AI artificial intelligence intelligent tin smelting control system and method

Abstract

The invention relates to the technical field of tin smelting, and discloses an AI intelligent tin smelting control system and method, and the system comprises a data collection and preprocessing module, an intelligent decision and control module, and a real-time monitoring and early warning module. The method comprises the following steps: collecting process data in a tin smelting production process in real time, and cleaning the collected process data; extracting and selecting characteristics of which the influence on the smelting quality and efficiency exceeds a preset threshold value, and carrying out standardization processing on the data; receiving the standardized process data, training the model, processing the data monitored in real time, and outputting the adjusted process parameters; various key parameters and data in the production process are displayed in real time, the real-time data are analyzed, abnormal conditions are detected, early warning is carried out on faults, and response measures are automatically triggered. Intelligent and automatic tin smelting is achieved, and the efficiency, quality and safety of the smelting process can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tin smelting, and in particular to an AI artificial intelligence intelligent tin smelting control system and method. Background Art

[0002] Tin smelting is the process of converting tin ore and other raw materials into high-purity tin through a series of chemical, physical reactions and mechanical processes. The traditional tin smelting process mainly includes roasting, reduction, refining and electrolysis. Although these steps have basically solved the problem of tin extraction, they still face some challenges in the production process; first, the grade of tin ore is unstable, and the composition of different batches of raw materials varies greatly, which makes the smelting process and effect difficult to predict and control; secondly, traditional tin smelting methods have high energy consumption and high pollution. The large amount of waste gas, waste slag and waste liquid generated during the smelting process has caused serious pollution to the environment; in addition, traditional tin smelting relies on the experience and manual control of operators, the process is unstable, the product quality is uneven, and it is also prone to operational errors, causing safety accidents; with the rapid development of artificial intelligence technology, especially deep learning and reinforcement learning, it has become possible to introduce AI into the intelligent control of the tin smelting process; AI technology can solve many shortcomings of traditional control systems. By introducing By introducing reinforcement learning algorithms and using AI to analyze and mine large amounts of historical and real-time data, the control strategy can be continuously optimized through machine learning models, the state changes in the smelting process can be perceived in real time, and the control parameters can be dynamically adjusted, thereby improving smelting efficiency and stability and achieving optimal production levels. Therefore, the AI-based smart tin smelting control system can break through the various limitations of traditional control systems, significantly improve the intelligence level of the smelting process, and has broad application prospects and great significance. At present, parameter adjustments in the production process mostly rely on experienced operators, and people's subjective judgments inevitably bring volatility, making it difficult to ensure product consistency and stability. Traditional smelting processes have high energy consumption and emissions, causing serious pollution to the environment, and have failed to effectively adopt energy-saving and consumption-reduction and pollution control technologies.

[0003] Prior art 1, Chinese patent, application number: 20221049039.3 discloses an intelligent crude tin smelting system and method, which monitors different aspects of crude tin smelting in the early, middle and late stages, collects and statistics data, and extracts and marks the collected data, so as to jointly calculate the data of different periods to obtain the evaluation value of the corresponding period. Based on the evaluation value, the state of crude tin smelting in the corresponding period can be analyzed and evaluated, and different methods of early warning and prompting can be performed, which can realize comprehensive monitoring and early warning analysis of crude tin smelting, so that crude tin smelting can be operated intelligently and efficiently, reducing manual intervention inspection and adjustment, so as to achieve the purpose of improving the effect of crude tin smelting. Although the existing scheme solves the technical problem that crude tin smelting cannot be intelligently and efficiently monitored and prompted, while taking into account the reasonable use of data processing resources to avoid high-load operation. However, the collected data is not fully utilized for optimization, the value of the data is underestimated, and the utilization rate of the data is low.

[0004] Prior art 2, Chinese patent, application number 202111130720.6 discloses an intelligent crude tin smelting system and method, which includes: an operating management layer for receiving an operating plan and sending the operating plan data to an execution manufacturing layer; the execution manufacturing layer includes: an MES management system, a smelting expert system, and a spray gun expert system; the control layer is used to receive real-time data to control the boiler equipment in the smelting process in real time; the equipment layer collects the real-time data of the boiler equipment in the smelting process and sends it to the control layer and the execution manufacturing layer. The method includes: real-time collection of real-time data of boiler equipment in the smelting process; calculation and analysis of operating plan data, production management data, and real-time data, so as to realize intelligent analysis and intelligent adjustment of the main operating parameters, automatic optimization of operating parameters, and automatic correction, and timely alarm in case of abnormalities. Although it solves the problems of manual intervention, independent control, and lack of unified management and control in each process in the original production process. However, it lacks real-time monitoring and fault warning functions, and is prone to safety accidents due to improper parameter settings or equipment failures.

[0005] Prior art three, Chinese patent, application number 202111130677.3 discloses an intelligent crude tin smelting system, including: planning management layer, execution management layer, data analysis and evaluation layer, DCS system management layer, DCS system control layer and field equipment layer; the planning management layer adopts the ERP system, and exchanges data with the execution management layer by establishing a private enterprise network; the execution management layer integrates all kinds of data of the entire crude tin production for use by the data analysis and evaluation layer; the data analysis and evaluation layer analyzes and evaluates the data of each layer; the DCS system management layer writes the production process control algorithm into the DCS controller through the engineer station and the DCS configuration server; the data of the DCS system control layer is transmitted to the DCS system management layer through the industrial Ethernet and the switch for data display; the field equipment layer transmits the collected data to the DCS system control layer. Although it can reduce the manual participation in each link of tin smelting, change the traditional semi-automatic smelting control status, and realize the unified coordination and control of each production management link. However, it is difficult to dynamically adjust the parameters and cannot quickly respond to the impact of raw material composition and environmental changes, resulting in a decrease in the quality of the smelted finished tin.

[0006] At present, the existing technologies 1, 2 and 3 fail to make full use of the collected data for optimization, and the value of the data is underestimated; lack of real-time monitoring and fault warning functions, easily causing safety accidents due to improper parameter settings or equipment failures; difficult to dynamically adjust parameters, and unable to quickly respond to the impact of raw material composition and environmental changes. Therefore, the present invention provides an AI artificial intelligence intelligent tin smelting control system. Summary of the invention

[0007] The main purpose of the present invention is to provide an AI artificial intelligence intelligent tin smelting control system and method to solve the problems in the prior art of failing to fully utilize the collected data for optimization, lacking real-time monitoring and fault warning functions, and difficulty in dynamically adjusting parameters.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] An AI artificial intelligence intelligent tin smelting control system, the AI ​​artificial intelligence intelligent tin smelting control system comprising:

[0010] The data acquisition and preprocessing module is used to collect process data in the tin smelting production process in real time and clean the collected process data; extract and select features that affect the smelting quality and efficiency beyond the preset threshold, and standardize the data; the process data includes: monitoring the smelting temperature, oxygen concentration in the furnace and tin flow; and detecting the reduction degree of the finished product at the terminal, and monitoring the temperature changes of the slag and flue gas in real time;

[0011] Intelligent decision-making and control module, which is used to receive standardized process data, train the model, process the real-time monitored data and output the adjusted process parameters, including gun position, compressed gas flow, oxygen concentration, feed amount of various materials, material ratio, furnace pressure and temperature;

[0012] The real-time monitoring and early warning module is used to display the key parameters and data of the production process in real time, analyze the real-time data, detect abnormal situations, warn of faults, and automatically trigger response measures.

[0013] As a further improvement of the present invention, the data acquisition and preprocessing module includes:

[0014] The sensor node submodule is used to obtain the target data of the process data acquisition equipment, match the target data with each node in the smelting equipment and production line, deploy the corresponding process data acquisition equipment according to the matching results, and obtain the process data of tin smelting through the process data acquisition equipment;

[0015] The data processing submodule is used to obtain the original process data collected by the sensor node submodule, clean the collected original process data, and fill in the data if there is a discontinuous phenomenon; extract and select target features, and standardize the collected process data;

[0016] The data cache submodule is used to set up a data cache area between the sensor node submodule and the data processing submodule, and is responsible for temporarily storing and buffering process data when the data processing submodule network is congested.

[0017] As a further improvement of the present invention, the data processing submodule includes:

[0018] The original process data processing unit performs standardization on the collected original process data to ensure that the mean of the data is 0 and the variance is 1. Based on the standardized data, the covariance matrix is ​​calculated;

[0019] Solve the eigenvector unit, solve the process eigenvalue and process eigenvector of the covariance matrix, sort the obtained process eigenvalues ​​in descending order, select the process eigenvectors corresponding to some process eigenvalues ​​as the main components, and form a new feature space;

[0020] The feature extraction unit forms a transformation matrix with the selected process feature vectors, multiplies the transformation matrix with the original data matrix, and projects it into a new feature space, thereby converting the original process data into a low-dimensional representation while retaining the main information of the data.

[0021] As a further improvement of the present invention, the intelligent decision-making and control module includes:

[0022] The reinforcement learning algorithm submodule is used to simulate an intelligent agent interacting with the environment and obtain observations and rewards from it; train the intelligent agent and make decisions; the intelligent agent takes actions such as adjusting the gun position and changing the feed amount in each state;

[0023] The training and strategy submodule is used to use the reward function to obtain positive rewards for correct actions and negative rewards for incorrect actions respectively; evaluate the value of the current state of the agent, obtain the relationship between the new action and the best action after sorting in the current strategy, and train the agent using a deep Q network;

[0024] The process parameter adjustment submodule is used to dynamically adjust the tin smelting process parameters, including gun position, compressed gas flow, oxygen concentration, feed amount of various materials, material ratio, furnace pressure and temperature, according to the real-time detected production status and the optimal action after sorting in the current strategy output by the reinforcement learning model.

[0025] As a further improvement of the present invention, the reinforcement learning algorithm submodule includes:

[0026] The data establishment unit is used to establish a process parameter table and initialize random numbers. The parameters in the table include the current temperature, pressure, gas concentration and other process parameter states of the smelting process, and actions and gains such as adjusting the gun position and changing the feed amount. The experience pool is established. During the training process, the intelligent agent collects experience through interaction with the environment and stores the collected experience in the experience pool.

[0027] The state update unit is used to update the state and gain of the process parameters. Based on the current process parameter table, the intelligent agent selects the action of adjusting the gun position and changing the feeding amount with the largest value in the process parameter table, executes and observes the new state and gain of the process parameters; and updates the process parameter table according to the new state and gain of the process parameters;

[0028] The stable training unit is used to stabilize the training process and use a fixed target network to calculate the target process parameter table values. It separates the process of selecting actions and calculating the target process parameter table values, and uses the back propagation algorithm to optimize the process parameters until the target network converges. The performance of the intelligent agent is regularly evaluated during the training process.

[0029] As a further improvement of the present invention, the training and strategy submodule includes:

[0030] Function building unit, which is used to design a compound reward function based on multi-objective optimization. The agent considers the simultaneous optimization of multiple objectives when optimizing the smelting process.

[0031] The dynamic adjustment unit is used to generate the action selection probability through the policy model, and adopts an adaptive learning rate to dynamically adjust the learning rate according to the learning performance of the intelligent agent;

[0032] The decision feedback unit is used to introduce a feedback mechanism at each time step. Based on the state monitoring data and environmental changes, it evaluates whether the process parameters under the current strategy are optimal and promptly discovers deviations from the reward function target. Based on historical decision performance feedback, self-supervised learning is introduced to allow the intelligent agent to continuously extract lessons from past experience.

[0033] As a further improvement of the present invention, the process parameter adjustment submodule includes:

[0034] The strategy network unit is used to select the action of adjusting the gun position and changing the feeding amount, and outputs an action of adjusting the gun position and changing the feeding amount based on the current state of multiple process parameters such as temperature, pressure and gas concentration;

[0035] The value network unit is used to evaluate the value of the current strategy. Based on the current state of various process parameters such as temperature, pressure, gas concentration, etc., the actions of adjusting the gun position and changing the feed amount, an estimated value or advantage function is output to guide the update direction of the strategy network subunit;

[0036] The optimal strategy unit is used to set a process probability value between (0,1) during the training process. At each decision moment, the action with the highest estimated value is selected by the complement of the process probability value, and an action is randomly selected by the process probability value. The relationship between exploring new actions and utilizing the optimal action after sorting in the current strategy is balanced; through continuous updating, the optimal strategy after sorting in the current strategy is gradually reached.

[0037] As a further improvement of the present invention, the real-time monitoring and early warning module includes:

[0038] A real-time data analysis unit is used to receive the current output action and various process parameter states, extract the target features, extract the key parameters from the target features, and display the key parameters;

[0039] The abnormality detection unit is used to detect the extracted current output target characteristics and key parameters, and compare them with the optimal process parameters after sorting in the current strategy in the optimal strategy process parameter table after sorting in the current strategy. If the current output target characteristics and key parameter values ​​are less than the optimal process parameter values ​​after sorting in the current strategy, an early warning signal is issued;

[0040] The early warning and response unit is used to automatically trigger the preset emergency response procedure after receiving the early warning signal, and remind relevant personnel through visual tools and alarm sounds.

[0041] As a further improvement of the present invention, the real-time data analysis unit comprises:

[0042] The target feature selection subunit is used to perform pre-processing operations such as cleaning and correction on the data. Based on the optimal process parameters after sorting in the current strategy in the process parameter table of the optimal strategy after sorting in the current strategy, the most relevant target features are selected from the original data to form an output feature set;

[0043] The association rule calculation subunit is used to scan the output feature set, calculate the support of each item, and obtain the candidate item set. If the calculated support of the candidate item set is greater than the preset threshold, it is marked as a frequent item set, and the frequent item set 1 is found; the candidate item set 2 is generated using the frequent item set 1, and the output feature set is scanned again, the support is calculated, and the frequent item set 2 is found; the above process is repeated until no new frequent item sets can be generated, and finally all the association rules between the key parameters and the target features that make the support meet the preset threshold are output;

[0044] The support calculation formula is:

[0045] Support = number of features in the item set / total number of features

[0046] The model training subunit is used to utilize the association rules, use the key parameters and target features whose support meets the preset threshold as the training set of the model to train the model, and evaluate the performance of the model through methods such as cross-validation. After the model is built and trained, the current output target features are input to visualize the output key parameters.

[0047] To achieve the above object, the present invention also provides the following technical solutions:

[0048] An AI artificial intelligence intelligent tin smelting control method comprises the following steps:

[0049] Collect process data from the tin smelting production process in real time and clean the collected process data; extract and select features that affect smelting quality and efficiency beyond the preset threshold, and standardize the data;

[0050] Receive standardized process data, train the model, process the real-time monitored data and output the adjusted process parameters, including gun position, compressed gas flow, oxygen concentration, feed amount of various materials, material ratio, furnace pressure and temperature;

[0051] Display key parameters and data of the production process in real time, analyze real-time data, detect abnormal situations, warn of faults, and automatically trigger response measures.

[0052] The data acquisition and preprocessing module of the present invention collects key process data (such as smelting temperature, oxygen concentration in the furnace, tin flow, slag and flue gas temperature, etc.) in the smelting process in real time to ensure the timeliness and accuracy of the data; the collected data is cleaned to remove noise and outliers, and important features related to smelting quality and efficiency are extracted, and converted into a standardized form for subsequent processing. Significance: Through cleaning and standardization, the data input to the intelligent decision-making module is ensured to be accurate and available, thereby improving the reliability of decision-making; extracting key features enables the decision-making system to more accurately understand and analyze changes in the smelting process, which helps to better predict and optimize the smelting effect. The intelligent decision-making and control module analyzes and trains the standardized process data, and uses machine learning or deep learning algorithms to form a prediction model; according to the data obtained by real-time monitoring, the adjusted process parameters (such as gun position, compressed gas flow, oxygen concentration, etc.) are output to realize the automatic control of the smelting process. Significance: Accurately adjusting the process parameters can improve the smelting efficiency and the yield of tin, while reducing the production of defective products; automated intelligent decision-making reduces the intervention of manual operation, reduces the risk of human error, and ensures the stability and consistency of production. The real-time monitoring and early warning module monitors and analyzes key parameters in real time. When abnormal conditions occur (such as excessive temperature, abnormal oxygen concentration, etc.), it automatically identifies and issues early warnings. When a fault or abnormal condition is detected, it can automatically trigger corresponding response measures (such as adjusting parameters or stopping production) to prevent accidents. Significance: The real-time monitoring and early warning system can promptly detect potential dangers in the production process, ensure the safety of operators and equipment, and prevent accidents. By timely adjusting and responding to abnormal conditions, the stability of the production line can be improved, downtime can be reduced, and overall production efficiency and economic benefits can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a functional module diagram of an embodiment of the AI ​​artificial intelligence intelligent tin smelting control system of the present invention;

[0054] Figure 2 This is a functional module schematic diagram of an embodiment of the AI ​​artificial intelligence intelligent tin smelting control system of the present invention;

[0055] Figure 3 This is a functional module diagram of a data acquisition and preprocessing module of an embodiment of the AI ​​artificial intelligence intelligent tin smelting control system of the present invention;

[0056] Figure 4 This is a functional module diagram of an intelligent decision-making and control module of an embodiment of the AI ​​artificial intelligence intelligent tin smelting control system of the present invention;

[0057] Figure 5This is a functional module diagram of a real-time monitoring and early warning module of an embodiment of the AI ​​artificial intelligence intelligent tin smelting control system of the present invention;

[0058] Figure 6 This is a schematic diagram of the steps of an embodiment of the AI ​​artificial intelligence intelligent tin smelting control method of the present invention;

[0059] Figure 7 It is a structural schematic diagram of an embodiment of an electronic device of the present invention;

[0060] Figure 8 It is a schematic structural diagram of an embodiment of the storage medium of the present invention. DETAILED DESCRIPTION

[0061] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] like Figure 1 As shown, this embodiment provides an embodiment of an AI artificial intelligence intelligent tin smelting control system. In this embodiment, the AI ​​artificial intelligence intelligent tin smelting control system specifically includes a data acquisition and preprocessing module 1, an intelligent decision-making and control module 2, and a real-time monitoring and early warning module 3 which are electrically connected in sequence;

[0063] Among them, the data acquisition and preprocessing module 1 is used to collect process data in the tin smelting production process in real time, and clean the collected process data; extract and select features that affect the smelting quality and efficiency beyond the preset threshold, and standardize the data; the process data includes: monitoring the smelting temperature, oxygen concentration in the furnace and tin flow; and detecting the reduction degree of the finished product at the terminal, real-time monitoring of the temperature changes of the slag and flue gas, etc.; the intelligent decision-making and control module 2 is used to receive the standardized process data, train the model, process the real-time monitored data and output the adjusted process parameters, including gun position, compressed gas flow, oxygen concentration, various material feed amounts, material ratios, furnace pressure and temperature, etc.; the real-time monitoring and early warning module 3 is used to display the key parameters and data in the production process in real time, analyze the real-time data, detect abnormal conditions, warn of faults, and automatically trigger response measures.

[0064] Preferably, the data acquisition and preprocessing module of this embodiment collects key process data (such as smelting temperature, oxygen concentration in the furnace, tin flow, slag and flue gas temperature, etc.) in the smelting process in real time to ensure the timeliness and accuracy of the data; cleans the collected data, removes noise and outliers, and extracts important features related to smelting quality and efficiency, and converts them into standardized forms for subsequent processing. Significance: Through cleaning and standardization, the data input to the intelligent decision-making module is ensured to be accurate and available, thereby improving the reliability of decision-making; extracting key features enables the decision-making system to more accurately understand and analyze changes in the smelting process, which helps to better predict and optimize the smelting effect. The intelligent decision-making and control module analyzes and trains the standardized process data, and uses machine learning or deep learning algorithms to form a prediction model; according to the data obtained by real-time monitoring, the adjusted process parameters (such as gun position, compressed gas flow, oxygen concentration, etc.) are output to realize the automatic control of the smelting process. Significance: Accurately adjusting the process parameters can improve the smelting efficiency and the yield of tin, while reducing the production of defective products; automated intelligent decision-making reduces the intervention of manual operation, reduces the risk of human error, and ensures the stability and consistency of production. The real-time monitoring and early warning module monitors and analyzes key parameters in real time. When abnormal conditions occur (such as excessive temperature, abnormal oxygen concentration, etc.), it automatically identifies and issues early warnings. When a fault or abnormal condition is detected, it can automatically trigger corresponding response measures (such as adjusting parameters or stopping production) to prevent accidents. Significance: The real-time monitoring and early warning system can promptly detect potential dangers in the production process, ensure the safety of operators and equipment, and prevent accidents. By timely adjusting and responding to abnormal conditions, it can improve the stability of the production line, reduce downtime, and thus improve overall production efficiency and economic benefits (for specific principles, please refer to the attached Figure 2 ).

[0065] The data acquisition and preprocessing module of this embodiment ensures that the system can obtain the latest data in the production process in a timely manner, provide a basis for intelligent decision-making, remove redundant and erroneous or inconsistent data, improve the accuracy of subsequent processing and analysis, screen out features that affect smelting quality and efficiency by more than a preset threshold from massive data, reduce the amount of calculation and improve the accuracy of the model, and convert data of different dimensions and ranges into a unified format for subsequent processing and analysis; the intelligent decision-making and control module uses historical data and real-time data to train the model and optimize the smelting process parameters. According to the standardized data, the process parameters are adjusted in real time to optimize the smelting process, and automatic control is achieved through AI algorithms, which reduces human errors and improves control accuracy; the real-time monitoring and early warning module displays various key parameters and data in the production process in real time through a visual interface, which is convenient for management personnel to monitor. Conduct in-depth analysis of real-time data to identify potential problems and anomalies, issue early warnings for detected anomalies, and automatically trigger corresponding response measures; provide high-quality and accurate data support for intelligent decision-making, improve the efficiency and accuracy of data processing, reduce human intervention, reduce costs, help quickly identify and correct potential problems in the production process, and improve production stability; realize intelligent control of the smelting process, improve production efficiency and product quality, adjust process parameters in real time, quickly respond to changes in the production process, improve production stability, reduce dependence on skilled workers, reduce labor costs, and increase production flexibility; improve the transparency and controllability of the production process, reduce production risks, timely warn and handle abnormal situations, avoid the expansion of faults, reduce losses, improve production safety and stability, and ensure smooth production.

[0066] In summary, this embodiment realizes the intelligence and automation of the tin smelting process, and can effectively improve the efficiency, quality and safety of the smelting process. Through real-time data monitoring, intelligent decision-making and rapid response, the system greatly reduces manual intervention, improves the production environment, optimizes resource utilization, reduces costs, and promotes the transformation of the tin smelting industry to intelligent manufacturing. This integrated control scheme is not only suitable for tin smelting, but also can provide a model for intelligent production management in other metallurgical and related industries, and has a high promotion value.

[0067] Furthermore, if Figure 3 As shown, the data acquisition and preprocessing module 1 specifically includes:

[0068] The sensor node submodule is used to obtain the target data of the process data acquisition equipment, match the target data with each node in the smelting equipment and production line, deploy the corresponding process data acquisition equipment according to the matching results, and obtain the process data of tin smelting through the process data acquisition equipment;

[0069] Among them, the process data collection equipment includes oxygen content sensors, temperature sensors, flow meters, component analyzers and infrared thermometers, etc.; the oxygen content sensor is installed in the flue of the smelting furnace to monitor the oxygen concentration in the furnace; the temperature sensor is arranged in the smelting furnace to monitor the smelting temperature in the furnace in real time; the flow meter is installed at the conveying entrance to monitor the tin flow; the component analyzer is arranged at the smelting terminal to detect the reduction degree of the finished product; the infrared thermometer is installed outside the smelting furnace to monitor and collect the temperature changes of the slag and flue gas in real time;

[0070] The data processing submodule is used to obtain the original process data collected by the sensor node submodule, clean the collected original process data, and fill in the data if there is a discontinuous phenomenon; extract and select target features, and standardize the collected process data;

[0071] The data cache submodule is used to set up a data cache area between the sensor node submodule and the data processing submodule, and is responsible for temporarily storing and buffering process data when the data processing submodule network is congested.

[0072] Preferably, the sensor node submodule of this embodiment acquires the key process data in the tin smelting process in real time by arranging oxygen content sensors, temperature sensors, flow meters, component analyzers, infrared thermometers and other equipment; ensures that the acquisition equipment matches the smelting equipment and each node of the production line, so as to correctly collect data related to the corresponding position and time. Significance: Realize real-time monitoring of key parameters in the smelting process and provide important operation and control information; through appropriate sensor selection and reasonable layout, ensure that the collected data has high accuracy and credibility, thereby providing a basis for subsequent data processing and decision-making. The data processing submodule cleans the original process data collected by the sensor, removes outliers and noise, and ensures data quality; for missing or intermittent process data, interpolation or estimation methods are used to fill in to ensure data continuity; extract and select key features that affect smelting efficiency and quality from the cleaned process data to facilitate subsequent analysis and decision-making; standardize the collected process data and convert it into a data format of a unified scale to facilitate subsequent analysis and model training. Significance: Data cleaning and filling ensure the integrity and reliability of the data, thereby supporting more accurate decision-making; through feature extraction and selection, the data dimension is reduced, the computational complexity is reduced, and subsequent intelligent decision-making and control are more efficient; standardized processing makes various data comparable, which can be more easily analyzed and modeled. When network congestion or processing delay occurs in the data processing submodule, the data cache submodule temporarily stores and buffers process data to reduce data loss or omission caused by instantaneous traffic surge or transmission problems; by setting up a data cache area, the stability of the data flow is improved to ensure that the data processing module can continuously and efficiently obtain data. Significance: The setting of the data cache area can increase the fault tolerance of the system, so that it can operate normally under high load conditions, and the failure of the entire system will not be caused by data loss; in complex smelting monitoring, the cache mechanism can balance data input and processing processes, improve overall system performance, and ensure real-time performance.

[0073] The sensor node submodule of this embodiment accurately obtains the key process parameters in the smelting process through different types of sensors, and intelligently matches and deploys corresponding data acquisition equipment according to the specific nodes of the smelting equipment and production line to ensure the comprehensiveness and pertinence of data acquisition; the data processing submodule removes noise and abnormal values ​​in the original data, reasonably fills in the missing data, ensures the integrity and consistency of the data, extracts features useful for process analysis from a large amount of data, reduces data dimensions, improves subsequent processing efficiency, and converts data of different dimensions and different ranges into a unified format to facilitate subsequent data analysis and modeling; the data cache submodule temporarily stores and buffers process data when the network is congested or the load of the data processing submodule is too high to prevent data loss. Through the cache mechanism, smooth data transmission is achieved to reduce network fluctuations or processing delays. It reduces the impact of delays on data collection and processing; improves the accuracy and real-time nature of data collection, provides a reliable data basis for subsequent process analysis and optimization, reduces the complexity and cost of data collection through intelligent matching and layout, and improves the efficiency and flexibility of the overall system; improves data quality, provides high-quality input for subsequent data analysis and mining, simplifies the complexity of the problem through feature extraction and selection, improves the efficiency and accuracy of data analysis, and data standardization enables data from different sources and at different time points to be directly compared and analyzed, enhancing the comparability of data; ensures the continuity and integrity of data collection, and ensures the complete preservation of data even in the event of a brief failure in the network or processing system, improves the stability and reliability of the system, and reduces the risk of production interruptions due to data loss or processing delays.

[0074] In summary, the data acquisition and preprocessing module of this embodiment can realize comprehensive and accurate process data monitoring and processing of the tin smelting process. It not only improves the reliability and availability of data, but also provides a solid foundation for subsequent intelligent decision-making and control, ensuring the quality, efficiency and safety of the smelting process. The configuration and functional design of this embodiment have greatly promoted the development of the tin smelting industry towards intelligence and automation, and have contributed to improving production efficiency, saving resources and reducing environmental impact.

[0075] Furthermore, the data processing submodule specifically includes:

[0076] The original process data processing unit performs standardization on the collected original process data to ensure that the mean of the data is 0 and the variance is 1. Based on the standardized data, the covariance matrix is ​​calculated;

[0077] The formula for standardization is as follows:

[0078]

[0079] Where: X is the original process data, μ is the mean of the original process data, and σ is the standard deviation of the original process data; the processed original process data can better perform subsequent covariance calculation;

[0080] The covariance matrix C is used to measure the linear relationship between multiple variables, and its expression is:

[0081]

[0082] Where: C is the covariance matrix, size is m×m, m is the dimension of the feature, X′ is the standardized input data matrix, size is n×m, n is the number of samples; n-1 is the degree of freedom adjustment to ensure that the calculated covariance estimate is unbiased;

[0083] Solve the eigenvector unit, solve the process eigenvalues ​​and process eigenvectors of the covariance matrix, sort the obtained process eigenvalues ​​in descending order, select the process eigenvectors corresponding to some process eigenvalues ​​as the main components, and form a new feature space:

[0084] Among them, solve the eigenvalues ​​and eigenvectors of the covariance matrix C. The characteristic equation is:

[0085] Cv=λv

[0086] Where: v is the eigenvector and λ is the corresponding eigenvalue.

[0087] The eigenvalue represents the degree of variation of the corresponding eigenvector in the covariance matrix; the eigenvalues ​​are selected and sorted from large to small, so that some eigenvectors corresponding to high eigenvalues ​​can be selected as principal components to construct a new feature space;

[0088] According to the selected process characteristic value and the corresponding process characteristic vector, the mapping matrix P is calculated using the following formula:

[0089]

[0090]

[0091] In the formula, Represents the mapping metric in the new feature space, calculating the sample f i and f j The distance between i and f j are two samples in the original process data, P is the mapping matrix, which contains the selected eigenvectors, and T represents the transposition operation; the mapping matrix update rule used according to the given conditions is:

[0092]

[0093] Where: Pt is the mapping matrix l obtained at the tth iteration ij is the sample f i and f j Some measure (such as distance) between them, α is a learning rate or adjustment parameter used to control the magnitude of the update, ∈ ij is some kind of parameter involved in the update, representing an error or a specific condition;

[0094] The feature extraction unit forms a transformation matrix with the selected process feature vectors, multiplies the transformation matrix with the original data matrix, and projects it into a new feature space, thereby converting the original process data into a low-dimensional representation while retaining the main information of the data;

[0095] The process is expressed as:

[0096] Y=X′P

[0097] Where: Y is the data representation after dimensionality reduction, X′ is the normalized original data matrix, and P is the transformation matrix consisting of the selected eigenvectors.

[0098] Preferably, the original process data processing unit of this embodiment standardizes the original process data through a formula so that the mean of the processed data is 0 and the variance is 1; it helps to eliminate the dimensional influence between features and ensure that different features are compared on the same scale; on the basis of standardized data, the covariance matrix is ​​calculated to measure the linear relationship between different variables. The covariance matrix can reveal the correlation, importance and variability between input features. Significance: Data standardization eliminates the instability of model training caused by different dimensions, making subsequent analysis more reliable and accurate; the covariance matrix lays the foundation for subsequent feature selection and data dimensionality reduction. Through in-depth analysis of the covariance matrix, the system can identify important features, thereby optimizing decision support. The eigenvector solving unit obtains eigenvalues ​​and eigenvectors by solving the characteristic equation of the covariance matrix. In this process, the eigenvalue represents the degree of variation of the corresponding feature, and the eigenvector indicates the most important direction in the data; the eigenvalues ​​are arranged in order from large to small, so that the eigenvector corresponding to the high eigenvalue can be selected as the principal component to identify the main source of variation in the data. Significance: By selecting the eigenvector with the largest variance, we can focus on the most important features for subsequent analysis and reduce the dimension of their impact on the results, thereby improving analysis efficiency; using the selected eigenvectors to form a new feature space can effectively convert the original data into a low-dimensional representation, while retaining key information and improving subsequent model performance. The feature extraction unit forms a transformation matrix with the selected process feature vectors and multiplies it with the standardized original data matrix to achieve data dimensionality reduction and map high-dimensional data to a low-dimensional space; through this projection operation, the main information is retained as much as possible, while noise and irrelevant information are compressed. Significance Through data dimensionality reduction, the computational complexity of the model is reduced, and the training and reasoning processes become more efficient; while maintaining the main information, reducing the number of features can improve the generalization ability of the model and reduce the risk of overfitting, thereby improving prediction results and decision quality.

[0099] In summary, in the monitoring of smelting and related industrial processes, efficient and accurate data processing in this embodiment can provide support for real-time monitoring and decision-making, and improve production efficiency; through feature selection and dimensionality reduction, enterprises can allocate resources more effectively and make accurate business strategies, thereby reducing production costs and improving economic benefits; the data processing module provides a high-quality data foundation for subsequent machine learning or intelligent algorithms, and promotes the intelligent and automated development of the smelting process. The data processing submodule plays a vital role in data analysis and intelligent decision-making in tin smelting, enhancing the operability and practicality of the system.

[0100] Furthermore, if Figure 4 As shown, the intelligent decision-making and control module 2 specifically includes:

[0101] The reinforcement learning algorithm submodule is used to simulate an intelligent agent interacting with the environment and obtain observations and rewards from it; train the intelligent agent and make decisions; the intelligent agent takes actions such as adjusting the gun position and changing the feed amount in each state;

[0102] Among them, in the tin smelting control system, the agent is the subject responsible for making decisions during the smelting process, which is used to control and optimize the smelting parameters;

[0103] Environment refers to the entire smelting process and its control system, which is perceived and influenced by the agent;

[0104] State, which describes the current operation status of the smelting process, including various process parameters such as temperature, pressure and gas concentration;

[0105] Action: the actions that the agent can take in each state, such as adjusting the gun position and changing the feed amount;

[0106] Reward: After the agent takes an action, the environment gives the agent a reward signal to measure the quality of the action.

[0107] The training and strategy submodule is used to use the reward function to obtain positive rewards for correct actions and negative rewards for incorrect actions respectively; evaluate the value of the current state of the agent, obtain the relationship between the new action and the best action after sorting in the current strategy, and train the agent using a deep Q network;

[0108] The process parameter adjustment submodule is used to dynamically adjust the tin smelting process parameters, including gun position, compressed gas flow, oxygen concentration, feed amount of various materials, material ratio, furnace pressure and temperature, etc., according to the real-time detected production status and the optimal action after sorting in the current strategy output by the reinforcement learning model.

[0109] Preferably, the reinforcement learning algorithm submodule of this embodiment constructs an interactive model between the agent and the smelting environment. The agent explores the state space by trying different actions (such as adjusting the gun position and changing the feed rate), and obtains observation values ​​and rewards based on the results of the attempts, combining the learning process with decision evaluation; the agent perceives in each state, determines the parameters of the current smelting process (such as temperature, pressure, gas concentration), and selects the best action on this basis, aiming to improve the efficiency and output of the smelting process. Significance: Through reinforcement learning, the system can make fast and effective decisions in a changing smelting environment to cope with different production needs and scenarios and improve the level of automation; the agent has the ability to continuously learn and self-adjust according to feedback in the actual production process, thereby optimizing parameter settings and improving overall production efficiency. The training and strategy submodule evaluates the action results of the agent by setting a reward function. Positive rewards are used to encourage correct actions, and negative rewards are used to punish wrong actions, forming an effective feedback mechanism; the value of the current state is evaluated, and decisions are made based on the relationship between the new action and the optimal action after sorting in the current strategy. The deep Q network (DQN) is used to train the agent to improve its decision-making ability. Significance: Through effective reward feedback, the intelligent agent can quickly identify actions that are beneficial or detrimental to the smelting process, combining "exploration and utilization" to make decisions more scientific and accurate; using the deep Q network to train the intelligent agent can continuously improve the quality of the decision-making strategy, ensure the best performance in the long-term process, and adapt to complex and changing production needs. The process parameter adjustment submodule adjusts the tin smelting process parameters (such as gun position, oxygen concentration, furnace pressure, etc.) in a timely manner according to the real-time monitoring of the production status; applies the strategy derived from the reinforcement learning model to practice; combines the decision output of reinforcement learning with real-time data to achieve intelligent control of the smelting process, and optimizes the efficiency of resource utilization while ensuring product quality and safety. Significance: By adjusting the process parameters in real time, it can quickly respond to production changes, ensure that the various parameters in the smelting process remain within a safe and optimal range, and improve the overall production stability; optimizing the process parameters can not only improve production efficiency, but also reduce the consumption of raw materials, reduce the emission of waste gas and waste, and meet the goals of green manufacturing and sustainable development.

[0110] The reinforcement learning algorithm submodule of this embodiment simulates the real-time interaction between the intelligent agent and the smelting environment, so that the system can conduct experiments and learning without interrupting actual production. By receiving the observation results of the environment and the corresponding reward signals, it provides the intelligent agent with basic data for learning. As the core of decision-making, the intelligent agent can select and execute appropriate actions according to the current state to achieve long-term optimization goals; the training and strategy submodule uses a clear reward function to distinguish between correct and incorrect actions, motivate the value of the intelligent agent in the current state, help select the optimal action after sorting in the current strategy, and use deep neural networks to approximate Q values, solve the dimensionality curse problem of traditional Q learning in high-dimensional state space, and achieve more complex and sophisticated decision-making; the process parameter adjustment submodule obtains production status information in real time through sensors and monitoring equipment, provides data support for the adjustment of process parameters, and adjusts the process parameters according to the current strategy output by the reinforcement learning model. The optimal action after sorting automatically adjusts the key process parameters in the tin smelting process, and optimizes the raw material utilization rate and product quality by adjusting the material ratio and feed amount; it realizes the intelligence and automation of the production process, reduces human intervention, improves production efficiency and stability, and gradually optimizes the production strategy through continuous trial and error and learning to adapt to different working conditions and market demands; it improves the efficiency and accuracy of strategy learning, so that the intelligent agent can more quickly find the optimal or close to the optimal production strategy after sorting in the current strategy, enhances the system's adaptability and robustness, and can maintain stable performance under different production conditions and external disturbances; it realizes the refined control and optimization of process parameters, improves production efficiency and product quality, reduces resource consumption and waste generation, reduces production costs and environmental impacts, improves the flexibility and response speed of the production process, and can adapt to market changes and customer needs more quickly.

[0111] In summary, the functional synergy of the intelligent decision-making and control module of this embodiment realizes the comprehensive intelligent control and optimization of the tin smelting process. Its comprehensive significance lies in: by adopting reinforcement learning technology and deep learning algorithms, the automation and intelligence of the tin smelting process are significantly improved, reducing the need for human intervention and labor intensity; the intelligent decision-making module can achieve precise control in a data-driven manner, reduce energy consumption and costs, and improve production efficiency, thereby enhancing the competitiveness of enterprises; flexibly respond to changes in market demand, adjust production strategies in real time, improve the ability of enterprises to adapt to market changes, and enhance the resilience and sustainable development capabilities of the industry. The intelligent decision-making and control system achieves the production goals of high efficiency, high quality and low loss by enhancing the intelligence and automation of the entire tin smelting process, and promotes the upgrading and development of industry technology.

[0112] Furthermore, the reinforcement learning algorithm submodule specifically includes:

[0113] The data establishment unit is used to establish a process parameter table and initialize random numbers. The parameters in the table include the current temperature, pressure, gas concentration and other process parameter states of the smelting process, and actions and gains such as adjusting the gun position and changing the feed amount. The experience pool is established. During the training process, the intelligent agent collects experience through interaction with the environment and stores the collected experience in the experience pool.

[0114] The state update unit is used to update the state and gain of process parameters. Based on the current process parameter table, the intelligent agent selects the action of adjusting the gun position and changing the feed rate with the largest value in the process parameter table, executes and observes the new state and gain of process parameters; and updates the process parameter table according to the new state and gain of process parameters;

[0115] Among them, the state is updated at each time step t, and the agent is updated according to the current state S t and the selected action A t Update the process parameter status, the status update expression is:

[0116] S t+1 =S t +ΔS+∈(S t , A t )

[0117] In the formula, S t Indicates the current state of process parameters, including important variables such as temperature, pressure, gas concentration, etc. t Indicates that the agent is in state S t The action selected by the agent, such as adjusting the gun position or changing the feed rate, ΔS represents the state change caused by the agent's action and environmental feedback, reflecting the immediate response of the process, ∈(S t , A t ) represents the random disturbance introduced to simulate the noise and uncertainty in the environment and enhance the adaptability of the model to the real situation; the dynamic update mechanism ensures that the agent can adjust to the real-time feedback and continuously improve the decision-making process;

[0118] The action selection process of the agent is based on the Q value of the current state, while considering the possible impact of state changes, and is implemented using the following formula:

[0119] A t = argmax a∈A (Q(S t , a)+αΔS(a, S t ))

[0120] In the formula, Q(S t , a) represents the expected return value of selecting action a in the current state, A represents the set of all possible actions, α is a tuning parameter used to balance the impact of Q value estimation and state change on action selection, ΔS(a, St ) represents the expected change of the state after executing action a, which is obtained by fitting or modeling historical data; the selection mechanism enables the agent to not only focus on short-term immediate rewards, but also consider the potential impact of the action on the future state, so as to make more informed decisions;

[0121] Apply incremental learning mechanism to update the Q value of the state:

[0122] Q(S t , A t )←Q(S t , A t )+β(R t +γmax a Q(S t+1 , a)-Q(S t , A t ))

[0123] In the formula, R t represents the reward received at time t, γ represents the discount factor, which determines the weight of future rewards in the current decision, and β represents the incremental learning rate, which can gradually decay as the learning process proceeds, making the early experience have a greater impact and the later experience less, thereby smoothing the learning process; the incremental update mechanism provides the agent with the flexibility to adjust learning in real time, which promotes learning efficiency in complex situations;

[0124] The state information is analyzed in the frequency domain through Fourier transform, focusing on periodic changes:

[0125]

[0126] In the formula, F(S t ) represents the frequency domain representation of the state vector, which is used to capture and analyze the periodic characteristics or patterns in the state; in the decision-making process, the formula for optimizing action selection by combining frequency domain features is:

[0127] A t = argmax a∈A (Q(F(S t ), a)+αΔS(a, S t ))

[0128] Helping the agent identify trends and cyclical fluctuations in process parameters, so that it can make more targeted adjustments when making decisions;

[0129] The stable training unit is used to stabilize the training process and use a fixed target network to calculate the target process parameter table values. It separates the process of selecting actions and calculating the target process parameter table values, and uses the back propagation algorithm to optimize the process parameters until the target network converges. The performance of the intelligent agent is regularly evaluated during the training process.

[0130] Preferably, the data establishment unit of this embodiment initializes and maintains a table containing key process parameters (such as temperature, pressure, gas concentration, etc.), and performs random number initialization to ensure the diversity of the initial state; collects the experience gained by the agent in the process of interacting with the environment, and stores these experiences in the experience pool for subsequent learning and use. Significance achieved: Through reasonable random initialization, the system can explore more possible states, thereby avoiding falling into local optimality too early; the establishment of the experience pool allows the agent to reuse experience in multiple learning tasks, improve learning efficiency, and reduce dependence on real-time data. State update unit The agent updates the state of the process parameters by executing the selected action, and observes the new state and gain; through the state update formula, the immediate reaction and external influence in the process are reflected. Significance achieved: The agent can continuously correct its own behavior according to environmental feedback, improve the flexibility and adaptability of decision-making; the introduction of random perturbations enables the system to better cope with uncertainty and noise in reality and avoid overfitting problems. Action selection mechanism When making decisions, the agent not only considers the Q value of the current state, but also integrates the potential impact of state changes to optimize action selection; to achieve this selection. Significance achieved: The intelligent agent will not only pursue short-term rewards when making decisions, but will focus on the potential benefits of future states, so as to make better decisions; the optimized action selection process improves the efficiency of learning and enables the intelligent agent to adapt quickly in complex environments. The incremental learning mechanism continuously optimizes the Q-value evaluation of the intelligent agent through the incremental update formula, and gradually improves the learning effect by using real-time feedback. Significance achieved: The incremental update strategy reduces the instability of learning, allowing the intelligent agent to learn smoothly and gradually converge to the optimal strategy; over time, the intelligent agent can better adjust its learning rate and pay attention to important historical experience. Fourier transform and frequency domain analysis obtain the frequency domain characteristics of state parameters through Fourier transform to help identify periodic changes. Significance achieved: The intelligent agent can make more targeted adjustments based on the trends and periodic fluctuations in the process parameters, thereby improving the accuracy of decision-making; frequency domain analysis provides the intelligent agent with richer information and promotes more complex decision-making processes. The stable training unit uses a fixed target network to calculate the target process parameter value, making the training process more stable; separate the steps of selecting actions and calculating the target process parameter value, and optimize the process parameters through the back propagation algorithm. Significance achieved: By fixing the target network, the shock caused by the target value update is reduced, and the convergence speed and stability of training are improved; the performance of the intelligent agent is regularly evaluated so that it can adapt to environmental changes in a timely manner and continuously optimize its strategy.

[0131] The data establishment subunit of this embodiment provides the system with initial smelting environment data, including key process parameters such as temperature, pressure and gas concentration, as well as operable actions such as adjusting the gun position and changing the feed amount and their corresponding gains. By randomly initializing the process parameters and action gains, the diversity and exploratory nature of the system are increased, which helps to cover a wider range of smelting conditions in the early stages of training. The experience pool stores each step of the operation and its results as experience during the interaction between the agent and the environment. The experience pool can efficiently manage and reuse these experiences for subsequent learning and optimization of the agent. The state update subunit is based on the current process parameter table. The agent selects actions such as adjusting the gun position and changing the feed amount that can bring the maximum gain after sorting. After executing the selected action, the new process parameter state and gain are observed and recorded. According to the new state and gain, the process parameter table is updated to reflect the latest state of the current smelting environment. When calculating the target process parameter table value, the stable training subunit uses a fixed target network to reduce fluctuations during the training process and separate the two processes so that the selection of actions is not affected by the current network parameter update. The back propagation algorithm is used to optimize the process parameters. By continuously adjusting the network parameters, the performance of the intelligent agent gradually approaches the optimal solution after sorting in the current strategy. The performance of the intelligent agent is regularly evaluated during the training process to ensure the correctness of the training direction. It provides a rich and explorable initial environment for the intelligent agent, which is the basis for intelligent control optimization. Random number initialization promotes the intelligent agent's exploration of unknown areas and avoids prematurely falling into the optimal solution after sorting in the local current strategy. It improves the utilization rate of data, enables the intelligent agent to learn from past experience, speeds up the learning speed, and alleviates the problems of data correlation and non-stationary distribution by storing and reusing experience, improving the stability and efficiency of learning. It realizes the real-time control and optimization of the smelting process by the intelligent agent. Through continuous trial and error and learning, the intelligent agent can gradually find the optimal smelting strategy after sorting in the current strategy, improving smelting efficiency and product quality. It improves the stability and convergence of training, enables the intelligent agent to learn effective smelting strategies more reliably, separates the process of selecting actions and calculating target values, avoids overfitting deviations during training, and regularly evaluates the performance of the intelligent agent, providing timely feedback and adjustment opportunities for the training process to ensure that the training effect meets expectations.

[0132] In summary, the entire reinforcement learning system of this embodiment can more efficiently build models, adapt to the environment, optimize decisions, and then perform effective intelligent control in complex smelting processes. It not only improves the flexibility and robustness of the system, but also provides strong support and theoretical basis for practical applications.

[0133] Furthermore, the training and strategy submodules specifically include:

[0134] Function building unit, which is used to design a compound reward function based on multi-objective optimization. The agent considers the simultaneous optimization of multiple objectives when optimizing the smelting process.

[0135] Components of the reward function:

[0136] Economic Rewards economic： Related to smelting efficiency and raw material cost, using the formula:

[0137] W economic =α·(Y output -C input )-β·C fuel

[0138] Where Y output is the product output, C input is the raw material cost, C fuel is the energy cost, α and β are weight coefficients;

[0139] Safety Reward W safety : Based on the safety performance indicators in the smelting process, negative scores are used for risk factors such as fire and overheating:

[0140] W safety =-γ·R risk

[0141] In the formula, R risk is the risk assessment index under the current state, γ is the weight coefficient;

[0142] Environmentally friendly reward environment : Used to encourage reduction of waste and emissions, using the formula:

[0143] W environment = -δ·E emissions

[0144] In the formula, E emissions is the emission, δ is the weight coefficient;

[0145] Combining the rewards of each part, the final reward function is obtained by weighted summation:

[0146] R(S t , A t , S t+1 )=W economic +W safety +W environment

[0147] The dynamic adjustment unit is used to generate the action selection probability through the policy model, and adopts an adaptive learning rate to dynamically adjust the learning rate according to the learning performance of the intelligent agent;

[0148] The decision feedback unit is used to introduce a feedback mechanism at each time step. Based on the state monitoring data and environmental changes, it evaluates whether the process parameters under the current strategy are optimal and promptly discovers deviations from the reward function target. Based on historical decision performance feedback, self-supervised learning is introduced to allow the intelligent agent to continuously extract lessons from past experience.

[0149] Preferably, the function construction unit of this embodiment combines economic, safety and environmental friendliness factors to form a comprehensive reward function, which can capture multiple goals in the smelting process and avoid the negative effects caused by the optimization of a single goal; it allows the weight coefficient in the reward function to be flexibly adjusted under different operation stages or environmental conditions to adapt to specific production requirements and external environmental changes. The significance achieved: through multi-objective optimization, it is ensured that the smelting process effectively controls safety risks and environmental impacts while improving economic benefits, and strives to achieve the best overall production effect; the intelligent agent no longer focuses on a single indicator in the decision-making process, but is based on comprehensive considerations, making the decision more scientific and reasonable, and improving the intelligence level of the overall system. The dynamic adjustment unit generates the action selection probability for a specific state through a strategy model, so that the intelligent agent can take the optimal or suboptimal action when facing the corresponding environmental changes, thereby achieving more accurate decision-making; according to the learning performance of the intelligent agent, the learning rate is adaptively adjusted to improve the convergence speed of the model and control the fluctuations in the training process. Significance achieved: The adaptive learning rate helps the intelligent agent adapt to the new production environment more quickly and avoids the impact caused by too large or too small learning rate in the initial learning stage; the action selection probability generated by the strategy model enables the intelligent agent to make flexible adjustments in complex and changing environments, improving the real-time control capability of the production process. At each time step, the decision feedback unit evaluates the effect of the current strategy execution through real-time status monitoring data, and promptly discovers possible deviations and deficiencies; based on historical decision performance, the intelligent agent can learn lessons from past experience and provide a basis for future decisions, thereby continuously optimizing its strategy. Significance achieved: The real-time feedback mechanism enables the intelligent agent to adjust its decision in time according to environmental changes and feedback on production status to ensure the optimality of process parameters; the self-supervised learning mechanism enables the intelligent agent to continuously learn and improve, support long-term automated decision-making and continuous optimization, and make the automation and intelligence of the smelting process more efficient and accurate.

[0150] In summary, the training and strategy submodules of this embodiment play a vital role in the tin smelting control system. Taking into account the three aspects of economy, safety and environment, various indicators of the production process can be comprehensively improved. The dynamic adjustment capability and feedback mechanism enable the process parameters to be adjusted in real time according to the actual situation, thereby ensuring the continuity and controllability of production; through feedback and self-supervision, the intelligent body has the ability to continuously learn and adapt, and continuously improve the accuracy and timeliness of decision-making. It promotes the intelligence level and efficiency of the intelligent decision-making and control module in the tin smelting process, and helps to achieve the goals of safe, environmentally friendly and efficient smelting production.

[0151] Furthermore, the process parameter adjustment submodule specifically includes:

[0152] The strategy network unit is used to select actions such as adjusting the gun position and changing the feed amount, and output an action such as adjusting the gun position and changing the feed amount based on the current state of multiple process parameters such as temperature, pressure and gas concentration;

[0153] The value network unit is used to evaluate the value of the current strategy. Based on the current state of various process parameters such as temperature, pressure, gas concentration, etc., and actions such as adjusting the gun position and changing the feed rate, an estimated value or advantage function is output to guide the update direction of the strategy network subunit.

[0154] The optimal strategy unit is used to set a process probability value between (0,1) during the training process. At each decision moment, the action with the highest estimated value is selected by the complement of the process probability value, and an action is randomly selected by the process probability value to balance the relationship between exploring new actions and using the optimal action after sorting in the current strategy; through continuous updating, the optimal strategy after sorting in the current strategy is gradually reached.

[0155] Preferably, the strategy network unit of this embodiment generates targeted decisions based on the input real-time process parameters (such as temperature, pressure, gas concentration, etc.), selects the most suitable action, including adjusting the gun position and changing the feed rate; can select the appropriate control strategy in a multivariate and dynamically changing state environment to improve the reaction speed and the effectiveness of the decision. The significance achieved: By real-time monitoring of process parameters, the strategy network can respond quickly, reduce reaction delays, and ensure the precise control of the smelting process; constantly adjust the strategy according to feedback and historical data to ensure that the selected action can achieve better production efficiency and product quality in the current state. The value network unit can evaluate the effect of the strategy based on the current state and action, thereby outputting an estimated value or advantage function to reflect the performance of the current strategy; by providing value evaluation feedback, it guides the update direction of the strategy network, so that the strategy is gradually adjusted in the direction of higher value. The significance achieved: provides a method to quantify the advantages and disadvantages of the strategy, helps to discover the shortcomings of the current strategy, and thus conducts targeted optimization; by evaluating the performance of the strategy, it promotes the benign interaction between the strategy and the environment, and helps the intelligent agent to continuously improve and find a better solution in learning. The optimal strategy unit will set a probability value between 0 and 1 for the strategy. When taking action, it can make a trade-off between exploration (randomly selecting actions) and utilization (selecting the best action) based on the probability value; by setting the exploration probability, it can explore unknown actions while not neglecting the utilization of the currently known best actions. Significance achieved: Encouraging the agent to try new actions during training helps to discover efficient operation strategies that may be overlooked, thereby reducing the risk of overfitting; by randomly selecting actions, it ensures that the strategy maintains a balance between diversity and stability to adapt to various changes and challenges that may arise during the smelting process.

[0156] The strategy network subunit of this embodiment selects and adjusts the gun position and changes the feed rate based on the current process parameter status, thereby realizing real-time control and optimization of the process. Through continuous updating and learning, the strategy network subunit can gradually adapt to different process conditions and changes, and improve the accuracy and stability of control. The value network subunit evaluates the current strategy and outputs an estimated value or advantage function to guide the update direction of the strategy network subunit. Through interaction and feedback with the strategy network subunit, the value network subunit can continuously optimize and update its own evaluation ability to more accurately reflect the quality of the strategy. The optimal strategy subunit after sorting in the current strategy sets a process probability value during the training process to balance the relationship between exploring new actions and using the optimal action after sorting in the current strategy. By using the process probability value at each decision moment, the value network subunit can continuously optimize and update its own evaluation ability to more accurately reflect the quality of the strategy. The complement selects the action with the highest estimated value at present, and randomly selects an action with the process probability value. The optimal strategy subunit after sorting in the current strategy can gradually converge to the optimal strategy after sorting in the current strategy; it realizes the intelligent control of the process, improves production efficiency and product quality, reduces manual intervention and dependence, and reduces production costs and labor costs; it provides an effective evaluation and guidance mechanism for the strategy network subunit, promotes the continuous optimization and improvement of the strategy, and improves the adaptability and robustness of the entire system, so that it can better cope with various process conditions and changes; it realizes an effective balance between exploration and utilization, avoids strategy instability or performance degradation caused by excessive exploration or excessive utilization, and improves the performance and efficiency of the entire system by gradually converging to the optimal strategy after sorting in the current strategy, providing strong support for the intelligent control of the process.

[0157] In summary, the process parameter adjustment submodule of this embodiment significantly improves the intelligence level and effect of the decision-making process through the coordinated operation of these three subunits. The strategy, value and optimal strategy units cooperate with each other to ensure that the intelligent agent can make flexible and accurate decisions in a dynamic environment and improve smelting efficiency and quality; through real-time data feedback and evaluation, continuously optimize and adjust measures to enhance the system's ability to adapt to different process conditions; while optimizing measures to increase output and efficiency, it also creates conditions for reducing safety risks and environmental impacts. Through flexible strategy selection, value assessment and exploration mechanisms, the process parameter adjustment submodule provides strong technical support for intelligent decision-making in the smelting process, helping to achieve efficient, safe and sustainable production goals.

[0158] Furthermore, if Figure 5 As shown in the figure, the real-time monitoring and early warning module specifically includes:

[0159] A real-time data analysis unit is used to receive the current output action and various process parameter states, extract the target features, extract the key parameters from the target features, and display the key parameters;

[0160] The abnormality detection unit is used to detect the extracted current output target characteristics and key parameters, and compare them with the optimal process parameters after sorting in the current strategy in the optimal strategy process parameter table after sorting in the current strategy. If the current output target characteristics and key parameter values ​​are less than the optimal process parameter values ​​after sorting in the current strategy, an early warning signal is issued;

[0161] The early warning and response unit is used to automatically trigger the preset emergency response procedure after receiving the early warning signal, and remind relevant personnel through visual tools and alarm sounds.

[0162] Preferably, the real-time data analysis unit of this embodiment receives the states of various process parameters (such as temperature, pressure, gas concentration, etc.) and the current output action in the smelting process in real time, and performs data parsing and analysis; extracts representative target features from the original data, and further identifies key parameters, which are the core information in the monitoring process. Significance achieved: By continuously collecting and processing data, ensure that the system can quickly reflect the current state of the production process and provide basic information for emergency response; visualize and display key parameters to help operators identify production dynamics in a timely manner, enhance the readability and operability of data, and help the decision-making process. The anomaly detection unit compares the optimal process parameters ranked after the current strategy, and evaluates the target features and key parameters of the current output in real time; by comparison, it can quickly identify the situation that deviates from the normal state; once it is detected that the target feature or key parameter is lower than the set optimal process parameter value, an early warning signal will be automatically issued. Significance achieved: Through real-time monitoring and anomaly detection, deviations can be detected in time at the early stage of the problem, potential accidents or failures can be prevented, and the safety of production can be ensured; having anomaly detection function helps to carry out predictive maintenance, reduce downtime and maintenance costs, and improve the overall reliability of the system. After receiving the warning signal, the early warning and response unit can quickly and automatically execute the preset emergency response procedures to ensure efficient handling of any abnormal situation; the early warning information is displayed through visual tools, and the relevant operators are reminded through sound alarms and other means to attract sufficient attention. Significance achieved: The automation function of the early warning and response unit can ensure that measures are taken in time when abnormalities occur, significantly improving the reaction and emergency handling capabilities of on-site personnel; the effective early warning and response mechanism enhances the safety awareness of operators, improves their alertness to potential risks, and helps to build a safer working environment.

[0163] The real-time data analysis unit of this embodiment can receive the current output action and various process parameter states in real time to ensure the timeliness and integrity of the data, and extract target-related features from a large amount of data. These features can reflect the key information in the process or production process; further select key parameters from the target features, and display the key parameters in an intuitive manner, so that relevant personnel can quickly understand the current process or production status; the abnormality detection unit can monitor the extracted current output target features and key parameters in real time to ensure timely detection of abnormal situations, and compare the current output target features and key parameters with the optimal process parameters after sorting in the current strategy in the optimal strategy process parameter table to evaluate whether the current process or production status deviates from the optimal state after sorting in the current strategy; when it is found that the current output target features and key parameter values ​​are less than the optimal process parameter values ​​after sorting in the current strategy, an early warning signal is automatically issued to remind relevant personnel to pay attention; early warning and response unit It can receive early warning signals from the abnormality detection unit to ensure the timely transmission of information, automatically trigger the preset emergency response program to deal with possible abnormal situations, remind relevant personnel of abnormal situations through visualization tools and alarm sounds, and guide them to take corresponding measures; improve the efficiency and accuracy of data processing, and provide a reliable basis for subsequent abnormality detection and early warning; through the display of key parameters, it enhances the transparency and controllability of the system, which helps relevant personnel to find problems and take measures in time; through real-time monitoring and comparative analysis, it can timely discover abnormal situations in the process or production process to avoid the occurrence of potential problems. The issuance of early warning signals helps relevant personnel to take measures in time to reduce losses and ensure production safety; through the triggering of the emergency response program, it can quickly respond to abnormal situations, reduce losses and ensure production safety. The visualization and alarm reminder functions enhance the ease of use and reliability of the system, which helps relevant personnel to quickly understand abnormal situations and respond.

[0164] In summary, the real-time monitoring and early warning module of this embodiment has established an efficient monitoring and response system through the synergy of these three units, ensuring the safety and stability of the tin smelting production process. The goals achieved include: through real-time data analysis and anomaly detection, the monitoring capability of the production process is effectively improved, so that the system has the intelligent features of self-regulation and real-time early warning; with the help of an effective early warning mechanism, potential problems are discovered and handled in a timely manner, the risk of accidents is reduced, and the safety and health of employees are ensured; through rapid decision-making and response, the stagnation time in the production process is reduced, the overall operation efficiency is ensured, and lean production is achieved. The efficient monitoring and early warning mechanism enables the tin smelting process to continue to advance in a safe, intelligent, and green direction, laying the foundation for achieving the goal of sustainable development.

[0165] Furthermore, the real-time data analysis unit specifically includes:

[0166] The target feature selection subunit is used to perform pre-processing operations such as cleaning and correction on the data. Based on the optimal process parameters after sorting in the current strategy in the process parameter table of the optimal strategy after sorting in the current strategy, the most relevant target features are selected from the original data to form an output feature set;

[0167] The association rule calculation subunit is used to scan the output feature set, calculate the support of each item, and obtain the candidate item set. If the calculated support of the candidate item set is greater than the preset threshold, it is marked as a frequent item set, and the frequent item set 1 is found; the candidate item set 2 is generated using the frequent item set 1, and the output feature set is scanned again, the support is calculated, and the frequent item set 2 is found; the above process is repeated until no new frequent item sets can be generated, and finally all the association rules between the key parameters and the target features that make the support meet the preset threshold are output;

[0168] The support calculation formula is:

[0169] Support = number of features in the item set / total number of features

[0170] The model training subunit is used to utilize the association rules, use the key parameters and target features whose support meets the preset threshold as the training set of the model to train the model, and evaluate the performance of the model through methods such as cross-validation. After the model is built and trained, the current output target features are input to visualize the output key parameters.

[0171] Preferably, the target feature selection subunit of this embodiment cleans and corrects the original data to ensure data quality, eliminate outliers and errors, and improve the reliability of subsequent analysis; based on the optimal process parameters sorted in the current strategy, the most relevant target features are selected from the standardized and cleaned original data to form an output feature set; the feature is the key indicator that has the greatest impact on the output result in the process parameters. The significance achieved: by cleaning the data and screening the features, it is ensured that the subsequent analysis and modeling are based on high-quality data, which improves the accuracy and reliability of the results; extracting the most relevant features can provide a more accurate basis to help subsequent decision-making and model optimization, and improve production efficiency and safety. The association rule calculation subunit scans the output feature set, calculates the support of each feature item, and finds out the frequent item sets that meet the preset threshold; the subunit will repeat this process to generate multiple frequent item sets, and finally output the association rules between all key parameters and target features whose support meets the requirements; the application of the calculation formula enables the quantification of the frequency of occurrence of different item sets, providing a basis for subsequent algorithms. Significance achieved: Through the mining of frequent item sets and association rules, the potential relationship between target features and key parameters can be discovered, helping to understand the mutual influence between process parameters; this analysis provides rich feature and parameter support for subsequent model building, so that the model has better prediction capabilities, thereby optimizing the control and management of the smelting process. The model training subunit uses key parameters and target features that meet the support threshold as training sets, and trains and optimizes the selected model through evaluation methods such as cross-validation; in this way, patterns and laws in the data are extracted; after training, the target features of the current output are input, and the system can visualize the impact of key parameters, which enables operators to understand the data more intuitively. Significance achieved: Through the construction and training of the model, the prediction and optimization of the smelting process can be achieved, providing guidance for subsequent actual production; the visual output helps decision makers better understand the relationship between key parameters and target features, enhance the initiative of process control, and improve the transparency and control capabilities of production.

[0172] In summary, each subunit in the real-time data analysis unit of this embodiment has constructed a scientific decision support system through refined data processing and feature analysis, which is closely integrated with the needs of the tin smelting process. The goals achieved include: being able to use real-time data to generate an extensible model, thereby effectively refreshing and improving the understanding and control of the production process; through the mining of features and the establishment of association rules, the system has the ability to make intelligent decisions, improving production efficiency and product quality; achieving accurate analysis and rapid response, which helps to optimize the smelting process, enhance the system's ability to respond to abnormal situations, and ensure production safety; the real-time data analysis unit can provide deep-level intelligent support for the tin smelting process, and promote a more efficient, intelligent and sustainable production process.

[0173] The target feature selection subunit of this embodiment ensures the accuracy and consistency of the data through operations such as cleaning and correction, providing a reliable basis for subsequent feature selection. Based on the optimal strategy process table after sorting in the current strategy, the target features most relevant to the optimal process parameters after sorting in the current strategy are screened out from the original data, and the output feature set is constructed to remove redundant information and retain features that have an important impact on model training. The association rule calculation subunit calculates the support of each item to screen out frequent item sets, thereby finding out the association rules between key parameters and target features, revealing potential associations in the data, and providing valuable information for model training. The candidate item sets are generated using frequent item sets, and the support rule calculation subunit is repeated. The degree is maintained until no new frequent item sets can be generated. All association rules that meet the preset threshold can be mined to provide comprehensive information support for the model. The model training subunit uses the training set generated by the association rules to train the model to ensure that the model can accurately predict key parameters and improve the prediction performance of the model so that it can cope with complex process. The performance of the model is evaluated through cross-validation and other methods to ensure the stability and reliability of the model, verify the applicability of the model, and provide strong support for the practical application of the model. The key parameters predicted by the model are output in a visual way to facilitate relevant personnel to quickly understand the process status, enhance the usability and comprehensibility of the system, and improve the overall performance of the system.

[0174] like Figure 6 As shown, this embodiment also provides an embodiment of an AI artificial intelligence intelligent tin smelting control method. In this embodiment, the AI ​​artificial intelligence intelligent tin smelting control method is applied to the AI ​​artificial intelligence intelligent tin smelting control system in the above embodiment. The AI ​​artificial intelligence intelligent tin smelting control method specifically includes the following steps:

[0175] Step S1: real-time collection of process data in the tin smelting production process, and cleaning of the collected process data; extraction and selection of features that affect smelting quality and efficiency exceeding a preset threshold, and standardization of the data;

[0176] Step S2: receiving standardized process data, training the model, processing the real-time monitored data and outputting the adjusted process parameters, including gun position, compressed gas flow, oxygen concentration, feed amount of various materials, material ratio, furnace pressure and temperature, etc.;

[0177] Step S3: Display key parameters and data in the production process in real time, analyze real-time data, detect abnormal situations, warn of faults, and automatically trigger response measures.

[0178] Preferably, step S1 of this embodiment ensures the timeliness and accuracy of the data, provides timely information support for subsequent analysis and decision-making, removes noise, duplication and outliers in the data, improves data quality, and lays a foundation for subsequent feature extraction and model training; screens out features that affect smelting quality and efficiency by more than a preset threshold, reduces redundant information, and improves model training efficiency; converts data into a unified standard form, eliminates dimensional differences between different features, and improves the stability and predictive performance of the model; step S2 uses standardized process data to train the model so that the model can learn the laws and patterns in the data; according to the real-time monitored data, adjusts the process parameters output by the model, including gun position, compressed gas flow rate and oxygen concentration, etc., to optimize the smelting process; step S3 displays various key parameters and data in the production process in a visual manner, so that operators can understand the production status in real time, analyze real-time data, detect abnormal conditions, and promptly discover potential problems in the production process. In the problem; warn of detected abnormal situations and automatically trigger response measures, such as shutdown inspection and adjustment of process parameters, to prevent the occurrence of failures; collect and preprocess process data in real time to provide timely and accurate information support for subsequent analysis and decision-making, and reduce the complexity and time cost of model training and improve the prediction accuracy of the model through feature extraction and selection; data standardization processing helps to eliminate dimensional differences between data and improve the generalization ability of the model; through model training, realize intelligent monitoring and prediction of the smelting process, improve production efficiency and product quality, adjust process parameters in real time, and dynamically optimize the smelting process according to production conditions to reduce energy consumption and costs; display key parameters and data in real time to improve the production monitoring capabilities of operators and ensure the smooth progress of the production process; through data analysis and early warning, timely discover and handle abnormal situations in the production process, reduce failure rate and downtime, automatically trigger response measures, and quickly respond to emergencies in the production process to ensure production safety.

[0179] like Figure 7 As shown, this embodiment provides an embodiment of an electronic device. In this embodiment, the electronic device 4 includes a processor 41 and a memory 42 coupled to the processor 41.

[0180] The memory 42 stores program instructions for implementing the AI ​​artificial intelligence intelligent tin smelting control method of any of the above embodiments.

[0181] The processor 41 is used to execute program instructions stored in the memory 42 to perform AI artificial intelligence intelligent tin smelting control.

[0182] The processor 41 may also be referred to as a CPU (Central Processing Unit). The processor 41 may be an integrated circuit chip having signal processing capabilities. The processor 41 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0183] Further, Figure 8 The schematic diagram of the structure of the storage medium of an embodiment of the present application is that the storage medium 5 of the embodiment of the present application stores program instructions 51 that can implement all the above methods, wherein the program instructions 51 can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0184] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0185] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software functional units. The above is only an implementation mode of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly used in other related technical fields, is also included in the patent protection scope of the present invention.

[0186] The specific implementation methods of the invention are described in detail above, but they are only examples, and the invention is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the invention, therefore, the equalization, modification, improvement, etc. made without departing from the spirit and principle of the invention should be included in the scope of the invention.

Claims

1. An AI artificial intelligence intelligent tin smelting control system, characterized in that: The AI ​​artificial intelligence intelligent tin smelting control system includes: The data acquisition and preprocessing module is used to collect process data in the tin smelting production process in real time and clean the collected process data; extract and select features that affect the smelting quality and efficiency beyond the preset threshold, and standardize the data; the process data includes: smelting temperature, oxygen concentration in the furnace and tin flow; and detect the reduction degree of the finished product at the terminal, and monitor the temperature changes of the slag and flue gas in real time; Intelligent decision-making and control module, which is used to receive standardized process data, train the model, process the real-time monitored data and output the adjusted process parameters, including gun position, compressed gas flow, oxygen concentration, feed amount of various materials, material ratio, furnace pressure and temperature; The real-time monitoring and early warning module is used to display the key parameters and data of the production process in real time, analyze the real-time data, detect abnormal situations, warn of faults, and automatically trigger response measures.

2. The AI ​​artificial intelligence intelligent tin smelting control system according to claim 1 is characterized in that: Data acquisition and preprocessing modules, including: The sensor node submodule is used to obtain the target data of the process data acquisition equipment, match the target data with each node in the smelting equipment and production line, deploy the corresponding process data acquisition equipment according to the matching results, and obtain the process data of tin smelting through the process data acquisition equipment; The data processing submodule is used to obtain the original process data collected by the sensor node submodule, clean the collected original process data, and fill in the data if there is a discontinuous phenomenon; extract and select target features, and standardize the collected process data; The data cache submodule is used to set up a data cache area between the sensor node submodule and the data processing submodule, and is responsible for temporarily storing and buffering process data when the data processing submodule network is congested.

3. The AI ​​artificial intelligence intelligent tin smelting control system according to claim 2 is characterized in that: Data processing submodule, including: The original process data processing unit performs standardization on the collected original process data to ensure that the mean of the data is 0 and the variance is 1. Based on the standardized data, the covariance matrix is ​​calculated; Solve the eigenvector unit, solve the process eigenvalue and process eigenvector of the covariance matrix, sort the obtained process eigenvalues ​​in descending order, select the process eigenvectors corresponding to some process eigenvalues ​​as the main components, and form a new feature space; The feature extraction unit forms a transformation matrix with the selected process feature vectors, multiplies the transformation matrix with the original data matrix, and projects it into a new feature space, thereby converting the original process data into a low-dimensional representation while retaining the main information of the data.

4. The AI ​​artificial intelligence intelligent tin smelting control system according to claim 1 is characterized in that: Intelligent decision-making and control module, including: The reinforcement learning algorithm submodule is used to simulate an intelligent agent interacting with the environment and obtain observations and rewards from it; train the intelligent agent and make decisions; the intelligent agent takes actions such as adjusting the gun position and changing the feeding amount in each state; The training and strategy submodule is used to use the reward function to obtain positive rewards for correct actions and negative rewards for incorrect actions respectively; evaluate the value of the current state of the agent, obtain the relationship between the new action and the best action after sorting in the current strategy, and train the agent using a deep Q network; The process parameter adjustment submodule is used to dynamically adjust the tin smelting process parameters, including gun position, compressed gas flow, oxygen concentration, feed amount of various materials, material ratio, furnace pressure and temperature, according to the real-time detected production status and the optimal action after sorting in the current strategy output by the reinforcement learning model.

5. The AI ​​artificial intelligence intelligent tin smelting control system according to claim 4 is characterized in that: Reinforcement learning algorithm submodule, including: The data establishment unit is used to establish a process parameter table and initialize random numbers. The parameters in the table include the current temperature, pressure, gas concentration and other process parameter states of the smelting process, adjust the gun position, change the feed amount action and gain; establish an experience pool. During the training process, the intelligent agent collects experience through interaction with the environment and stores the collected experience in the experience pool; The state update unit is used to update the state and gain of the process parameters. Based on the current process parameter table, the intelligent agent selects the action of adjusting the gun position and changing the feeding amount with the largest value in the process parameter table, executes and observes the new state and gain of the process parameters; and updates the process parameter table according to the new state and gain of the process parameters; The stable training unit is used to stabilize the training process and use a fixed target network to calculate the target process parameter table values. It separates the process of selecting actions and calculating the target process parameter table values, and uses the back propagation algorithm to optimize the process parameters until the target network converges. The performance of the intelligent agent is regularly evaluated during the training process.

6. The AI ​​artificial intelligence intelligent tin smelting control system according to claim 4 is characterized in that: Training and strategy submodules, including: Function building unit, which is used to design a compound reward function based on multi-objective optimization. The agent considers the simultaneous optimization of multiple objectives when optimizing the smelting process. The dynamic adjustment unit is used to generate the action selection probability through the policy model, and adopts an adaptive learning rate to dynamically adjust the learning rate according to the learning performance of the intelligent agent; The decision feedback unit is used to introduce a feedback mechanism at each time step. Based on the state monitoring data and environmental changes, it evaluates whether the process parameters under the current strategy are optimal and promptly discovers deviations from the reward function target. Based on historical decision performance feedback, self-supervised learning is introduced to allow the intelligent agent to continuously extract lessons from past experience.

7. The AI ​​artificial intelligence intelligent tin smelting control system according to claim 4 is characterized in that: Process parameter adjustment submodule, including: The strategy network unit is used to select the action of adjusting the gun position and changing the feeding amount, and outputs an action of adjusting the gun position and changing the feeding amount based on the current state of multiple process parameters such as temperature, pressure and gas concentration; The value network unit is used to evaluate the value of the current strategy. Based on the current state of various process parameters such as temperature, pressure, gas concentration, etc., the actions of adjusting the gun position and changing the feed amount, an estimated value or advantage function is output to guide the update direction of the strategy network subunit; The optimal strategy unit is used to set a process probability value between (0,1) during the training process. At each decision moment, the action with the highest estimated value is selected using the complement of the process probability value, and an action is randomly selected using the process probability value to balance the relationship between exploring new actions and utilizing the optimal action after sorting in the current strategy. Through continuous updating, the optimal strategy after sorting in the current strategy is gradually reached.

8. The AI ​​artificial intelligence intelligent tin smelting control system according to claim 1 is characterized in that: Real-time monitoring and early warning modules, including: A real-time data analysis unit is used to receive the current output action and various process parameter states, extract the target features, extract the key parameters from the target features, and display the key parameters; The abnormality detection unit is used to detect the extracted current output target characteristics and key parameters, and compare them with the optimal process parameters after sorting in the current strategy in the optimal strategy process parameter table after sorting in the current strategy. If the current output target characteristics and key parameter values ​​are less than the optimal process parameter values ​​after sorting in the current strategy, an early warning signal is issued; The early warning and response unit is used to automatically trigger the preset emergency response procedure after receiving the early warning signal, and remind relevant personnel through visual tools and alarm sounds.

9. The AI ​​artificial intelligence intelligent tin smelting control system according to claim 8 is characterized in that: Real-time data analysis unit, including: The target feature selection subunit is used to perform cleaning and correction preprocessing operations on the data, and select the most relevant target features from the original data based on the optimal process parameters after sorting in the current strategy in the optimal strategy process parameter table to form an output feature set; The association rule calculation subunit is used to scan the output feature set, calculate the support of each item, and obtain the candidate item set. If the calculated support of the candidate item set is greater than the preset threshold, it is marked as a frequent item set, and the frequent item set 1 is found; the candidate item set 2 is generated using the frequent item set 1, and the output feature set is scanned again, the support is calculated, and the frequent item set 2 is found; the above process is repeated until no new frequent item sets can be generated, and finally all the association rules between the key parameters and the target features that make the support meet the preset threshold are output; The support calculation formula is: Support = number of features in the item set / total number of features The model training subunit is used to utilize the association rules, use the key parameters and target features whose support meets the preset threshold as the training set of the model to train the model, and evaluate the performance of the model through the cross-validation method. After the model is built and trained, the current output target features are input to visualize the output key parameters.

10. An AI artificial intelligence intelligent tin smelting control method, characterized in that: The following steps are involved: Collect process data from the tin smelting production process in real time and clean the collected process data; extract and select features that affect smelting quality and efficiency beyond the preset threshold, and standardize the data; Receive standardized process data, train the model, process the real-time monitored data and output the adjusted process parameters, including gun position, compressed gas flow, oxygen concentration, feed amount of various materials, material ratio, furnace pressure and temperature; Display key parameters and data of the production process in real time, analyze real-time data, detect abnormal situations, warn of faults, and automatically trigger response measures.

Images