input by the user through the natural language processing module, conduct multi-dimensional analysis in combination with the welding process assessment report in the database, select the best process assessment plan, and generate detailed process optimization suggestions. At the same time, the system also has an adaptive learning function, which can continuously optimize the assessment strategy based on historical data, user feedback and welding simulation results to ensure that the recommended welding process has high adaptability and superiority.

[0005] The main functional modules of the system include: Natural language input module: allows users to input welding requirements through natural language. The system uses a large language model to parse these requirements and convert them into structured parameters that can be used for database retrieval.

[0006] Data extraction module: extract welding process qualification report from the factory database to obtain key information such as welding parameters, material properties, welding quality assessment, etc.

[0007] Multi-dimensional analysis module: Based on user needs and extracted assessment reports, a large language model is used to conduct multi-dimensional analysis of different assessment schemes, taking into account factors such as material properties, structural complexity, welding methods, environmental conditions, etc., to intelligently select the best solution.

[0008] Optimal report generation module: The system generates the best welding process assessment report based on the analysis results and provides optimization suggestions, detailing the selection of specific welding parameters such as welding current, welding speed, heat input, etc.

[0009] Adaptive learning module: The system optimizes process assessment strategies and improves the accuracy of future process recommendations by learning historical assessment data and simulation results.

[0010] Through these functional modules, the system of the present invention can significantly improve the efficiency and accuracy of welding process assessment, reduce manual intervention, and improve the adaptability and reliability of welding process. Beneficial Effects

[0011] Compared with the prior art, the present invention has the following beneficial effects: Efficiency and automation: The system parses natural language input through a large language model, automatically screens and optimizes welding process assessment reports, and greatly improves the efficiency of process assessment.

[0012] Multi-dimensional comprehensive analysis: By considering multiple factors (materials, welding parameters, structural complexity, etc.), the system can generate a more accurate and comprehensive welding process assessment plan.

[0013] Adaptive learning: Based on historical data and user feedback, the system can continuously improve the assessment strategy to ensure that future process assessments are more adaptive and superior.

[0014] Reduce subjectivity: The system avoids subjective bias in the traditional manual evaluation process through big data analysis and automatic recommendation, ensuring the fairness and consistency of the selection process.

[0015] Integration with simulation technology: The system can be integrated with welding simulation tools (such as ANSYS, SYSWELD) to further verify and optimize the recommended process through simulation results, thereby improving the feasibility and actual effect of the process. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 : Framework diagram of intelligent welding process assessment and optimization system based on large language model.

[0017] Figure 2: Multi-dimensional analysis flow chart of welding process assessment report.

[0018] Figure 3 : Workflow diagram of the adaptive learning module. DETAILED DESCRIPTION

[0019] The following is a detailed description of a specific implementation of the present invention, explaining the functions and workflow of each module.

[0020] 1. Natural language input module: Users can input welding requirements through natural language, such as welding materials, welding positions, geometric structures, working environments, usage standards, etc. The large language model is responsible for parsing the input content and extracting welding-related structured parameters. These parameters are then used for subsequent welding process assessment report screening and analysis.

[0021] 2. Data extraction module: The system extracts all welding process qualification reports that match user requirements from the welding process qualification database. The key parameters in each report (such as material type, welding process, welding quality assessment, etc.) are extracted through this module and stored in the multidimensional analysis module.

[0022] 3. Multi-dimensional analysis module: This module is the core part of the system. It uses a large language model to perform multi-dimensional intelligent analysis on the extracted welding process assessment report. The analysis dimensions include: Material properties: such as the welding performance of the material, changes in the heat-affected zone, etc.

[0023] Process parameters: such as welding current, welding speed, heat input, etc.

[0024] Structural complexity: the effects of different geometries and structures on welding.

[0025] Environmental conditions: such as welding temperature, pressure, climate environment, etc.

[0026] The system intelligently sorts the evaluation reports and recommends the best process based on user needs and the analysis results of these factors.

[0027] 4. Optimal report generation module: Based on the results of multi-dimensional analysis, the system automatically generates a welding process assessment report, which contains detailed welding parameter optimization suggestions. These suggestions include key parameters such as welding current, speed, heat input, etc., and provide optimization solutions based on actual operating conditions to ensure welding quality and stability.

[0028] 5. Adaptive learning module: The system has adaptive learning capabilities and can continuously optimize the assessment strategy by continuously learning from successful cases, user feedback and simulation results of historical assessment reports. For example, the system can identify which process parameter combinations are most effective and give priority to similar solutions in future assessments. The adaptive learning module can also use welding simulation tools (such as ANSYS, SYSWELD) for verification and adjust the recommended strategy of the assessment report in real time.

[0029] 6. Integration with welding simulation tools: The system of the present invention can be integrated with welding simulation tools (such as ANSYS, SYSWELD), and the simulation tools can be used to theoretically verify the recommended welding process. The simulation results can help optimize the process parameters, such as verifying the stress distribution in the heat-affected zone, the strength change of the material, etc. Through simulation feedback, the system can dynamically adjust the process parameters to ensure that the recommended welding scheme is highly feasible in actual operation. Example

[0030] Example 1: Welding procedure assessment for pressure vessels In a welding factory, the best welding process needs to be selected for a high-pressure container. The operator describes the material type, welding position, environmental conditions and other requirements of the container through the system's natural language input module. The system uses a large language model to parse the requirements, automatically selects relevant welding process assessment reports from the database, and recommends the best welding process assessment report by considering material properties, welding process parameters and working conditions through a multi-dimensional analysis module.

[0031] The system also automatically generates an optimization proposal for parameters including welding current, welding speed and heat input. Subsequently, the system verifies the feasibility and safety of the process by combining it with welding simulation tools (such as ANSYS and SYSWELD) to ensure that it meets the welding requirements of pressure vessels.

[0032] Example 2: Welding assessment of complex geometric components A factory needs to select a welding process for a group of complex geometric parts. Based on the input welding requirements, the system analyzes the characteristics of the geometric structure and extracts the relevant process evaluation report.

Claims

1. An intelligent welding process evaluation and optimization system based on artificial intelligence large language model, characterized in that: The system includes: a natural language input module, which is used to receive welding requirements input by users and perform natural language processing through a large language model to parse the requirements; a data extraction module, which is used to extract welding parameters and process information from a welding process assessment report database; a multidimensional analysis module, which is used to perform multidimensional analysis on the extracted welding process assessment report based on the user's welding requirements and perform intelligent screening through the calculation and reasoning capabilities of the large language model; an optimal report generation module, which is used to generate welding process optimization suggestions and recommended welding process assessment reports based on the multidimensional analysis results; and an adaptive learning module, which is used to perform continuous learning and optimization through a large language model based on historical data and user feedback, so as to improve the screening and recommendation strategies of welding process assessment reports.

2. The system according to claim 1, characterized in that The natural language input module can parse the welding requirements input by the user, including but not limited to information such as welding materials, welding positions, welding structures, environmental conditions and welding quality requirements.

3. The system according to claim 1 or 2, characterized in that: The data extraction module can extract welding parameters from the welding process assessment report, including but not limited to welding current, welding speed, heat input, material properties, welding process method and welding assessment results.

4. The system according to claim 1, characterized in that The multi-dimensional analysis module can comprehensively consider multiple factors, including material properties, welding parameters, structural complexity, stress distribution and heat-affected zone, and perform intelligent screening and recommendation.

5. The system according to claim 1, characterized in that The optimal report generation module can automatically generate a welding process assessment report containing detailed welding parameter optimization suggestions to improve the quality and stability of the welding process.

6. The system according to claim 1, characterized in that The adaptive learning module can continuously learn and optimize through a large language model based on user historical input, process evaluation reports and welding simulation results, and update the recommended rules for the welding process.

7. The system according to claim 1, characterized in that The system can be integrated with welding simulation software to verify the feasibility of the welding process assessment report through simulation results, and adjust the welding process parameters to optimize the actual welding process.

8. The system according to claim 1, characterized in that The system can dynamically adjust the process recommendation strategy according to the specific welding task to ensure the applicability of the selected process under different working conditions.

9. The system according to claim 1, characterized in that The system is capable of processing welding requirement input in multiple languages ​​and is applicable to international welding standards and welding procedure assessment in different regions.

10. The system according to claim 1, characterized in that The system can automatically generate welding procedure qualification documents that meet industry standards, simplify the preparation process of qualification reports and reduce manual intervention.

Images