Casting defect diagnosis and optimization system based on artificial intelligence large language model

Abstract

The invention relates to a casting defect diagnosis and optimization system based on an artificial intelligence large language model. The system combines an Internet of Things sensor, computer vision, real-time data analysis, global data integration and a large language model (such as Festival Starry, oneself, Tonyuan and GPT-4) in artificial intelligence to realize real-time monitoring, diagnosis, prediction and process optimization of defects in the casting production process. The system comprises a plurality of modules, wherein a real-time data acquisition module is responsible for acquiring production parameters; the knowledge base and global data integration module learns defect data and production cases of the global casting industry through a large language model, and constructs a dynamically updated knowledge base; the vision and text analysis module performs defect identification and root cause analysis in combination with the image and text data; the predictive diagnosis and prevention module predicts potential defects and gives out early warning by analyzing production data in real time; the generative process optimization module is combined with a simulation tool to generate optimal process parameters, and defects are reduced; and the real-time monitoring and feedback module automatically adjusts production parameters to ensure the stability and high quality of the production process. Through the cooperation and knowledge sharing module, the system can share data with global casting enterprises and experts, and the self-learning ability of the system is improved. The intelligent level of casting production is greatly improved, the defect occurrence rate is effectively reduced, and the production efficiency and the product quality are improved.

Description

Technical Field

[0001] The present invention relates to casting technology and artificial intelligence technology, in particular to an intelligent system for casting defect diagnosis, prediction and process optimization based on large language models in artificial intelligence (such as iFlytek Spark, Wenxin Yiyan, Tongyi Qianwen, and GPT-4). Background Art

[0002] The casting process is a complex multi-step physical and chemical process, which is affected by many factors such as temperature, material composition, cooling rate, etc. Even in mature processes, defects are inevitable. Traditional casting defect diagnosis methods usually rely on expert experience or preset rule systems. These methods have great limitations in dealing with real-time defect prevention and diagnosis under complex production conditions.

[0003] With the rapid development of technologies such as the Internet of Things, artificial intelligence, and big data, especially the widespread application of large language models (such as iFlytek Spark, Wenxin Yiyan, Tongyi Qianwen, and GPT-4), these advanced technologies help to break through the limitations of traditional casting defect diagnosis systems and provide more intelligent, real-time, and dynamic diagnosis and process optimization solutions. However, there is currently no complete solution that organically combines large artificial intelligence language models with casting defect diagnosis and process optimization systems. Summary of the invention

[0004] Technical problem: The purpose of the present invention is to provide a casting defect diagnosis and optimization system based on artificial intelligence large language model, which aims to monitor the casting process in real time, diagnose defects, predict potential problems, and optimize process parameters by combining Internet of Things sensors, computer vision, global data integration and large language models to reduce the occurrence of defects, improve production efficiency and enhance product quality.

[0005] Technical solution: The present invention proposes an intelligent casting defect diagnosis and optimization system, which includes the following modules: Real-time data acquisition module: used to collect key parameters in the casting process, such as temperature, pressure, material composition, cooling rate and other data through IoT sensors, and transmit these data to the system background through wireless or wired networks for further analysis and processing.

[0006] Knowledge base and global data integration module: Through large-scale language models such as iFlytek Spark, Wenxin Yiyan, Tongyi Qianwen, and GPT-4, we learn from the global casting defect database, literature, production cases, etc., build a dynamically updated knowledge base, and generate intelligent diagnosis and process optimization strategies.

[0007] Vision and text analysis module: Through computer vision technology, combined with real-time collected casting defect image data, the defect type is automatically identified; at the same time, combined with production records, sensor data and other text information, a large language model is used to perform root cause analysis on the data and provide corresponding solutions.

[0008] Predictive diagnosis and prevention module: By analyzing sensor data and using large language models, it predicts possible defects in the production process and prevents the occurrence of defects by suggesting adjustments to production parameters.

[0009] Generative process optimization module: By combining large language models with casting simulation tools (such as ProCAST and AnyCasting), it simulates different process conditions and generates the optimal process adjustment plan to reduce the occurrence of defects.

[0010] Real-time monitoring and feedback module: monitor production parameters in real time and automatically adjust production process parameters according to the prediction results to ensure the stability of the casting process and product quality.

[0011] Collaboration and knowledge sharing module: By sharing data with global foundries, research institutions and experts, the system's knowledge base is continuously enriched, and diagnostic rules and process optimization strategies are dynamically updated by learning new data. Beneficial Effects

[0012] The casting defect diagnosis and optimization system of the present invention has the following beneficial effects: Intelligent learning and dynamic updating: The system integrates and updates the global casting defect knowledge base in real time through the dynamic learning capability of large language models, ensuring the cutting-edge nature of diagnosis and optimization strategies.

[0013] Real-time prediction and prevention: The system can monitor production data in real time, predict potential casting defects, and automatically provide process adjustment suggestions to reduce the occurrence of defects.

[0014] Process optimization and feedback: Through the combination of simulation tools and large language models, the system can generate the optimal process parameter adjustment plan to improve production efficiency and product quality.

[0015] Collaboration and knowledge sharing: Through the global data sharing platform, the system can dynamically obtain new cases and data from the foundry industry, and continuously optimize and improve system performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 : Schematic diagram of the overall system structure.

[0017] Figure 2 : Workflow diagram of the real-time data acquisition module.

[0018] Figure 3: Defect identification and analysis flow chart of the visual and text analysis modules.

[0019] Figure 4 : Schematic diagram of the simulation and optimization process of the generative process optimization module.

[0020] Figure 5 : Schematic diagram of automatic control of real-time monitoring and feedback module. DETAILED DESCRIPTION Specific implementation method: Example 1: Operation of real-time data acquisition module ( Figure 1 ) On a foundry production line, the system's real-time data acquisition module collects key parameters of the production process, such as temperature, pressure, cooling rate, and material composition, through IoT sensors. Sensors are installed at different locations on the casting equipment, and the real-time collected data is transmitted to the system backend via a wireless network. The system can monitor the real-time changes of each key parameter and provide this data to the predictive diagnosis and prevention module for further analysis.

[0022] Example 2: Defect Identification by Visual and Text Analysis Modules ( Figure 2 ) This module combines high-definition cameras and computer vision technology to collect images of casting surfaces in real time, and uses large language models to analyze image data to identify surface defects such as pores and cracks. At the same time, the system also combines sensor data and text data recorded by operators, uses large language models to analyze abnormal conditions in the production process, and generates defect root cause analysis reports.

[0023] Example 3: Application of predictive diagnosis and prevention module ( Figure 3 ) In the actual production process, the system analyzes real-time data such as temperature, pressure and cooling rate, combined with a large language model, to predict possible defects in production. When the system predicts that cracks may occur, the system automatically alerts the operator and recommends adjusting the temperature or cooling rate to prevent defects.

[0024] Example 4: Operation of the Generative Process Optimization Module ( Figure 4 ) The system combines large language models and casting simulation tools (such as ProCAST, AnyCasting) to simulate different production conditions. In the casting task, the system generates an optimal combination of process parameters based on different temperature and cooling rate conditions, reducing the occurrence of casting defects and improving production efficiency.

[0025] Example 5: Automatic control of real-time monitoring and feedback module ( Figure 5 ) During the production process, the system monitors the key parameters of the casting equipment in real time and predicts possible future production problems through large-scale language model analysis. If the temperature is too high or the pressure is unstable, the system will automatically adjust the production parameters to ensure the quality stability of the casting. At the same time, all adjustments and feedback information will be recorded and stored in the system's knowledge base for future optimization.

Claims

1. A casting defect diagnosis and optimization system based on artificial intelligence large language model, characterized in that: The system includes: a real-time data acquisition module, which is used to collect key parameters in the production process, including temperature, pressure, material composition and cooling rate, through the Internet of Things sensors integrated on the casting equipment, and transmit the collected data to the system background; a knowledge base and global data integration module, which is used to learn global casting defect data, literature and production cases through a large language model, build a dynamically updated casting defect knowledge base, and generate diagnosis and optimization strategies; a visual and text analysis module, which is used to analyze casting defect images in combination with computer vision technology, and perform text analysis in combination with production records and sensor data, identify defect types and perform root cause analysis; a predictive diagnosis and prevention module, which is used to analyze real-time production data through a large language model, predict possible defects, and recommend adjustment of process parameters for prevention; a generative process optimization module, which is used to combine a large language model with a casting simulation tool to generate the optimal process adjustment plan to reduce the occurrence of defects; a real-time monitoring and feedback module, which is used to monitor production parameters in real time, and automatically adjust production parameters according to prediction and diagnosis results to ensure the quality and stability of the production process; a collaboration and knowledge sharing module, which is used to share data through the global foundry industry and expert network, and dynamically update the knowledge base and strategy of the casting defect diagnosis and optimization system.

2. The system according to claim 1, characterized in that The real-time data acquisition module can collect multiple parameters related to the casting process, such as temperature, pressure, material composition, etc., and transmit the data to the system central processing unit via a wireless or wired network.

3. The system according to claim 1, characterized in that The knowledge base and global data integration module learns defect diagnosis cases, literature and production data of global foundries through a large language model, and continuously optimizes and updates the system's diagnostic rules to ensure that the knowledge base always keeps the latest defect solutions and process improvement suggestions.

4. The system according to claim 1, characterized in that The visual and text analysis module uses computer vision technology to analyze casting images in real time, and combines production records and sensor data to use large language models to perform root cause analysis and generate improvement suggestions.

5. The system according to claim 1, characterized in that The predictive diagnosis and prevention module analyzes the production data collected by sensors, uses a large language model to predict possible defects in the casting production process, and recommends that operators make process adjustments to reduce the probability of defects.

6. The system according to claim 1, characterized in that The generative process optimization module combines a large language model with a casting simulation tool to simulate different production conditions, generate optimal process parameters, reduce defects, and improve production efficiency.

7. The system according to claim 1, characterized in that The real-time monitoring and feedback module monitors production parameters in real time through sensors, analyzes production data through a large language model, and automatically adjusts parameters on the production line, such as temperature, cooling rate, and pressure, to ensure the stability and high quality of the casting production process.

8. The system according to claim 1, characterized in that The collaboration and knowledge sharing module allows sharing of data and experience with foundries, research institutes and expert networks around the world. The system continuously improves its diagnostic and process optimization capabilities by learning from new data and global cases.

9. A casting defect diagnosis and optimization method based on artificial intelligence large language model, characterized in that: The following steps are involved: Collect real-time data in casting production through IoT sensors; use large language models to learn defect data, literature and production cases in the global foundry industry and update the knowledge base; analyze casting defects through visual and text analysis modules, identify defect types and analyze their root causes; Analyze real-time data through large language models, predict potential casting defects, issue warnings and suggest production process adjustments; combine large language models with simulation tools to generate optimal process parameter adjustment plans to reduce defects; The casting process is monitored in real time and process parameters are automatically adjusted through a feedback mechanism to ensure casting quality.

10. The method according to claim 9, characterized in that In the steps described above, defect data of the global foundry industry is learned through a large language model, and the system's diagnostic rules and process optimization strategies are continuously optimized to improve the system's adaptability and accuracy.

Images