Photovoltaic solar cell and module production and test system based on AI large model

Abstract

The invention provides a photovoltaic solar cell and module production and test system based on an AI large model, and belongs to the technical field of new energy, the photovoltaic solar cell and module production and test system based on the AI large model comprises a four-layer structure: bottom layer equipment is connected with an upper computer through a PLC language, the signal transmission accuracy is not less than 98%, the instruction response time is not more than 0.8 second, and the bottom layer equipment is connected with the upper computer through a PLC language; the number of signal transmission errors within 48 hours of continuous operation is not more than 3, and the number of signal transmission errors within 72 hours is not more than 2; the test equipment sends data and pictures to the MES system through WID signals, the data sending delay time does not exceed 0.4 second, the picture resolution is not lower than 1920 * 1080 pixels, the data sending success rate within one hour is not lower than 99%, and the data sending success rate within 30 minutes is not lower than 99.5%; the system can improve the production efficiency and the product quality, reduces the downtime of equipment and the labor cost, and has strong adaptability and continuous optimization capability.

Description

Technical Field

[0001] This invention belongs to the field of new energy technology, specifically relating to a production and testing system for photovoltaic solar cells and modules based on AI large-scale models. Background Technology

[0002] In the photovoltaic new energy field, as the market for photovoltaic solar power generation continues to expand and the industry accelerates production expansion, a large amount of new capacity is being released. Although existing machines have MES tracking systems, they basically only provide data feedback and cannot automatically process or perform simple analysis of abnormal data. When a certain standard is reached before shutting down, it is often too late. Moreover, human processing is slow, lacks periodicity, and cannot accurately control the corresponding area, affecting the efficiency of machine operation and resulting in poor analysis and processing of anomalies. Although it is a standardized operation, there are still problems such as differences in human techniques.

[0003] Therefore, based on the aforementioned technical problems, it is necessary to design an AI-based large-scale model-based photovoltaic solar cell and module production and testing system. This invention achieves the testing of photovoltaic solar cells and systems through a four-layer large-scale model. The first layer connects to the host computer via the PLC language of the underlying equipment to transmit signals and instructions; the second layer of the large-scale model sends the tested data and images to the MES via WID signals through the testing equipment; the third layer embeds the AI ​​large-scale model into the machine's cloud platform, providing accurate analysis solutions and reports to notify personnel of the next steps through machine learning, robot adaptive learning, and deep learning from historical data; the fourth layer integrates a Q&A assistant (chat-GPT, OPPO-AI) onto the server to answer on-site questions. This reduces equipment downtime and manual intervention, increases machine output, reduces manpower, and enables timely processing. Summary of the Invention

[0004] The purpose of this invention is to provide an AI-based large-scale model-based photovoltaic solar cell and module production and testing system, aiming to solve many problems existing in the current technology. On the one hand, this system can basically only provide data feedback and cannot automatically process or perform simple anomaly analysis. Often, by the time a certain standard is reached and the system is shut down, it is already too late. On the other hand, human intervention in handling anomalies is slow, lacks periodicity, and makes it difficult to accurately control the corresponding areas. This not only affects the efficiency of the equipment but also results in poor anomaly analysis and handling. Although the operation is standardized, differences in human technique still make it difficult to guarantee consistent processing results.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A photovoltaic solar cell and module production and testing system based on AI large-scale model, characterized by comprising the following structure:

[0007] The first layer of the structure connects to the host computer via a simple PLC language interface at the lower level for transmitting signals and instructions. The accuracy of signal transmission is no less than 98%, and the instruction response time is no more than 0.8 seconds. The number of signal transmission errors does not exceed 3 within 48 hours of continuous operation.

[0008] The second layer structure: The test equipment sends the test data and images to the MES system via WID signal. The data transmission delay time is no more than 0.4 seconds, and the image resolution is no less than 1920×1080 pixels. The data transmission success rate is no less than 99% within 1 hour.

[0009] The third layer of the structure is a database based on the MES system as one of the simplest data sources. It embeds large AI models into the cloud of the machine and provides accurate analysis solutions and reports to notify personnel of the next steps through machine learning, robot adaptive learning and deep learning of historical data. The accuracy of the analysis solution is no less than 92%, and the report generation time is no more than 1.5 minutes. The accuracy of the analysis of specific complex anomalies is no less than 85%.

[0010] The fourth layer structure integrates a Q&A assistant (chat-GPT, OPPO-AI) onto the server to answer on-site questions, thereby reducing equipment downtime and manual intervention, increasing machine output, reducing manpower while enabling timely processing, with a response time of no more than 25 seconds; and reducing equipment downtime by no less than 40%.

[0011] As a preferred embodiment of the present invention, the underlying device PLC language can stably and accurately connect to the host computer for signal and instruction transmission, and the number of signal transmission errors does not exceed 2 within 72 hours of continuous operation.

[0012] As a preferred embodiment of the present invention, the testing equipment can efficiently and accurately send the test data and images to the MES system via WID signal, with a data transmission success rate of no less than 99.5% within 30 minutes.

[0013] As a preferred embodiment of the present invention, the AI ​​large model in the third layer structure can quickly and accurately provide on-site analysis solutions and reports through machine learning, robot adaptive learning and historical data deep learning, with an analysis accuracy of no less than 95% for specific minor anomalies.

[0014] As a preferred embodiment of the present invention, the question-and-answer assistant (chat-GPT, OPPO-AI) in the fourth layer structure can accurately and promptly answer on-site questions, effectively reducing equipment downtime and manual actions, with a reduction in equipment downtime of no less than 45%.

[0015] As a preferred embodiment of the present invention, in S3, the accuracy of the key performance parameter data sent by the test equipment is not less than ±0.5%.

[0016] As a preferred embodiment of the present invention, the AI ​​large model in the third layer structure achieves an accuracy rate of no less than 91% when processing more than 10,000 historical data entries.

[0017] As a preferred embodiment of the present invention, the question-answering assistant (chat-GPT, OPPO-AI) in the fourth layer structure can handle more than 5 on-site questions at the same time, with a response time of no more than 30 seconds and an accuracy rate of no less than 90%.

[0018] As a preferred embodiment of the present invention, the connection stability between the first layer structure and the host computer is such that the number of interruptions does not exceed 2 within one week of continuous operation.

[0019] As a preferred embodiment of the present invention, the data transmission delay of the test equipment of the second layer structure does not vary by more than ±0.1 seconds under different ambient temperatures (0℃-40℃), the AI ​​big model of the third layer structure has an analysis scheme universality of no less than 80% for different types of batteries and components, and the question-answering assistant (chat-GPT, OPPO-AI) of the fourth layer structure improves the accuracy of answering questions by no less than 5% after upgrades and maintenance.

[0020] Compared with the prior art, the beneficial effects of the present invention are:

[0021] 1. This solution improves the stability and accuracy of signal transmission: By connecting the PLC language of the underlying device to the host computer, the signal transmission accuracy rate is not less than 98%, the instruction response time is not more than 0.8 seconds, the number of signal transmission errors does not exceed 3 times in 48 hours of continuous operation, and the number of signal transmission errors does not exceed 2 times in 72 hours of continuous operation, thus ensuring the stability and accuracy of signal and instruction transmission during the production process and providing a guarantee for the smooth operation of production.

[0022] Ensuring efficient and accurate data and image transmission: The testing equipment sends the tested data and images to the MES system via WID signals. The data transmission delay is no more than 0.4 seconds, the image resolution is no less than 1920×1080 pixels, the data transmission success rate is no less than 99% within 1 hour, and no less than 99.5% within 30 minutes. The accuracy of key performance parameters is no less than ±0.5%. Simultaneously, under different ambient temperatures (0℃-40℃), the data transmission delay varies by no more than ±0.1 seconds, guaranteeing the efficiency, accuracy, and stability of data and image transmission, providing reliable information support for production process monitoring and management.

[0023] Providing precise analysis solutions and reports: Using the MES system as a basic and simple data source, a database is built into the cloud-based machine tool. Through machine learning, robot adaptive learning, and deep learning from historical data, it provides precise on-site analysis solutions and reports to inform personnel of the next steps. The accuracy rate of the analysis solutions is no less than 92%, the accuracy rate for specific complex anomalies is no less than 85%, and the accuracy rate for specific minor anomalies is no less than 95%. Report generation time is no more than 1.5 minutes. When processing more than 10,000 historical data entries, the accuracy rate of the analysis solutions is no less than 91%, and the general applicability of the AI ​​model to analysis solutions for different types of batteries and components is no less than 80%. This enables the system to quickly and accurately identify various anomalies in the production process and provide effective solutions, improving production efficiency and product quality.

[0024] 2. This solution reduces equipment downtime and manual intervention: A Q&A assistant (chat-GPT, OPPO-AI) is integrated into the server to answer on-site questions. The response time for each question is no more than 25 seconds, and when handling more than 5 on-site questions simultaneously, the response time is no more than 30 seconds, with an accuracy rate of no less than 90%. Equipment downtime is reduced by no less than 40% and no less than 45%. This effectively reduces equipment downtime and manual intervention, increases machine output, reduces labor costs, and improves production efficiency.

[0025] Enhancing system adaptability and continuous optimization capabilities: The AI ​​large-scale model demonstrates strong adaptability with at least 80% universality in its analysis solutions for different types of batteries and components. After upgrades and maintenance, the accuracy of the Q&A assistant (chat-GPT, OPPO-AI) improved by at least 5%, showcasing the system's continuous optimization capabilities. This allows the system to better adapt to different production environments and needs, providing enterprises with higher-quality services. Attached Figure Description

[0026] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0027] Figure 1 This is a flowchart illustrating the process distribution of the AI-based large-scale model photovoltaic solar cell and module production and testing system of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Example 1

[0030] Please see Figure 1 The present invention provides the following technical solutions:

[0031] A photovoltaic solar cell and module production and testing system based on AI large-scale model, characterized by comprising the following structure:

[0032] The first layer of the structure connects to the host computer via a simple PLC language interface at the lower level for transmitting signals and instructions. The accuracy of signal transmission is no less than 98%, and the instruction response time is no more than 0.8 seconds. The number of signal transmission errors does not exceed 3 within 48 hours of continuous operation.

[0033] The second layer structure: The test equipment sends the test data and images to the MES system via WID signal. The data transmission delay time is no more than 0.4 seconds, and the image resolution is no less than 1920×1080 pixels. The data transmission success rate is no less than 99% within 1 hour.

[0034] The third layer of the structure is a database based on the MES system as one of the simplest data sources. It embeds large AI models into the cloud of the machine and provides accurate analysis solutions and reports to notify personnel of the next steps through machine learning, robot adaptive learning and deep learning of historical data. The accuracy of the analysis solution is no less than 92%, and the report generation time is no more than 1.5 minutes. The accuracy of the analysis of specific complex anomalies is no less than 85%.

[0035] The fourth layer structure integrates a Q&A assistant (chat-GPT, OPPO-AI) onto the server to answer on-site questions, thereby reducing equipment downtime and manual intervention, increasing machine output, reducing manpower while enabling timely processing, with a response time of no more than 25 seconds; and reducing equipment downtime by no less than 40%.

[0036] In a specific embodiment of the present invention, the first layer structure is as follows: In this layer structure, the bottom-level devices are connected to the host computer via PLC language, and the signal transmission accuracy is set to no less than 98%. This parameter ensures high reliability of data interaction. For example, in every 100 signal transmissions, the number of errors is at most 2. The instruction response time is strictly controlled to no more than 0.8 seconds, like a responsive assistant. When control instructions are issued during production, the system can respond in a very short time. Assuming a production scenario where equipment parameters need to be frequently adjusted, this fast instruction response time can greatly improve production efficiency. Furthermore, during a relatively long period of continuous operation (48 hours), the number of signal transmission errors does not exceed 3, fully demonstrating the stability of the system. Even during long-term continuous operation, good signal transmission quality can be maintained, providing a solid guarantee for the smooth operation of production.

[0037] Second-layer structure: The testing equipment sends data and images to the MES system via WID signals. The data transmission delay is no more than 0.4 seconds, meaning information can be transmitted at extremely high speeds, ensuring data timeliness. For example, when monitoring products in real time, the test results can be quickly fed back to the system for timely adjustments. The image resolution is no less than 1920×1080 pixels; high-resolution images provide clearer and more detailed product information, helping to accurately judge product quality. Within one hour, the data transmission success rate is no less than 99%, guaranteeing the reliability of data transmission. It's conceivable that almost all data can be successfully transmitted within one hour, greatly reducing the risk of data loss and providing an accurate data foundation for subsequent analysis and decision-making.

[0038] Third-layer structure: A database based on the MES system as one of the fundamental data sources, combined with AI large-scale models embedded in the machine cloud. The accuracy rate of the analysis solutions is no less than 92%, meaning that in most cases, the analysis solutions provided by the system are accurate and reliable. For example, when problems occur during production, the system can provide solutions with high accuracy, guiding on-site personnel to take the next step. Report generation time is no more than 1.5 minutes, ensuring that on-site personnel can understand the production situation in a timely manner. In a fast-paced production environment, quickly generated reports allow staff to make rapid decisions, improving production efficiency. The accuracy rate of analysis for specific complex anomalies is no less than 85%, indicating that the system also has strong analytical capabilities when facing complex problems. Even for those complex anomalies that are difficult to judge, the system can analyze them with high accuracy, providing strong support for problem-solving.

[0039] Fourth-layer structure: The Q&A assistants (chat-GPT, OPPO-AI) are integrated into the server, ensuring a response time of no more than 25 seconds. In actual production, when on-site personnel encounter problems, they can receive rapid answers, avoiding production stoppages caused by unresolved issues. Equipment downtime is reduced by at least 40%, demonstrating the system's significant effect on improving equipment availability. By reducing downtime, not only is production efficiency improved, but production costs are also reduced. Simultaneously, reduced manpower and timely problem-solving further enhance the company's competitiveness, giving it an advantage in the fierce market competition.

[0040] Please refer to the details. Figure 1 The underlying device's PLC language can stably and accurately connect to the host computer for signal and instruction transmission, and the number of signal transmission errors does not exceed 2 within 72 hours of continuous operation.

[0041] In this embodiment, the PLC language used by the underlying device demonstrates excellent performance in transmitting signals and instructions to the host computer. Firstly, its stability is fully guaranteed, meaning that frequent connection interruptions or instability will not occur during long-term operation. This connection remains reliable regardless of complex production environments or various external interferences.

[0042] Accurate signal and command transmission is crucial for smooth production. It ensures that commands issued by the host computer are accurately received and executed by the underlying devices, while signals fed back from the underlying devices are precisely transmitted to the host computer for analysis and processing.

[0043] The requirement of no more than two signal transmission errors within 72 hours of continuous operation is a very stringent metric. It reflects the system's high reliability and robustness. Such a low error rate over such a long period significantly reduces production interruptions and quality issues caused by signal transmission errors.

[0044] For example, on a continuously operating photovoltaic solar cell production line, such a stable and accurate connection ensures coordinated operation of all production stages. Even under high production pressure, accurate signal transmission is guaranteed, thereby improving production efficiency and product quality. At the same time, the low error rate reduces maintenance costs and manpower input, bringing greater economic benefits to the enterprise.

[0045] Please refer to the details. Figure 1 The testing equipment can efficiently and accurately send the test data and images to the MES system via WID signal, with a data transmission success rate of no less than 99.5% within 30 minutes.

[0046] In this embodiment, the testing equipment plays a crucial role in data and image transmission. It utilizes WID signals to efficiently and accurately transmit post-test information to the MES system. This efficiency is reflected in the rapid data transmission speed, ensuring that test results can be promptly utilized by subsequent stages, avoiding production delays or decision-making lags caused by transmission latency.

[0047] Accurate transmission of data and images means that every data point and every image is transmitted completely and faithfully to the MES system, without data loss, corruption, or blurry images. This is crucial for production processes that rely on precise data and clear images for analysis and decision-making.

[0048] A data transmission success rate of no less than 99.5% within a 30-minute timeframe is an extremely high standard. This indicates that almost all test data and images can be successfully transmitted within a short period. For example, in a high-pressure production environment, multiple tests and data transmissions may occur every 30 minutes; such a high success rate ensures production continuity and stability. Even under heavy workloads, reliable data transmission is guaranteed, providing a solid foundation for production management and quality control. Simultaneously, the high success rate reduces repetitive testing and repair work required due to data transmission failures, improving production efficiency and reducing costs.

[0049] Please refer to the details. Figure 1 The AI ​​big model in the third layer structure can quickly and accurately provide on-site analysis solutions and reports through machine learning, robot adaptive learning and historical data deep learning, with an accuracy rate of no less than 95% for analyzing specific minor anomalies.

[0050] In this embodiment: In the third-layer structure, the large AI model demonstrates powerful analytical capabilities. Machine learning, robot adaptive learning, and deep learning from historical data work together to provide strong support for the generation of on-site analysis plans and reports.

[0051] Machine learning enables large AI models to automatically learn patterns and rules from massive amounts of data, continuously optimizing their analytical capabilities. As data accumulates and learning continues, it becomes able to more accurately understand various situations in the production process, providing more targeted analytical solutions for on-site operations.

[0052] Robot adaptive learning enables the system to automatically adjust its analysis strategies based on different production environments and task requirements. For example, when new product types or process changes occur on the production line, the robot can quickly adapt and adjust its analysis methods to ensure the accuracy of the analysis results.

[0053] Historical data deep learning makes full use of past production data to uncover potential information. Through in-depth analysis of historical data, large AI models can discover hidden trends and problems, providing early warnings and solutions for on-site operations.

[0054] A minimum accuracy rate of 95% in analyzing specific minor anomalies demonstrates the exceptional ability of large-scale AI models to handle subtle issues. In actual production, minor anomalies are often easily overlooked, but if not addressed promptly, they can gradually escalate into serious problems. Large-scale AI models can identify these minor anomalies with high accuracy, providing timely alerts to on-site personnel and enabling them to take effective intervention measures before the problem escalates.

[0055] For example, in the production process of photovoltaic solar cells, AI-powered big data models can quickly detect subtle anomalies, such as voltage fluctuations and temperature increases, through real-time analysis of equipment operating parameters and product quality data. Based on these anomalies, the models can rapidly provide accurate analysis solutions and reports, guiding on-site personnel to make adjustments and repairs, thereby ensuring stable production and reliable product quality.

[0056] Please refer to the details. Figure 1 The question-and-answer assistant (chat-GPT, OPPO-AI) in the fourth layer structure can accurately and promptly answer on-site questions, effectively reducing equipment downtime and manual actions, with a reduction of equipment downtime of no less than 45%.

[0057] In this embodiment, the Q&A assistant (chat-GPT, OPPO-AI) plays a crucial role in the fourth-layer structure. Leveraging its powerful language understanding and question-answering capabilities, it accurately comprehends questions posed by on-site personnel. Whether the questions concern equipment operation, troubleshooting, or production process optimization, the Q&A assistant (chat-GPT, OPPO-AI) can quickly provide clear and accurate answers.

[0058] Its timeliness is also crucial. On the production floor, time is money. When problems arise, on-site personnel need quick solutions to avoid production stoppages. The Q&A assistants (chat-GPT, OPPO-AI) can respond to questions in a very short time, providing timely support to on-site personnel. This not only improves production efficiency but also reduces potential losses caused by unresolved issues.

[0059] By accurately and promptly answering on-site questions, the Q&A assistants (chat-GPT, OPPO-AI) effectively reduce equipment downtime and manual intervention. Equipment downtime often leads to significant production losses, including decreased output and increased costs. The Q&A assistants (chat-GPT, OPPO-AI) can help on-site personnel quickly diagnose problems and provide solutions, thereby shortening equipment downtime.

[0060] A reduction of at least 45% in equipment downtime is a very significant figure. For example, in the past, equipment might frequently stop due to various problems, each downtime causing production interruptions. However, with the help of the Q&A assistants (chat-GPT, OPPO-AI), downtime is drastically reduced. Assuming that the original weekly equipment downtime was 10 hours, a 45% reduction would lower it to below 5.5 hours. This will greatly improve equipment utilization and production efficiency, bringing significant economic benefits to the company.

[0061] At the same time, reducing manual actions also means lower labor costs and the risk of human error. On-site personnel no longer need to spend a lot of time searching for information, consulting experts, or performing trial and error operations. Instead, they can obtain accurate guidance directly from the Q&A assistants (chat-GPT, OPPO-AI), thereby completing work tasks more efficiently.

[0062] Please refer to the details. Figure 1 S3. The accuracy of the key performance parameter data sent by the test equipment shall not be less than ±0.5%.

[0063] In this embodiment, the data transmitted by the testing equipment in this system is of high quality. Specifically, the accuracy of the key performance parameter data is no less than ±0.5%, reflecting the accuracy and reliability of the data.

[0064] Such high precision means that the error range is strictly controlled within a very small range when measuring and recording key performance parameters. For example, for a specific key performance parameter, if the actual value is 100, then the measured value will be between 99.5 and 100.5. This high-precision data is crucial for accurately evaluating the performance of photovoltaic solar cells and modules.

[0065] It can provide a reliable basis for subsequent analysis and decision-making. During the production process, accurate key performance parameter data can help engineers identify problems in a timely manner, adjust production processes, and improve product quality. At the same time, for quality control departments, high-precision data can ensure that products meet relevant standards and requirements.

[0066] Furthermore, high-precision data can help optimize the entire production system. By accurately monitoring and analyzing key performance parameters, bottlenecks and areas for optimization in the production process can be identified, thereby improving production efficiency and reducing costs.

[0067] For example, in the production of photovoltaic solar cells, the accuracy of key performance parameters such as voltage, current, and power directly affects the product's performance and quality. Low measurement accuracy of these parameters can lead to incorrect judgments and decisions, impacting the product's market competitiveness. Therefore, the high accuracy of key performance parameter data transmitted by testing equipment is a crucial factor in ensuring the stable operation of the entire production system and the reliability of product quality.

[0068] Please refer to the details. Figure 1 When the AI ​​large model in the third layer structure processes more than 10,000 historical data, the accuracy of the analysis scheme is no less than 91%.

[0069] In this embodiment, the AI ​​large model demonstrates powerful data analysis capabilities in the third-layer structure. When processing more than 10,000 historical data entries, the accuracy of its analysis scheme remains at a high level of no less than 91%, which fully reflects the reliability and efficiency of the model.

[0070] This vast amount of historical data provides rich information resources for large-scale AI models. Through in-depth mining and analysis of this data, the models can learn production patterns, problem characteristics, and solutions under various circumstances. Even when faced with complex and ever-changing production environments and problems, they can provide accurate analytical solutions based on their rich experience.

[0071] When the amount of historical data reaches this scale, large AI models can perform more comprehensive pattern recognition and trend prediction. For example, various types of anomalies may occur during the production of photovoltaic solar cells and modules. By analyzing a large amount of historical data, the model can identify the patterns and characteristics of these anomalies, thereby enabling it to discover problems more quickly and provide accurate solutions in new production processes.

[0072] Meanwhile, highly accurate analytical solutions provide strong support for production decisions. Enterprises can adjust their production strategies, optimize processes, and improve product quality and production efficiency in a timely manner based on the analytical solutions provided by the AI ​​model. In a highly competitive market environment, this accurate analytical capability can help enterprises gain an advantage and achieve sustainable development.

[0073] For example, suppose a photovoltaic solar cell manufacturer generates a large amount of production data every day. Over time, as historical data accumulates, once it reaches over 10,000 records, a large AI model can accurately analyze the current production status, predict potential problems, and provide corresponding solutions. This will significantly improve the company's production management level and market competitiveness.

[0074] Please refer to the details. Figure 1 The question-and-answer assistant (chat-GPT, OPPO-AI) in the fourth layer structure can handle more than 5 on-site questions at the same time, with a response time of no more than 30 seconds and an accuracy rate of no less than 90%.

[0075] In this embodiment, the question-answering assistant (chat-GPT, OPPO-AI) demonstrates excellent multitasking capabilities in the fourth layer structure. Even when faced with more than five on-site questions simultaneously, it maintains a high response speed and accuracy.

[0076] For scenarios involving multiple questions simultaneously, the Q&A assistant (chat-GPT, OPPO-AI) needs powerful computing capabilities and rapid logical reasoning. It must analyze and understand multiple questions quickly and provide accurate answers. With a response time of no more than 30 seconds, this means that in actual production, even if multiple urgent issues arise on-site, the Q&A assistant (chat-GPT, OPPO-AI) can quickly provide solutions to staff, preventing the problems from escalating and production from halting.

[0077] An accuracy rate of at least 90% ensures the reliability of the answers. Maintaining such a high accuracy rate even when handling multiple questions indicates that the Q&A assistant (chat-GPT, OPPO-AI) has undergone thorough training and optimization, accurately understanding the core of the questions and providing appropriate solutions. For example, in photovoltaic solar cell and module production sites, multiple different types of problems may occur simultaneously, such as equipment malfunctions, process parameter adjustments, and quality inspections. The Q&A assistant (chat-GPT, OPPO-AI) can accurately determine the nature of each problem and provide targeted solutions, providing strong support for smooth production. This Q&A assistant combines Chat-GPT and OPPO-AI, achieving higher response efficiency than either Chat-GPT or OPPO-AI alone. Furthermore, it can answer questions that Chat-GPT and OPPO-AI cannot.

[0078] Furthermore, this highly efficient multitasking capability saves businesses significant time and labor costs. Staff can obtain answers to problems without long waiting times, enabling them to take swift action and improve productivity. Simultaneously, it reduces reliance on specialized technical personnel, lowering overall operating costs.

[0079] Imagine a busy production workshop where multiple employees simultaneously ask different questions. The chat-GPT or OPPO-AI assistant can respond quickly, providing accurate answers within 30 seconds, helping employees resolve issues and ensuring continuous production. This efficient service will significantly enhance a company's competitiveness and production efficiency.

[0080] Please refer to the details. Figure 1 The connection stability between the first layer structure and the host computer is such that the number of interruptions does not exceed 2 within one week of continuous operation.

[0081] In this embodiment: In the first-layer structure, the connection stability between the lower-level devices and the host computer exhibits excellent performance. A continuous operation spanning one week is a challenging test for the production system, and the number of interruptions during this period did not exceed two, fully demonstrating the reliability of the connection.

[0082] This highly stable connection provides a solid foundation for continuous production. During extended week-long operations, whether facing complex production tasks or potential external interference, the connection remains stable, ensuring smooth transmission of signals and commands.

[0083] For example, in the production process of photovoltaic solar cells and modules, continuous data monitoring and control commands are required. Unstable connections and frequent interruptions can lead to production disruptions, data loss, and other problems, severely impacting production efficiency and product quality. The regulation stipulating no more than two interruptions per week ensures smoother production processes and reduces the risks and losses caused by connectivity issues.

[0084] This also reflects the rigor in the design and implementation of the system. Achieving such high connection stability necessitates careful consideration and optimization in hardware selection, software system optimization, and communication protocol design. By employing advanced technology and reliable equipment, the system ensures stable connections between the underlying devices and the host computer over extended periods, providing strong support for the efficient operation of the production system.

[0085] Please refer to the details. Figure 1The data transmission delay of the second-layer test equipment does not vary by more than ±0.1 seconds under different ambient temperatures (0℃-40℃). The AI ​​big model of the third layer has an analysis scheme universality of no less than 80% for different types of batteries and components. After upgrades and maintenance, the accuracy of the question-answering assistant (chat-GPT, OPPO-AI) of the fourth layer improves by no less than 5%.

[0086] In this embodiment, for the second-layer structure, the test equipment exhibited highly stable data transmission performance under different ambient temperatures ranging from 0℃ to 40℃. The data transmission delay time variation did not exceed ±0.1 seconds, meaning that the test equipment could maintain a relatively consistent data transmission speed regardless of whether it was in a relatively cold environment of 0℃ or a relatively hot environment of 40℃. In practical applications, the production environment of photovoltaic solar cells and modules may be affected by factors such as seasonal changes and indoor temperature regulation. This data transmission delay characteristic, which is insensitive to temperature changes, ensures that data can be transmitted to the MES system in a timely and accurate manner under various actual operating conditions, providing reliable information support for the monitoring and management of the production process.

[0087] The AI-powered large-scale model in the third layer of the structure demonstrates strong adaptability, with its analysis solutions for different types of cells and modules exhibiting at least 80% general applicability. Different types of photovoltaic solar cells and modules may differ in structure and performance characteristics, but the AI-powered large-scale model can overcome these differences to a certain extent, providing effective analysis solutions for various product types. This indicates that the model has undergone thorough training and optimization, enabling it to extract common features and patterns from large amounts of data, thus allowing its application in different product analyses. For example, in a company that simultaneously produces multiple specifications of photovoltaic products, the AI-powered large-scale model can provide highly universal analysis solutions for different production lines, reducing the workload of individual analysis and adjustments for different products, and improving production efficiency and management convenience.

[0088] After upgrades and maintenance, the accuracy of the fourth-layer question-answering assistants (chat-GPT, OPPO-AI) improved by at least 5%, demonstrating the system's continuous optimization capabilities. With ongoing technological advancements and changing production demands, upgrading and maintaining the question-answering assistants (chat-GPT, OPPO-AI) is essential. Such upgrades not only fix potential vulnerabilities and issues but also enhance performance by introducing new algorithms and increasing data volume. An accuracy improvement of at least 5% means that after the upgrade, the assistants can answer on-site questions more accurately, providing more reliable support to staff. For example, after a major upgrade, the accuracy of the assistants in answering some complex technical questions significantly improved, helping companies resolve production challenges more quickly and further improving production efficiency and quality.

[0089] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A photovoltaic solar cell and module production and testing system based on AI large-scale model, characterized in that, Includes the following structures: The first layer of the structure connects to the host computer via a simple PLC language interface at the lower level for transmitting signals and instructions. The accuracy of signal transmission is no less than 98%, and the instruction response time is no more than 0.8 seconds. The number of signal transmission errors does not exceed 3 within 48 hours of continuous operation. The second layer structure: The test equipment sends the test data and images to the MES system via WID signal. The data transmission delay time is no more than 0.4 seconds, and the image resolution is no less than 1920×1080 pixels. The data transmission success rate is no less than 99% within 1 hour. The third layer of the structure is a database based on the MES system as one of the simplest data sources. It embeds large AI models into the cloud of the machine and provides accurate analysis solutions and reports to notify personnel of the next steps through machine learning, robot adaptive learning and deep learning of historical data. The accuracy of the analysis solution is no less than 92%, and the report generation time is no more than 1.5 minutes. The accuracy of the analysis of specific complex anomalies is no less than 85%. The fourth layer structure integrates a Q&A assistant (chat-GPT, OPPO-AI) onto the server to answer on-site questions, thereby reducing equipment downtime and manual intervention, increasing machine output, reducing manpower while enabling timely processing, with a response time of no more than 25 seconds; and reducing equipment downtime by no less than 40%.

2. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 1, characterized in that: The underlying device's PLC language can stably and accurately connect to the host computer for signal and instruction transmission, with no more than 2 signal transmission errors occurring within 72 hours of continuous operation.

3. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 2, characterized in that: The testing equipment can efficiently and accurately send the test data and images to the MES system via WID signal, with a data transmission success rate of no less than 99.5% within 30 minutes.

4. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 3, characterized in that: The AI ​​big model in the third layer structure can quickly and accurately provide on-site analysis solutions and reports through machine learning, robot adaptive learning and deep learning of historical data, with an accuracy rate of no less than 95% for analyzing specific minor anomalies.

5. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 4, characterized in that: The question-and-answer assistant (chat-GPT, OPPO-AI) in the fourth layer structure can accurately and promptly answer on-site questions, effectively reducing equipment downtime and manual actions, with a reduction in equipment downtime of no less than 45%.

6. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 5, characterized in that: S3. The accuracy of the key performance parameter data sent by the test equipment shall not be less than ±0.5%.

7. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 6, characterized in that: When the AI ​​large model in the third layer structure processes more than 10,000 historical data entries, the accuracy of the analysis scheme is no less than 91%.

8. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 6, characterized in that: When the question-and-answer assistant (chat-GPT, OPPO-AI) in the fourth layer structure handles more than 5 on-site questions at the same time, the response time for answering questions shall not exceed 30 seconds and the accuracy rate shall not be less than 90%.

9. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 6, characterized in that: The connection stability between the first layer structure and the host computer is such that the number of interruptions does not exceed 2 within one week of continuous operation.

10. The AI-based large-scale model photovoltaic solar cell and module production and testing system according to claim 6, characterized in that: The data transmission delay of the second-layer test equipment varies by no more than ±0.1 seconds under different ambient temperatures (0℃-40℃). The AI ​​big model of the third layer has an analysis solution universality of no less than 80% for different types of batteries and components. After upgrades and maintenance, the accuracy of the question-answering assistant (chat-GPT, OPPO-AI) of the fourth layer improves by no less than 5%.

Images