A remote monitoring system for heat treatment furnace groups based on digital twins

Abstract

This invention relates to the field of industrial control and remote monitoring technology, and particularly to a remote monitoring system for heat treatment furnace groups based on digital twins. The invention deploys data acquisition units at each physical furnace and its supporting equipment, periodically collecting and preprocessing multi-source data; a digital twin modeling unit constructs a twin model containing at least a physical model and a behavioral model, and periodically iterates and optimizes parameters; a model update unit transmits the preprocessed data to the twin model in real time through an edge computing architecture with a preset compression rate to complete the update; a control unit, using a predictive model, determines the faults and their types within a preset time period based on the updated model data and generates process parameter optimization instructions; an analysis unit determines the monitoring status based on the error characterization values ​​of the standardized errors of key process points, thereby adjusting the model iteration optimization cycle and adapting the amount of historical operating data collected according to the adjusted monitoring status. This invention improves the operation and maintenance management efficiency of furnace groups.

Description

Technical Field

[0001] This invention relates to the field of industrial control and remote monitoring technology, and in particular to a remote monitoring system for heat treatment furnace groups based on digital twins. Background Technology

[0002] Heat treatment furnaces are key equipment in industries such as machinery manufacturing and metallurgy, and their operating status directly affects product quality, performance, and production efficiency. With the large-scale development of industrial production, the application of heat treatment furnace groups is becoming increasingly widespread, placing higher demands on the monitoring and management of these furnace groups.

[0003] Existing heat treatment furnace group monitoring systems mostly adopt the traditional distributed monitoring mode, which collects operational data by deploying sensors in each furnace and transmits it to the monitoring center for display and analysis via wired or wireless networks. However, these systems have the following shortcomings: First, they lack accurate digital mapping of the physical entities of the furnace group, making it impossible to intuitively and comprehensively reflect the operating status and internal changes of the furnaces, making it difficult for maintenance personnel to quickly understand the operating mechanism of the furnace group; second, their data processing and analysis capabilities are limited, mainly focusing on data display and simple alarms, and failing to achieve predictive analysis of the furnace group's operating status and optimization suggestions for process parameters; third, the coordination between monitoring and control is insufficient, making it difficult to achieve precise closed-loop control based on the overall operating status of the furnace group, and the remote operation and maintenance and fault diagnosis capabilities are weak, resulting in untimely handling of equipment failures and affecting production continuity.

[0004] Chinese Patent Publication No. CN112947327A discloses an intelligent monitoring and management system for industrial furnace groups based on WINCC, including a processor. The processor is communicatively connected to a furnace temperature detection module, an environmental detection module, a hazard analysis module, a risk avoidance analysis module, an early warning module, a controller, and a data center. The early warning module includes purple warning lights, yellow warning lights, and red warning lights. The risk avoidance analysis module is used to plan escape routes for furnace group workers.

[0005] It is evident that existing technologies have the following problems: the monitoring system for heat treatment furnace groups suffers from low digitalization, weak analytical capabilities, and poor collaborative control, resulting in low efficiency in the operation and maintenance management of the furnace groups. Summary of the Invention

[0006] To address this, the present invention provides a remote monitoring system for heat treatment furnace groups based on digital twins, which overcomes the problems of low digitization, weak analytical capabilities, and poor collaborative control in existing heat treatment furnace group monitoring systems, resulting in low efficiency in the operation and maintenance management of the furnace groups.

[0007] To achieve the above objectives, the present invention provides a remote monitoring system for heat treatment furnace groups based on digital twins, comprising:

[0008] The data acquisition unit is deployed on each physical furnace body and supporting equipment of the heat treatment furnace group to periodically collect furnace body operating status data, heat treatment process parameter data and environmental data, as well as to preprocess the collected multi-source data.

[0009] A digital twin modeling unit, connected to the data acquisition unit, is used to construct a digital twin model of the heat treatment furnace group. The digital twin model includes at least a physical model and a behavioral model. The physical model is constructed based on the finite element analysis method, and the behavioral model is constructed based on historical furnace operating status data and machine learning algorithms. The parameters of the digital twin model are periodically iteratively optimized.

[0010] The model update unit is connected to the data acquisition unit and the digital twin modeling unit respectively, and is used to transmit the pre-processed multi-source data to the digital twin model in real time using an edge computing architecture with a preset compression rate, so as to update the digital twin model in real time.

[0011] The control unit, which is connected to the model update unit, is used to determine whether a fault occurs and the type of fault within a preset time period using the predictive model based on the operating status data of the updated digital twin model, and to generate process parameter optimization instructions based on the fault type.

[0012] An analysis unit, connected to the control unit, is used to determine the monitoring status based on the error characterization value determined by the standardized error of multiple key process points, and to adjust the iterative optimization cycle of the parameters of the digital twin model based on the monitoring status, and to adjust the number of furnace operating status data collected at historical moments based on the monitoring status after adjusting the iterative optimization cycle of the parameters of the digital twin model.

[0013] Furthermore, the digital twin modeling unit also includes a model adaptive update module, which is used to calculate the model error based on the simulation results of the digital twin model and the state operation data of the physical furnace group, and to iteratively optimize the parameters of the digital twin model using the gradient descent algorithm.

[0014] Furthermore, the data acquisition unit also includes a sampling period adjustment module, which is used to adjust the acquisition period of multi-source data based on the temperature change data of the furnace body within a preset time.

[0015] Furthermore, the analysis unit is also used to plot a time-model error curve based on the model errors corresponding to multiple historical moments when the monitoring status is unqualified; the analysis unit is also used to calculate the integral of the curve, and if the integral is greater than a preset integral, adjust the iterative optimization cycle of the parameters of the digital twin model based on the ratio of the integral to the preset integral; wherein, if the error characterization value is greater than a preset error, the monitoring status is determined to be unqualified.

[0016] Furthermore, the analysis unit is also used to reduce the iterative optimization cycle of the parameters of the digital twin model based on the ratio of the integral to the preset integral, and the reduction in the iterative optimization cycle is proportional to the ratio.

[0017] Furthermore, the analysis unit is also used to obtain the prediction residual sequence of the behavior model within a preset time period when the monitoring status is unqualified after adjusting the iterative optimization cycle of the parameters of the digital twin model, and calculate the time series autocorrelation coefficient of the prediction residual of the behavior model based on the prediction residual sequence; the analysis unit is also used to adjust the amount of furnace operation status data at historical moments based on the ratio of the time series autocorrelation coefficient to the preset threshold when the time series autocorrelation coefficient is greater than the preset threshold.

[0018] Furthermore, the analysis unit is also used to increase the number of furnace operating status data at historical times based on the ratio of the time series autocorrelation coefficient to the preset threshold, and the increase in the number of furnace operating status data at historical times is proportional to the ratio.

[0019] Furthermore, the analysis unit is also used to, if the monitoring status is unqualified after adjusting the number of furnace operating status data at historical time moments, repeatedly adjust the number of furnace operating status data at historical time moments at least once, until the number of adjustments is less than a preset number and the monitoring status is qualified, or the number of adjustments is equal to the preset number, and then stop adjusting; the analysis unit is also used to, if the monitoring status is unqualified after stopping the adjustment, calculate the average end-to-end latency of multiple data packets, and if the latency is greater than a preset latency, adjust the preset compression rate based on the ratio of the latency to the preset latency.

[0020] Furthermore, the analysis unit is also used to increase the preset compression ratio based on the ratio of the latency to the preset latency, and the increase in the preset compression ratio is proportional to the ratio.

[0021] Furthermore, the analysis unit is also used to calculate the difference between the error characterization value and the preset error when the monitoring status is unqualified after adjusting the preset compression rate; the analysis unit is also used to reduce the acquisition cycle of multi-source data based on the ratio of the difference to the preset difference when the difference is greater than the preset difference, and the reduction in the acquisition cycle is proportional to the ratio.

[0022] Compared with existing technologies, the advantages of this invention are as follows: This invention deploys data acquisition units in each physical furnace and supporting equipment, periodically collecting and preprocessing multi-source data; a digital twin modeling unit constructs a twin model containing at least a physical model and a behavioral model, and periodically iterates and optimizes parameters; a model update unit transmits the preprocessed data to the twin model in real time through an edge computing architecture with a preset compression rate to complete the update; the control unit, using a predictive model, determines the faults and their types within a preset time period based on the updated model data and generates process parameter optimization instructions; the analysis unit determines the monitoring status based on the error characterization value of the standardized error of key process points, thereby adjusting the model iteration optimization cycle and adapting the amount of historical operating data collected according to the adjusted monitoring status. This invention improves the operation and maintenance management efficiency of the furnace group.

[0023] Furthermore, this invention determines the model error based on the model adaptive update module, and uses the gradient descent algorithm to iteratively optimize the parameters of the digital twin model based on the model error. This enables the digital twin model and the physical furnace group to be updated synchronously, allowing for more accurate monitoring of the operating status of the physical furnace group, thereby further improving the efficiency of furnace group operation and maintenance management.

[0024] Furthermore, this invention adjusts the acquisition cycle of multi-source data based on the temperature change data of the furnace body within a preset time, which can more effectively collect multi-source data, thereby enabling the digital twin model to be more effectively synchronized with the physical furnace group, and further improving the operation and maintenance management efficiency of the furnace group.

[0025] Furthermore, this invention determines the cause of unqualified monitoring status based on the integral of the time-model error curve, which can more accurately determine the cause of unqualified monitoring status, thereby enabling more effective adjustment of relevant parameters and further improving the operation and maintenance management efficiency of the furnace group.

[0026] Furthermore, this invention adjusts the iterative optimization cycle of the parameters of the digital twin model based on the ratio of the integral to the preset integral, which enables the digital twin model to more accurately simulate the actual behavior of the physical furnace group, thereby enabling the digital twin model to synchronize with the physical furnace group more effectively, and further improving the operation and maintenance management efficiency of the furnace group.

[0027] Furthermore, this invention determines the reasons for unqualified monitoring status based on the time series autocorrelation coefficient of the residual predicted by the behavioral model. This can more accurately determine the reasons for unqualified monitoring status, thereby enabling more effective adjustments based on the reasons and further improving the operation and maintenance management efficiency of the furnace group.

[0028] Furthermore, the present invention adjusts the amount of furnace operating status data at historical moments based on the ratio of the time series autocorrelation coefficient to the preset threshold, which enables the behavioral model to predict more accurately, further fully capture the dynamic patterns in the time series, thereby making the results output by the machine learning algorithm more accurate, and further improving the operation and maintenance management efficiency of the furnace group.

[0029] Furthermore, this invention determines the cause of monitoring status failure based on the average end-to-end latency of multiple data packets, which can more accurately identify the cause and enable more effective adjustments to be made based on the cause, thereby further improving the operation and maintenance management efficiency of the furnace group.

[0030] Furthermore, the present invention adjusts the preset compression rate based on the ratio of latency to preset latency, which enables the edge computing architecture to transmit more effectively, further reducing transmission latency and thus further improving the operation and maintenance management efficiency of the furnace group.

[0031] Furthermore, this invention adjusts the acquisition cycle of multi-source data based on the ratio of the difference between the error characterization value and the preset error, which can more effectively acquire multi-source data, thereby making the standardized error more accurate, further making the determined monitoring status more precise, and thus further improving the operation and maintenance management efficiency of the furnace group. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the structure of the remote monitoring system for heat treatment furnace groups based on digital twins according to an embodiment of the present invention;

[0033] Figure 2 This is a flowchart illustrating the steps of a remote monitoring method for heat treatment furnace groups based on digital twins, as described in an embodiment of the present invention.

[0034] Figure 3 This is a flowchart illustrating the steps of determining the error characterization value based on the comparison result with the preset error in an embodiment of the present invention.

[0035] Figure 4 This is a flowchart illustrating the steps of determining the monitoring status based on adjusting the amount of historical furnace operating status data in an embodiment of the present invention. Detailed Implementation

[0036] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0037] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0038] It should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0039] Please see Figure 1 As shown, it is a schematic diagram of the structure of the remote monitoring system for heat treatment furnace groups based on digital twins according to an embodiment of the present invention.

[0040] The system includes a data acquisition unit, a digital twin modeling unit, a model update unit, a control unit, and an analysis unit.

[0041] The data acquisition unit is deployed on each physical furnace body and supporting equipment of the heat treatment furnace group to periodically collect furnace body operating status data, heat treatment process parameter data and environmental data, as well as to preprocess the collected multi-source data.

[0042] The digital twin modeling unit is connected to the data acquisition unit and is used to construct a digital twin model of the heat treatment furnace group. The digital twin model includes at least a physical model and a behavioral model. The physical model is constructed based on the finite element analysis method, and the behavioral model is constructed based on historical furnace operating status data and machine learning algorithms. The parameters of the digital twin model are periodically iteratively optimized.

[0043] The model update unit is connected to the data acquisition unit and the digital twin modeling unit respectively. It is used to transmit the pre-processed multi-source data to the digital twin model in real time using an edge computing architecture with a preset compression rate, so as to update the digital twin model in real time.

[0044] The control unit is connected to the model update unit, which is used to determine whether a fault occurs and the type of fault within a preset time period using the predictive model based on the operating status data of the updated digital twin model, and to generate process parameter optimization instructions based on the fault type.

[0045] The analysis unit is connected to the control unit. It is used to determine the monitoring status based on the error characterization value determined by the standardized error of multiple key process points, and to adjust the iterative optimization cycle of the parameters of the digital twin model based on the monitoring status. It also adjusts the number of furnace operation status data collected at historical moments based on the monitoring status after adjusting the iterative optimization cycle of the parameters of the digital twin model.

[0046] Specifically, the data acquisition unit includes a temperature sensor, a pressure sensor, a vibration sensor, a current sensor, a voltage sensor, and an ambient temperature and humidity sensor. Each sensor has a data preprocessing function, which can perform noise reduction, filtering, and format standardization on the acquired raw data.

[0047] Specifically, the digital twin model includes four core sub-models: (1) Geometric model: Based on the three-dimensional design drawings, CAD models and laser scanning data of the furnace body, it is constructed using three-dimensional modeling software to accurately restore the geometric features of the furnace body, such as the external structure, internal cavity, heating elements and insulation layer, with a geometric accuracy error of no more than 0.1 mm; (2) Physical model: Based on theories such as heat transfer and fluid mechanics, it is established using finite element analysis software to simulate the distribution of physical fields such as temperature field, pressure field and flow field inside the furnace body, and accurately reflect the heat transfer and mass transfer process of the furnace body; (3) Behavioral model: Based on the historical operation data of the furnace group, it is constructed using the LSTM (Long Short-Term Memory Network) machine learning algorithm to reproduce the operating behavior of the furnace body under different process parameters, such as temperature change curves and pressure fluctuation patterns; (4) Rule model: Integrating heat treatment process specifications, equipment operation and maintenance standards, safety operation guidelines and related industry standards, a rule base is constructed to provide a basis for furnace group operation monitoring, fault diagnosis and process optimization.

[0048] Specifically, the edge computing architecture mainly consists of an edge gateway, a 5G communication module, and a data encryption unit. Data preprocessed by the data acquisition unit is first transmitted to the edge gateway, where it performs local aggregation, caching, and preliminary analysis of the data, filtering out invalid data and reducing subsequent transmission pressure. Then, the valid data is transmitted in real-time to the digital twin modeling module and the remote monitoring center via the 5G communication module. 5G communication technology ensures high bandwidth and low latency data transmission, meeting the needs of real-time monitoring and interaction. The data encryption unit uses the AES-256 encryption algorithm to encrypt the transmitted data throughout the entire process, preventing data theft and tampering during transmission and ensuring data transmission security.

[0049] Specifically, the control unit has simulation analysis capabilities, which can simulate and analyze the operating status of the furnace group through a digital twin model, predict the changing trends of parameters such as furnace temperature and pressure, and include fault diagnosis. This process is based on deep learning algorithms and combines historical fault data and real-time operating data of the furnace to automatically identify common faults such as furnace heating element damage, insulation layer failure, and fan failure, and can accurately locate the fault location, providing fault handling suggestions for maintenance personnel.

[0050] Specifically, the model adaptive update module, based on the error between the current model output and the actual observation data, uses a gradient descent algorithm to repeatedly adjust the model's internal parameters, such as weights, coefficients, or structural parameters, along the direction of error reduction with a set step size. This process, through multiple iterative feedback loops, gradually reduces the deviation between the model simulation results and the actual state of the physical system, ensuring that the dynamic characteristics of the digital twin model continuously approximate the actual operating patterns of the furnace group. This process is existing technology and will not be elaborated upon here.

[0051] Please see Figure 2 The diagram shown is a flowchart illustrating the steps of a remote monitoring method for a heat treatment furnace group based on digital twins, according to an embodiment of the present invention.

[0052] The specific steps for remote monitoring of heat treatment furnace groups based on digital twins are as follows:

[0053] S1 periodically collects furnace operating status data, heat treatment process parameter data, and environmental data through data acquisition units deployed on each physical furnace body and supporting equipment of the heat treatment furnace group, and preprocesses the collected multi-source data.

[0054] S2, A digital twin model of the heat treatment furnace group is constructed through a digital twin modeling unit connected to the data acquisition unit. The digital twin model includes at least a physical model and a behavioral model. The physical model is constructed based on the finite element analysis method, and the behavioral model is constructed based on historical furnace operating status data and machine learning algorithms. The parameters of the digital twin model are periodically iteratively optimized.

[0055] S3, the model update unit, which is connected to the data acquisition unit and the digital twin modeling unit respectively, uses an edge computing architecture with a preset compression rate to transmit the preprocessed multi-source data to the digital twin model in real time, so as to update the digital twin model in real time.

[0056] S4, the control unit connected to the model update unit uses the prediction model based on the updated digital twin model's operating status data to determine whether a fault occurs within a preset time period and the fault type, and generates process parameter optimization instructions based on the fault type.

[0057] S5, the monitoring status is determined by the analysis unit connected to the control unit based on the error characterization value determined by the standardized error of multiple key process points, and the iterative optimization cycle of the parameters of the digital twin model is adjusted based on the monitoring status, and the number of furnace operation status data collected at historical moments is adjusted based on the monitoring status after adjusting the iterative optimization cycle of the parameters of the digital twin model.

[0058] Please see Figure 3 As shown, it is a flowchart of the steps for determining the error characterization value based on the comparison result with the preset error in an embodiment of the present invention.

[0059] Specifically, taking the remote monitoring system for heat treatment furnace groups as an example, and based on the hardware measurement accuracy, response speed limit, and quality tolerance requirements of the furnace group sensors and actuators, as well as the massive amount of operating condition data and optimization records accumulated by combining historical production process statistics and fault diagnosis analysis, the corresponding preset or critical parameter values ​​are set.

[0060] Specifically, the error characterization value is the maximum value of the standardized error of multiple key process points. With a preset error L0 = 1, the comparison process between the error characterization value L and the preset error L0 is as follows:

[0061] If the error characterization value L is less than or equal to the preset error L0, then the monitoring status is determined to be qualified.

[0062] If the error characterization value L is greater than the preset error L0, then the monitoring status is determined to be unqualified.

[0063] Specifically, when the monitoring status is unqualified, a time-model error curve is plotted based on the model errors corresponding to multiple historical moments. The integral of the curve is calculated. If the integral is greater than the preset integral, it indicates that the model error calculated by the adaptive update module is consistently high, suggesting that the physical or behavioral model in the digital twin model cannot accurately simulate the actual behavior of the physical furnace group. This leads to the analysis unit deriving unreliable standardized errors based on the erroneous model. Therefore, the iterative optimization cycle of the digital twin model parameters is adjusted based on the ratio of the integral to the preset integral. The preset ratio of the integral to the preset integral is P0 = 1.2. The comparison process between the integral and the preset integral ratio P and the preset ratio P0 is as follows:

[0064] If the ratio P of the integral to the preset integral is less than or equal to the preset ratio P0, the iterative optimization period of the parameters of the digital twin model will be adjusted to 0.91 times the original iterative optimization period, where the adjusted value will be rounded up.

[0065] If the ratio P of the integral to the preset integral is greater than the preset ratio P0, the iterative optimization period of the parameters of the digital twin model will be adjusted to 0.79 times the original iterative optimization period, where the adjusted value will be rounded up.

[0066] Specifically, if the monitoring status is unqualified after adjusting the iterative optimization cycle of the digital twin model's parameters, the predicted residual sequence within a preset time period is obtained, and the time series autocorrelation coefficient of the behavioral model's predicted residual is calculated based on the predicted residual sequence. If the time series autocorrelation coefficient is greater than a preset threshold, it indicates that the behavioral model has a systematic prediction bias and has failed to fully capture the dynamic patterns in the time series, meaning the machine learning algorithm's output is inaccurate. Therefore, the amount of furnace operating status data at historical moments is adjusted based on the ratio of the time series autocorrelation coefficient to the preset threshold. The preset ratio of the time series autocorrelation coefficient to the preset threshold is Q0 = 1.5. The comparison process between the ratio Q of the time series autocorrelation coefficient to the preset threshold and the preset ratio Q0 is as follows:

[0067] If the ratio Q of the time series autocorrelation coefficient to the preset threshold is less than or equal to the preset ratio Q0, then the number of furnace operating status data at historical moments will be adjusted to 1.9 times the original number, where the adjusted value will be rounded up.

[0068] If the ratio Q of the time series autocorrelation coefficient to the preset threshold is greater than the preset ratio Q0, the number of furnace operating status data at historical moments will be adjusted to 2.5 times the original number, with the adjusted value rounded up.

[0069] Please see Figure 4 The diagram shows the steps of determining the monitoring status based on adjusting the amount of furnace operating status data at historical times, according to an embodiment of the present invention.

[0070] Specifically, if the monitoring status is unqualified after adjusting the number of historical furnace operating status data, the number of historical furnace operating status data should be adjusted at least once until the number of adjustments is less than the preset number and the monitoring status is qualified, or the number of adjustments is equal to the preset number. If the monitoring status is still unqualified after stopping the adjustment, the average end-to-end latency of multiple data packets is calculated. If the latency is greater than the preset latency, it indicates that there is a delay in data transmission or processing under the edge computing architecture, causing the digital Li Sheng model to fail to update in real time. The standardized error calculated by the analysis unit based on outdated model data cannot reflect the current actual status. Then, the preset compression ratio is adjusted based on the ratio of latency to preset latency. The preset ratio of latency to preset latency is R0 = 1.3. The comparison process between the ratio R of latency to preset latency and the preset ratio R0 is as follows:

[0071] If the ratio R of the delay to the preset delay is less than or equal to the preset ratio R0, the preset compression ratio will be adjusted to 1.43 times the original preset compression ratio, and the adjusted value will be retained to one decimal place.

[0072] If the ratio R of the delay to the preset delay is greater than the preset ratio R0, the preset compression ratio will be adjusted to 1.83 times the original preset compression ratio, and the adjusted value will be retained to one decimal place.

[0073] Specifically, if the monitoring status is unqualified after adjusting the preset compression ratio, the difference between the error characterization value and the preset error is calculated. If the difference is greater than the preset difference, it indicates that the acquisition frequency was not dynamically and accurately adjusted according to the furnace temperature change rate. During rapid process changes, such as heating and cooling, there are insufficient data points, resulting in the calculated standardized error failing to reflect the true deviation. Therefore, the acquisition cycle of multi-source data is adjusted based on the ratio of the difference to the preset difference. The preset ratio of the difference to the preset difference is T0 = 1.35. The comparison process between the ratio T0 and the preset difference is as follows:

[0074] If the ratio T of the difference to the preset difference is less than or equal to the preset ratio T0, the acquisition period of the multi-source data will be adjusted to 0.92 times the original acquisition period, and the adjusted value will be rounded up.

[0075] If the ratio T of the difference to the preset difference is greater than the preset ratio T0, the acquisition period of the multi-source data will be adjusted to 0.85 times the original acquisition period, and the adjusted value will be rounded up.

[0076] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0077] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A remote monitoring system for heat treatment furnace groups based on digital twins, characterized in that, include: The data acquisition unit is deployed on each physical furnace body and supporting equipment of the heat treatment furnace group to periodically collect furnace body operating status data, heat treatment process parameter data and environmental data, as well as to preprocess the collected multi-source data. A digital twin modeling unit, connected to the data acquisition unit, is used to construct a digital twin model of the heat treatment furnace group. The digital twin model includes at least a physical model and a behavioral model. The physical model is constructed based on the finite element analysis method, and the behavioral model is constructed based on historical furnace operating status data and machine learning algorithms. The parameters of the digital twin model are periodically iteratively optimized. The model update unit is connected to the data acquisition unit and the digital twin modeling unit respectively, and is used to transmit the pre-processed multi-source data to the digital twin model in real time using an edge computing architecture with a preset compression rate, so as to update the digital twin model in real time. The control unit, which is connected to the model update unit, is used to determine whether a fault occurs and the type of fault within a preset time period using the predictive model based on the operating status data of the updated digital twin model, and to generate process parameter optimization instructions based on the fault type. An analysis unit, connected to the control unit, is used to determine the monitoring status based on the error characterization value determined by the standardized error of multiple key process points, and to adjust the iterative optimization cycle of the parameters of the digital twin model based on the monitoring status, and to adjust the number of furnace operation status data collected at historical moments based on the monitoring status after adjusting the iterative optimization cycle of the parameters of the digital twin model, wherein the error characterization value is the maximum value of the standardized error of multiple key process points. The analysis unit is also used to plot a time-model error curve based on the model error corresponding to multiple historical time points when the monitoring status is unqualified. The analysis unit is also used to calculate the integral of the curve, and if the integral is greater than a preset integral, adjust the iterative optimization cycle of the parameters of the digital twin model based on the ratio of the integral to the preset integral. Wherein, if the error characterization value is greater than the preset error, the monitoring status is determined to be unqualified; The analysis unit is also used to obtain the prediction residual sequence of the behavior model within a preset time period when the monitoring status is unqualified after adjusting the iterative optimization cycle of the parameters of the digital twin model, and to calculate the time series autocorrelation coefficient of the prediction residual of the behavior model based on the prediction residual sequence. The analysis unit is also used to adjust the amount of furnace operating status data at historical moments based on the ratio of the time series autocorrelation coefficient to the preset threshold when the time series autocorrelation coefficient is greater than the preset threshold.

2. The remote monitoring system for heat treatment furnace groups based on digital twins according to claim 1, characterized in that, The digital twin modeling unit also includes a model adaptive update module, which is used to calculate the model error based on the simulation results of the digital twin model and the status operation data of the physical furnace group, and to iteratively optimize the parameters of the digital twin model using the gradient descent algorithm.

3. The remote monitoring system for heat treatment furnace groups based on digital twins according to claim 2, characterized in that, The data acquisition unit also includes a sampling period adjustment module, which is used to adjust the acquisition period of multi-source data based on the temperature change data of the furnace body within a preset time.

4. The remote monitoring system for heat treatment furnace groups based on digital twins according to claim 1, characterized in that, The analysis unit is also used to reduce the iterative optimization cycle of the parameters of the digital twin model based on the ratio of the integral to the preset integral, and the reduction in the iterative optimization cycle is proportional to the ratio.

5. The remote monitoring system for heat treatment furnace groups based on digital twins according to claim 1, characterized in that, The analysis unit is also used to increase the number of furnace operating status data at historical times based on the ratio of the time series autocorrelation coefficient to the preset threshold, and the increase in the number of furnace operating status data at historical times is proportional to the ratio.

6. The remote monitoring system for heat treatment furnace groups based on digital twins according to claim 5, characterized in that, The analysis unit is also used to, if the monitoring status is not qualified after adjusting the number of furnace operating status data at historical time, repeatedly adjust the number of furnace operating status data at historical time at least once, until the number of adjustments is less than the preset number and the monitoring status is qualified or the number of adjustments is equal to the preset number, and then stop adjusting. The analysis unit is also used to calculate the average end-to-end latency of multiple data packets when the monitoring status is unqualified after the adjustment is stopped, and to adjust the preset compression rate based on the ratio of the latency to the preset latency when the latency is greater than the preset latency.

7. The remote monitoring system for heat treatment furnace groups based on digital twins according to claim 6, characterized in that, The analysis unit is also used to increase the preset compression ratio based on the ratio of the latency to the preset latency, and the increase in the preset compression ratio is proportional to the ratio.

8. The remote monitoring system for heat treatment furnace groups based on digital twins according to claim 7, characterized in that, The analysis unit is also used to calculate the difference between the error characterization value and the preset error when the monitoring status is unqualified after adjusting the preset compression rate. The analysis unit is also used to reduce the acquisition cycle of multi-source data based on the ratio of the difference to the preset difference when the difference is greater than the preset difference, and the reduction in the acquisition cycle is proportional to the ratio.

Images