Valve body casting process intelligent control method based on industrial internet
By introducing digital twin modules and closed-loop control into the valve body casting process, the problems of delayed anomaly detection and low fault diagnosis efficiency in traditional methods have been solved. Real-time monitoring and prediction of key parameters have been achieved, improving the efficiency of the casting process and the stability of the equipment.
Patent Information
- Application Number
- CN202511343871.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-19
- Publication Date
- 2025-12-05
AI Technical Summary
In the traditional valve body casting process, the detection of anomalies is delayed, the troubleshooting efficiency is low, and there is a lack of preventive control. This results in long anomaly handling time, high rate of unplanned equipment downtime, and untimely system response to fluctuations in key parameters, which affects casting quality and equipment stability.
By embedding a digital twin module and combining real-time data acquisition with a closed-loop feedback mechanism, a dynamic and interactive digital twin model is established to enable rapid identification and response to potential abnormal conditions. Process parameters are adjusted through closed-loop control commands, forming a closed-loop management process of prediction, early warning, and feedback.
It significantly improves the efficiency of anomaly handling and equipment operation stability, enables real-time monitoring and accurate prediction of key parameters, and enhances the efficiency and reliability of the casting process.
Smart Images
Figure CN121069884A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of industrial control system technology, specifically a smart control method for valve body casting process based on the Industrial Internet. Background Technology
[0002] With the continuous integration of industrial intelligence and manufacturing processes, the application of industrial internet technology in complex production scenarios is gradually deepening. This is especially true in the control and optimization of valve body casting processes, where real-time monitoring and precise control of key process parameters are of paramount importance.
[0003] Traditional valve body casting process control relies heavily on manual experience and static rule settings, which are significantly inadequate when dealing with dynamically changing production environments. On the one hand, anomaly detection often lags behind the actual occurrence of problems, leading to prolonged processing time. On the other hand, troubleshooting efficiency is low, making it difficult to quickly pinpoint the root cause and implement effective measures. Furthermore, traditional methods lack preventative control measures and cannot intervene in advance by predicting potential abnormal conditions.
[0004] This approach results in insufficient timeliness in the system's response to fluctuations in key parameters such as furnace temperature, casting robot flow rate, and cooling rate, thus affecting casting quality and equipment stability. Consequently, existing technologies perform poorly in terms of anomaly handling time and unplanned equipment downtime, exhibiting both long anomaly handling times and high unplanned downtime rates. Therefore, there is an urgent need for an intelligent method capable of real-time monitoring, accurate prediction, and closed-loop control to improve the efficiency and reliability of the valve body casting process. Summary of the Invention
[0005] To address the technical problems of delayed anomaly detection, low fault diagnosis efficiency, and lack of preventative control in valve body casting processes within traditional industrial control systems, this invention provides an intelligent control method and system for valve body casting processes based on the Industrial Internet. This method significantly improves anomaly handling efficiency and equipment operational stability by embedding a digital twin module and combining real-time data acquisition with a closed-loop feedback mechanism. Specifically, this invention establishes a dynamically interactive digital twin model to simulate changes in key parameters during the casting process, and achieves rapid identification and response to potential anomalies through real-time monitoring and predictive analysis. Simultaneously, this invention achieves precise adjustment of process parameters through the generation and execution of closed-loop control commands, and synchronously updates the status data of the digital twin model, thus forming a complete closed-loop management process of prediction, early warning, and feedback.
[0006] According to one aspect of the present invention, an intelligent control method for valve body casting process based on the Industrial Internet is provided, the method comprising the following steps: A valve body casting digital twin module is embedded in the industrial control system architecture. The operating status data in the physical production scenario is collected in real time through the industrial Internet. The operating status data includes the melting furnace temperature, the flow rate of the casting robot arm, and the cooling rate of the cooling system. The digital twin module is used to model and process the operating status data to generate a digital twin model. The casting process under different combinations of process parameters is simulated using the digital twin model to obtain prediction results. Based on the prediction results, potential abnormal conditions are analyzed, and abnormal thresholds are defined. The abnormal thresholds include a furnace temperature deviation range of ±5 degrees Celsius and a casting flow rate deviation range of ±8%. When the deviation between the actual operating parameters and the prediction results of the digital twin model exceeds the abnormal threshold, a closed-loop control command is generated. The closed-loop control command is sent to the actuator to adjust the process parameters, and the status data of the digital twin model is updated at the same time to complete the closed-loop management of predictive early warning control feedback.
[0007] Furthermore, as one embodiment of the present invention, the step of using the digital twin module to model the operating status data and generate a digital twin model includes: The real-time data streams from the smelting furnace temperature sensor, the casting robot arm flow meter, and the cooling system rate monitoring device are acquired, and the real-time data streams are preprocessed to remove noisy data. An initial digital twin model is constructed based on the preprocessed real-time data stream. The initial digital twin model includes the heat transfer equation of the smelting furnace, the flow distribution function of the casting robot arm, and the heat dissipation curve of the cooling system. By continuously receiving new real-time data streams through the Industrial Internet, the initial digital twin model is dynamically updated to generate an optimized digital twin model. The accuracy of the optimized digital twin model is verified. If the model error is less than a preset error threshold, the optimized digital twin model is used as the final digital twin model for subsequent predictive analysis.
[0008] Furthermore, as one embodiment of the present invention, the step of generating a closed-loop control command when the deviation between the actual operating parameters and the prediction result of the digital twin model exceeds the anomaly threshold includes: Calculate the deviation between the actual operating parameters and the prediction results of the digital twin model, and determine whether the deviation exceeds the anomaly threshold; If the deviation value exceeds the abnormal threshold, a corresponding closed-loop control strategy is determined according to the deviation type. The closed-loop control strategy includes heating device power adjustment rules and flow regulating valve opening adjustment rules. Based on the closed-loop control strategy, a closed-loop control command is generated and sent to the actuator. The execution results of the closed-loop control commands are recorded, and the execution results are synchronized to the digital twin model to update the model state data.
[0009] Furthermore, as one embodiment of the present invention, the step of generating closed-loop control commands based on the closed-loop control strategy includes: Based on the magnitude and direction of the deviation value, determine the power adjustment amount of the heating device or the opening adjustment amount of the flow regulating valve; The industrial control protocol interface is invoked to convert the adjustment amount into a control signal that conforms to industrial communication standards; A closed-loop control command containing the control signal is generated and transmitted to the corresponding actuator via the Industrial Internet; During the execution of closed-loop control commands, the response status of the actuator is monitored in real time. If the response status does not meet the expected target, the adjustment amount is recalculated and the closed-loop control command is updated.
[0010] Furthermore, as one embodiment of the present invention, verifying the accuracy of the optimized digital twin model includes: A test dataset is extracted from historical operating status data, which includes the furnace temperature, casting robot flow rate, and cooling system rate at multiple time points. The test dataset is input into the optimized digital twin model to generate a prediction output; Calculate the error value between the predicted output and the actual test data. If the error value is less than a preset error threshold, then the optimized digital twin model is determined to meet the accuracy requirements. If the error value is greater than or equal to the preset error threshold, the optimized digital twin model is further corrected until the error value meets the accuracy requirements.
[0011] Furthermore, as one embodiment of the present invention, the real-time collection of operational status data in the physical production scenario via the Industrial Internet includes: Establish communication connections between the Industrial Internet and the temperature sensors of the smelting furnace, the flow meters of the casting robotic arm, and the rate monitoring devices of the cooling system to ensure the real-time performance and reliability of data transmission; Configure the data acquisition frequency to match the dynamic changes of key process parameters in the casting process. The data acquisition frequency ranges from once per second to once every ten seconds, and the specific frequency is dynamically adjusted according to process requirements. The received running status data is timestamped to ensure the consistency of the data over time, and the tagged running status data is stored in a distributed database; Perform preliminary verification on the stored operational status data, remove outliers caused by sensor malfunctions or signal interference, and generate a verified operational status dataset. Based on the verified operational status dataset, a real-time data stream pipeline is constructed to provide continuous and stable data input to the digital twin module.
[0012] Furthermore, as one embodiment of the present invention, the construction of the initial digital twin model based on the preprocessed real-time data stream includes: Analyze the mathematical expressions of the heat transfer equation of the smelting furnace, the flow distribution function of the casting robot arm, and the heat dissipation curve of the cooling system to determine the correlation between various physical quantities; The preprocessed real-time data stream is used to fit the parameters of the correlation to generate an initial model parameter set. Based on the initial model parameter set, a framework structure for the initial digital twin model is constructed, which includes an input layer, a computation layer, and an output layer. Embed process parameter constraints in the computation layer, including the upper limit of the melting furnace temperature, the lower limit of the casting flow rate, and the maximum gradient of the cooling rate. The initial digital twin model was preliminarily validated. By comparing the actual operating data with the model output results, it was confirmed that the model's basic predictive ability meets the design requirements.
[0013] Furthermore, as one embodiment of the present invention, recording the execution result of the closed-loop control command and synchronizing the execution result to the digital twin model to update the model state data includes: Receive feedback from the actuator regarding the actual process parameter adjustment results, including changes in heating device power and changes in flow control valve opening; Time series analysis was performed on the adjustment results to extract key change points in the process parameter adjustment process; The key change points are compared with the prediction results of the digital twin model, and the deviation value between the two is calculated. The state data of the digital twin model is incrementally updated based on the deviation value. The incremental update includes adjusting the model parameter weights and optimizing the model input-output mapping relationship. The updated digital twin model is quickly validated to ensure that the model's prediction accuracy is within the preset error range. If it exceeds the range, the model retraining process is triggered.
[0014] Furthermore, as one embodiment of the present invention, the step of calling the industrial control protocol interface to convert the adjustment amount into a control signal conforming to industrial communication standards includes: Analyze the communication specifications of industrial control protocol interfaces to determine the data format, transmission rate, and encoding rules of control signals; Based on the specific numerical range of the adjustment amount, select the appropriate industrial control protocol type, which includes Modbus, OPC UA, or Profibus; The adjustment amount is converted according to the encoding rules of the selected protocol to generate a control signal that conforms to industrial communication standards; The control signals are transmitted to the target actuator via the Industrial Internet, and the integrity and accuracy of the signal transmission are monitored in real time. If a signal transmission failure or delay exceeding a preset threshold is detected, a signal retransmission mechanism is triggered to ensure reliable delivery of control signals.
[0015] Furthermore, as one embodiment of the present invention, the further correction of the optimized digital twin model includes: A supplementary test dataset is extracted from historical operating status data, and the supplementary test dataset covers multiple abnormal operating condition scenarios. The supplementary test dataset is input into the optimized digital twin model to generate supplementary prediction output. Calculate the error between the supplementary prediction output and the actual test data, and determine the main sources of the error; Based on the main sources of the error values, the parameter optimization strategy of the digital twin model is adjusted. The optimization strategy includes increasing the number of training samples, introducing a nonlinear mapping function, or adjusting the model convergence condition. Repeat the calibration and verification process until the error value meets the preset error threshold to ensure that the optimized digital twin model has high-precision prediction capabilities.
[0016] According to another aspect of the present invention, an intelligent control system for valve body casting process based on the Industrial Internet is provided, the system comprising: The data acquisition and integration module is used to acquire operational status data in the physical production scenario, including furnace temperature, casting robot flow rate, and cooling rate of the cooling system. The digital twin modeling module is used to model and process the operating status data, generate a digital twin model, and simulate the casting process under different combinations of process parameters through the digital twin model to obtain prediction results. The anomaly detection and closed-loop control module is used to analyze potential anomaly conditions based on the prediction results, define anomaly thresholds, and generate closed-loop control commands when the deviation between the actual operating parameters and the prediction results exceeds the anomaly thresholds. The actuator control module is used to receive the closed-loop control command and adjust the process parameters, while feeding back the execution result to the digital twin modeling module to update the model status data.
[0017] Furthermore, as one embodiment of the present invention, the data acquisition and integration module specifically includes: Temperature sensors, flow meters, and rate monitoring devices are used to acquire the temperature of the smelting furnace, the flow rate of the casting robot arm, and the rate of the cooling system, respectively. The communication interface unit is used to establish a communication connection between the Industrial Internet and the aforementioned sensors, ensuring the real-time performance and reliability of data transmission. The data storage unit is used to timestamp the received running status data and store it in the distributed database; The data verification unit is used to perform preliminary verification on the stored running status data, remove outliers, and generate a verified running status dataset.
[0018] Furthermore, as one embodiment of the present invention, the digital twin modeling module specifically includes: A data preprocessing unit is used to perform noise removal and parameter fitting on the operating status data; The model building unit is used to generate an initial digital twin model based on preprocessed data, and to generate an optimized digital twin model through dynamic updates; The model verification unit is used to verify the accuracy of the optimized digital twin model and generate the final digital twin model after correction.
[0019] Furthermore, as one embodiment of the present invention, the anomaly detection and closed-loop control module specifically includes: The deviation calculation unit is used to calculate the deviation between the actual operating parameters and the prediction results of the digital twin model; A threshold determination unit is used to determine whether the deviation value exceeds an abnormal threshold. The instruction generation unit is used to generate closed-loop control instructions when the deviation value exceeds the abnormal threshold. The result synchronization unit is used to synchronize the adjustment results fed back by the actuator to the digital twin modeling module to update the model status data.
[0020] Furthermore, as one embodiment of the present invention, the actuator control module specifically includes: The signal conversion unit is used to convert the adjustment quantity in the closed-loop control command into a control signal that conforms to industrial communication standards. The signal transmission unit is used to transmit control signals to the target actuator via the Industrial Internet; The signal monitoring unit is used to monitor the integrity and accuracy of signal transmission in real time, and to trigger a signal retransmission mechanism when a transmission failure or delay is detected.
[0021] Compared with the prior art, the beneficial effects of the present invention are: The data acquisition and integration module connects to the industrial internet platform via a communication interface unit, ensuring the real-time performance and reliability of data transmission. The digital twin modeling module receives operational status data from the data acquisition and integration module through the industrial internet platform and outputs the generated digital twin model to the anomaly detection and closed-loop control module. The anomaly detection and closed-loop control module sends closed-loop control commands to the actuator control module through the industrial internet platform, while simultaneously receiving adjustment results from the actuator control module to update the digital twin model. All modules collaborate efficiently through the industrial internet platform to jointly complete the intelligent control task of the valve body casting process.
[0022] Through the above technical solution, the present invention not only solves the technical problems existing in the valve body casting process in traditional industrial control systems, but also significantly improves the efficiency of anomaly handling and the stability of equipment operation, providing important technical support for intelligent industrial manufacturing. Attached Figure Description
[0023] Figure 1 This is a schematic diagram of the overall system architecture of an embodiment of the present invention.
[0024] Figure 2 This is a flowchart illustrating the construction process of a digital twin model according to an embodiment of the present invention.
[0025] Figure 3 This is a flowchart illustrating the closed-loop control command generation and execution process according to an embodiment of the present invention.
[0026] Figure 4 This is a schematic diagram of industrial internet communication connection according to an embodiment of the present invention.
[0027] Figure 5 This is a schematic diagram of the dynamic update mechanism of the digital twin model in an embodiment of the present invention. Detailed Implementation
[0028] This invention provides an intelligent control method and system for valve body casting process based on the Industrial Internet, the specific implementation of which is as follows. Figure 1As shown, the overall system architecture includes a data acquisition and integration module, a digital twin modeling module, an anomaly detection and closed-loop control module, and an actuator control module. These modules achieve data interaction and functional collaboration through an industrial internet platform, jointly completing the intelligent control of the valve body casting process.
[0029] In a preferred embodiment, the data acquisition and integration module consists of a smelting furnace temperature sensor, a casting robotic arm flow meter, and a cooling system rate monitoring device. The smelting furnace temperature sensor is installed at a key location in the smelting furnace to monitor real-time temperature changes inside the furnace; the casting robotic arm flow meter is located at the output end of the casting robotic arm to measure the flow rate of molten metal during casting; and the cooling system rate monitoring device is positioned near the radiator of the cooling system to record changes in the cooling rate. These sensors connect to an industrial internet platform via a communication interface unit to ensure the real-time performance and reliability of data transmission. Figure 4 As shown, the communication interface unit is responsible for timestamping the data collected by the sensors and storing the tagged data in a distributed database. The data verification unit performs preliminary verification on the stored data, removing outliers caused by sensor malfunctions or signal interference, thereby generating a verified operational status dataset. This dataset is transmitted to the digital twin modeling module through a real-time data stream pipeline, providing basic data support for subsequent modeling.
[0030] As a preferred implementation, the digital twin modeling module, such as Figure 2 As shown, its core function is to generate a digital twin model by modeling and processing operational status data. First, the data preprocessing unit performs noise removal and parameter fitting operations on the received operational status data to improve data quality. Real-time data streams collected by the smelting furnace temperature sensor, the casting robot flow meter, and the cooling system rate monitoring device are preprocessed and then input into the model building unit. Based on the preprocessed data, the model building unit generates an initial digital twin model, which includes the smelting furnace heat transfer equation, the casting robot flow distribution function, and the cooling system heat dissipation curve.
[0031] The heat transfer equation of the smelting furnace describes the heat transfer within the furnace, and its mathematical expression includes physical quantities such as thermal conductivity, specific heat capacity, and temperature gradient. The flow distribution function of the casting robot reflects the relationship between the molten metal flow rate and the opening of the casting robot. The heat dissipation curve of the cooling system simulates the cooling rate over time. To ensure model accuracy, the model building unit embeds process parameter constraints in the computational layer, such as the upper limit of the smelting furnace temperature, the lower limit of the casting flow rate, and the maximum gradient of the cooling rate. After the initial digital twin model is generated, it continuously receives new real-time data streams through the industrial internet platform to dynamically update the model and generate an optimized digital twin model. The model validation unit verifies the accuracy of the optimized digital twin model by extracting test datasets from historical operating data and inputting them into the model to generate predictive output.
[0032] If the error between the predicted output and the actual test data is less than the preset error threshold, the model is deemed to meet the accuracy requirements; otherwise, the model needs to be further corrected until the error value meets the requirements.
[0033] As a preferred implementation, the anomaly detection and closed-loop control module, such as Figure 3 As shown, its main function is to analyze potential abnormal conditions based on the prediction results of the digital twin model and generate closed-loop control commands. The deviation calculation unit first calculates the deviation value between the actual operating parameters and the prediction results of the digital twin model, and determines whether the deviation value exceeds the abnormal threshold. The abnormal threshold includes a furnace temperature deviation range of ±5 degrees Celsius and a casting flow rate deviation range of ±8%.
[0034] When the deviation exceeds the abnormal threshold, the threshold judgment unit triggers the instruction generation unit to generate a closed-loop control instruction. The instruction generation unit determines the corresponding closed-loop control strategy based on the deviation type, such as the heating device power adjustment rule and the flow control valve opening adjustment rule. The heating device power adjustment rule specifies the magnitude and direction of the power adjustment, while the flow control valve opening adjustment rule defines the range of the flow control valve opening adjustment. The instruction generation unit calls the industrial control protocol interface to convert the adjustment amount into a control signal conforming to industrial communication standards, generates a closed-loop control instruction, and transmits it to the actuator control module via the industrial internet platform. The result synchronization unit receives the adjustment result feedback from the actuator and synchronizes the adjustment result to the digital twin modeling module to update the model state data.
[0035] As a preferred implementation, the actuator control module, such as Figure 3As shown, its core task is to receive closed-loop control commands and adjust process parameters. The signal conversion unit parses the communication specifications of the industrial control protocol interface, converts the adjustment quantities in the closed-loop control commands according to the encoding rules of the selected protocol, and generates control signals that conform to industrial communication standards. The signal transmission unit transmits the control signals to the target actuator, such as a heating device or a flow control valve, through an industrial internet platform.
[0036] The signal monitoring unit monitors the integrity and accuracy of signal transmission in real time. If a signal transmission failure or delay exceeding a preset threshold is detected, a signal retransmission mechanism is triggered to ensure reliable delivery of control signals. Upon receiving the control signal, the actuator adjusts the power or flow regulating valve opening of the heating device according to the signal content, and simultaneously feeds the adjustment result back to the result synchronization unit of the anomaly detection and closed-loop control module. The result synchronization unit compares the adjustment result with the prediction result of the digital twin model, calculates the deviation between the two, and incrementally updates the state data of the digital twin model based on the deviation. Incremental updates include adjusting model parameter weights and optimizing the model input-output mapping relationship to ensure that the model's prediction accuracy always remains within a preset error range.
[0037] As a preferred implementation method, such as Figure 5 As shown, the dynamic update mechanism of the digital twin model runs throughout the entire system operation. While incrementally updating the model via real-time data streams, a rapid validation unit verifies the updated model. If the model's prediction accuracy exceeds a preset error range, a model retraining process is triggered to ensure the model always possesses high-precision prediction capabilities. Furthermore, the digital twin model can cover multiple abnormal operating scenarios; by introducing nonlinear mapping functions or adjusting the model's convergence conditions, the model's adaptability and robustness are further improved.
[0038] In the above implementation, the connections between the modules and components are close and clearly defined. The data acquisition and integration module is connected to the industrial internet platform through a communication interface unit to ensure the real-time performance and reliability of data transmission. The digital twin modeling module receives the operating status data provided by the data acquisition and integration module through the industrial internet platform and outputs the generated digital twin model to the anomaly detection and closed-loop control module. The anomaly detection and closed-loop control module sends closed-loop control commands to the actuator control module through the industrial internet platform, and simultaneously receives adjustment results from the actuator control module to update the digital twin model. The modules achieve efficient collaboration through the industrial internet platform to jointly complete the intelligent control task of the valve body casting process.
[0039] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the term "inclusion" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.
[0040] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. An intelligent control method for valve body casting process based on industrial internet, characterized in that, The method comprises the following steps: embedding a valve body casting digital twin module in an industrial control system architecture, collecting real-time operation state data in a physical production scene through an industrial internet, the operation state data including smelting furnace temperature, pouring mechanical arm flow rate, and cooling system cooling rate; modeling and processing the operation state data using the digital twin module, generating a digital twin model, and simulating the casting process under different process parameter combinations through the digital twin model to obtain a prediction result; analyzing potential abnormal conditions according to the prediction result, defining an abnormal threshold, the abnormal threshold including a smelting furnace temperature deviation range of plus or minus 5 degrees Celsius and a pouring flow deviation range of plus or minus 8 percent; generating a closed-loop control instruction when detecting that the deviation between the actual operation parameters and the prediction result of the digital twin model exceeds the abnormal threshold; sending the closed-loop control instruction to an execution mechanism to adjust the process parameters, while updating the state data of the digital twin model, to realize closed-loop management of prediction and early warning control feedback.
2. The method of claim 1, wherein, The modeling and processing of the operation state data using the digital twin module to generate a digital twin model comprises: acquiring real-time data streams of smelting furnace temperature sensors, pouring mechanical arm flow meters, and cooling system rate monitoring devices, and performing preprocessing operations on the real-time data streams to eliminate noise data; constructing an initial digital twin model based on the preprocessed real-time data streams, the initial digital twin model including a smelting furnace heat transfer equation, a pouring mechanical arm flow distribution function, and a cooling system heat dissipation curve; continuously receiving new real-time data streams through the industrial internet, dynamically updating the initial digital twin model, and generating an optimized digital twin model; verifying the accuracy of the optimized digital twin model, and if the model error is less than a preset error threshold, using the optimized digital twin model as a final digital twin model for subsequent prediction analysis.
3. The method of claim 2, wherein, The generation of a closed-loop control instruction when detecting that the deviation between the actual operation parameters and the prediction result of the digital twin model exceeds the abnormal threshold comprises: calculating the deviation value between the actual operation parameters and the prediction result of the digital twin model, and determining whether the deviation value exceeds the abnormal threshold; if the deviation value exceeds the abnormal threshold, determining a corresponding closed-loop control strategy according to the deviation type, the closed-loop control strategy including heating device power adjustment rules and flow regulation valve opening adjustment rules; generating a closed-loop control instruction based on the closed-loop control strategy, and sending the closed-loop control instruction to an execution mechanism; recording the execution result of the closed-loop control instruction, and synchronizing the execution result to the digital twin model to update the model state data.
4. The method of claim 3, wherein, The generation of a closed-loop control instruction based on the closed-loop control strategy comprises: determining the heating device power adjustment amount or the flow regulation valve opening adjustment amount according to the size and direction of the deviation value; calling an industrial control protocol interface, converting the adjustment amount into a control signal conforming to an industrial communication standard; generating a closed-loop control instruction containing the control signal, and transmitting the closed-loop control instruction to the corresponding execution mechanism through the industrial internet; During the closed-loop control instruction execution process, the response state of the actuator is monitored in real time, and if the response state does not reach the expected target, the adjustment amount is recalculated and the closed-loop control instruction is updated.
5. The method of claim 2, wherein, The accuracy of the optimized digital twin model is verified, including: Extracting a test data set from historical operating state data, the test data set including smelting furnace temperature, pouring mechanical arm flow rate, and cooling system rate at multiple time points; Inputting the test data set into the optimized digital twin model to generate a predicted output; Calculating the error value between the predicted output and the actual test data, and if the error value is less than a preset error threshold, determining that the optimized digital twin model meets the accuracy requirement; If the error value is greater than or equal to the preset error threshold, further correcting the optimized digital twin model until the error value meets the accuracy requirement.
6. The method of claim 1, wherein, The real-time collection of operating state data in the physical production scene through the industrial internet includes: Establishing a communication connection between the industrial internet and the smelting furnace temperature sensor, the pouring mechanical arm flowmeter, and the cooling system rate monitoring device to ensure the real-time and reliability of data transmission; Configuring a data collection frequency to match the dynamic change characteristics of key process parameters in the casting process, the data collection frequency ranging from once per second to once per ten seconds, and the specific frequency being dynamically adjusted according to process requirements; Timestamping the received operating state data to ensure the time sequence consistency of the data, and storing the marked operating state data into a distributed database; Performing preliminary verification on the stored operating state data to eliminate abnormal values caused by sensor failure or signal interference, and generating a verified operating state data set; Based on the verified operating state data set, a real-time data stream pipeline is constructed to provide continuous and stable data input to the digital twin module.
7. The method of claim 2, wherein, The initial digital twin model is constructed based on the preprocessed real-time data stream, including: Analyzing the mathematical expression forms of the smelting furnace heat transfer equation, the pouring mechanical arm flow distribution function, and the cooling system heat dissipation curve to determine the correlation between the physical quantities; Using the preprocessed real-time data stream to perform parameter fitting on the correlation to generate an initial model parameter set; Based on the initial model parameter set, the framework structure of the initial digital twin model is constructed, including an input layer, a calculation layer, and an output layer; Embedding process parameter constraint conditions in the calculation layer, including the upper limit of the smelting furnace temperature, the lower limit of the pouring flow rate, and the maximum change gradient of the cooling rate; Preliminary verification of the initial digital twin model is performed by comparing the actual operating state data with the model output results to confirm that the basic prediction ability of the model meets the design requirements.
8. The method of claim 3, wherein, The execution results of the closed-loop control instruction are recorded and synchronized to the digital twin model to update the model state data, including: Receiving actual process parameter adjustment results fed back by the actuator, including the power change of the heating device and the opening change of the flow regulating valve; Performing time series analysis on the adjustment results to extract key change points in the process parameter adjustment process; The key change point is compared with the prediction result of the digital twin model, and a deviation value between the two is calculated; Based on the deviation value, the state data of the digital twin model is incrementally updated, which includes adjusting the model parameter weight and optimizing the model input-output mapping relationship; The updated digital twin model is quickly verified to ensure that the prediction accuracy of the model is within the preset error range, and if it exceeds the range, the model retraining process is triggered.
9. The method of claim 4, wherein, The industrial control protocol interface is called to convert the adjustment amount into a control signal that meets the industrial communication standard, including: Parsing the communication specification of the industrial control protocol interface to determine the data format, transmission rate, and encoding rules of the control signal; According to the specific numerical range of the adjustment amount, select the appropriate industrial control protocol type, including Modbus, OPC UA, or Profibus; Convert the adjustment amount according to the encoding rules of the selected protocol to generate a control signal that meets the industrial communication standard; Through the industrial internet, the control signal is transmitted to the target execution mechanism, and the integrity and accuracy of signal transmission are monitored in real time; If signal transmission failure or delay is detected beyond the preset threshold, a signal retransmission mechanism is triggered to ensure reliable delivery of the control signal.
10. The method of claim 5, wherein, The optimized digital twin model is further corrected, including: Extracting a supplementary test data set from historical operating state data, which covers multiple abnormal working condition scenarios; Input the supplementary test data set into the optimized digital twin model to generate a supplementary prediction output; Calculate the error value between the supplementary prediction output and the actual test data to determine the main source of the error value; Based on the main source of the error value, adjust the parameter optimization strategy of the digital twin model, including increasing the number of training samples, introducing a nonlinear mapping function, or adjusting the model convergence condition; Repeat the correction and verification process until the error value meets the preset error threshold to ensure that the optimized digital twin model has high-precision prediction capability.