Parameter adjustment and prediction model acquisition method, device and storage medium

Abstract

This disclosure provides a method, apparatus, and storage medium for parameter adjustment and prediction model acquisition, relating to artificial intelligence fields such as deep learning and big data processing, and applicable to various scenarios requiring intelligent temperature control and intelligent thermal management. The method may include: during the heating process of a heated object using heating equipment, determining the production stage to be entered and acquiring a prediction model corresponding to the production stage; using the prediction model to predict the parameter adjustment method for the production stage, wherein the parameters are predetermined heating-related parameters; and adjusting the parameters of the production stage according to the parameter adjustment method. Applying the solution described in this disclosure can improve the accuracy of the adjustment results, etc.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to methods, apparatus and storage media for parameter adjustment and prediction model acquisition in the fields of deep learning and big data processing. Background Technology

[0002] Galvanized steel strip is widely used in industries such as automobiles, home appliances, construction, and agricultural machinery. Continuous hot-dip galvanizing technology is the main production process for galvanized steel strip. In the continuous hot-dip galvanizing production process, cold-rolled or hot-rolled steel strip runs continuously on the production line at a certain speed, and undergoes two main processes: annealing and hot-dip galvanizing.

[0003] Among them, the temperature of the annealing furnace and the running speed of the steel strip determine the temperature of the steel strip at the outlet of the annealing furnace, which in turn directly affects the mechanical properties of the steel strip after annealing, the adhesion and surface quality of the hot-dip galvanized layer, the production line capacity, and the energy consumption per unit capacity. Therefore, the temperature of the annealing furnace and the running speed of the steel strip are the most critical control factors in the entire production process. Summary of the Invention

[0004] This disclosure provides a method, apparatus, and storage medium for parameter adjustment and prediction model acquisition.

[0005] A parameter adjustment method, comprising:

[0006] During the process of heating an object using heating equipment, the production stage to be entered is determined, and the prediction model corresponding to the production stage is obtained.

[0007] The prediction model is used to predict the parameter adjustment method for the production stage, where the parameters are predetermined heating-related parameters;

[0008] The parameters for the production stage are adjusted according to the parameter adjustment method described above.

[0009] A method for obtaining a prediction model, comprising:

[0010] For a predetermined production stage in the process of heating an object using a heating device, training data corresponding to the production stage is obtained, and the training data is training data generated based on real historical production data.

[0011] A prediction model corresponding to the production stage is trained using the training data. The prediction model is used to predict the parameter adjustment method of the production stage during the heating process so that the parameters of the production stage can be adjusted according to the parameter adjustment method. The parameters are predetermined heating-related parameters.

[0012] A parameter adjustment device includes: a first acquisition module, a prediction module, and an adjustment module;

[0013] The first acquisition module is used to determine the production stage to be entered during the process of heating the object using a heating device, and to acquire the prediction model corresponding to the production stage.

[0014] The prediction module is used to predict the parameter adjustment method of the production stage using the prediction model, wherein the parameters are predetermined heating-related parameters;

[0015] The adjustment module is used to adjust the parameters in the production stage according to the parameter adjustment method.

[0016] A predictive model acquisition device includes: a second acquisition module and a training module;

[0017] The second acquisition module is used to acquire training data corresponding to a predetermined production stage in the process of heating an object using a heating device. The training data is training data generated based on real historical production data.

[0018] The training module is used to train a prediction model corresponding to the production stage using the training data. The prediction model is used to predict the parameter adjustment method of the production stage during the heating process so that the parameters of the production stage can be adjusted according to the parameter adjustment method. The parameters are predetermined heating-related parameters.

[0019] An electronic device, comprising:

[0020] At least one processor; and

[0021] A memory communicatively connected to the at least one processor; wherein,

[0022] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described above.

[0023] A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the methods described above.

[0024] A computer program product includes a computer program / instructions that, when executed by a processor, implement the method described above.

[0025] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0026] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0027] Figure 1 This is a flowchart of an embodiment of the parameter adjustment method described in this disclosure;

[0028] Figure 2 This is a flowchart of an embodiment of the prediction model acquisition method described in this disclosure;

[0029] Figure 3 This is a schematic diagram of the physical information neural network architecture of the first prediction model of this disclosure;

[0030] Figure 4 This is a schematic diagram of the physical information neural network architecture of the second prediction model of this disclosure;

[0031] Figure 5 This is a schematic diagram of the composition structure of Embodiment 500 of the parameter adjustment device described in this disclosure;

[0032] Figure 6 This is a schematic diagram of the composition structure of Embodiment 600 of the prediction model acquisition device described in this disclosure;

[0033] Figure 7 A schematic block diagram of an electronic device 700 that can be used to implement embodiments of the present disclosure is shown. Detailed Implementation

[0034] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0035] Furthermore, it should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0036] Figure 1 This is a flowchart illustrating an embodiment of the parameter adjustment method described in this disclosure. Figure 1 As shown, the specific implementation methods are as follows.

[0037] In step 101, during the heating process of the object being heated using a heating device, the production stage to be entered is determined, and the prediction model corresponding to the production stage is obtained.

[0038] In step 102, the prediction model is used to predict the parameter adjustment method for the production stage, where the parameters are predetermined heating-related parameters.

[0039] In step 103, the parameters of the production stage are adjusted according to the parameter adjustment method.

[0040] In one embodiment of this disclosure, the heating device may be an annealing furnace, the object to be heated may be a steel strip, and the parameters may be the furnace temperature of the open flame heating section of the annealing furnace and the running speed of the steel strip.

[0041] In the traditional method, the temperature of the annealing furnace (such as the furnace temperature of the open flame heating section of the annealing furnace mentioned above) and the running speed of the steel strip are mainly adjusted by human experience, which has poor accuracy.

[0042] The scheme described in the embodiments of this disclosure can use a predictive model to predict the parameter adjustment method, such as the adjustment method of the furnace temperature of the open flame heating section and the running speed of the steel strip. Then, the furnace temperature of the open flame heating section and the running speed of the steel strip can be adjusted accordingly, thereby improving the accuracy of the adjustment results, eliminating the reliance on manual experience, and improving the product qualification rate and product performance stability.

[0043] Furthermore, the heating equipment and heating objects described in this disclosure are not limited to annealing furnaces and steel strips; they are equally applicable to other heating equipment and heating objects.

[0044] The solution described in this disclosure is applicable to various scenarios requiring intelligent temperature control and intelligent thermal management. For example, the solution described in this disclosure is applicable to intelligent temperature control of continuous hot-dip galvanizing annealing furnaces in the metallurgical industry, intelligent temperature control of industrial furnaces (smelting furnaces, melting furnaces, calcining furnaces, heating furnaces, heat treatment furnaces, drying furnaces, coke ovens, gas generators, etc.), intelligent temperature control of petrochemical reactors, intelligent temperature control of power plant heating equipment, and intelligent thermal management of spacecraft equipment. Accordingly, the specific forms of heating equipment and heating objects will differ in different scenarios.

[0045] The following example, taking an annealing furnace as the heating equipment, a steel strip as the heating object, and the furnace temperature and steel strip running speed of the open flame heating section of the annealing furnace as parameters, will further illustrate the scheme described in this disclosure.

[0046] In one embodiment of this disclosure, the production stage may include a steady-state stage in which the steel strip specifications do not change. Accordingly, the method of predicting the parameter adjustment of the production stage using a prediction model includes: using a first prediction model corresponding to the steady-state stage to predict the parameter value of the steady-state stage. Adjusting the parameter of the production stage may include: adjusting the parameter to the parameter value.

[0047] In one embodiment of this disclosure, the steel strip specifications in the steady-state phase can be used as input to a first prediction model to obtain the output parameter values.

[0048] In one embodiment of this disclosure, the production stage may further include a transition stage where the steel strip specifications change. Accordingly, predicting the parameter adjustment method for the production stage using a predictive model may include: using a second predictive model corresponding to the transition stage to predict the dynamic adjustment curve of the parameters in the transition stage; adjusting the parameters in the production stage may include: adjusting the parameters in the transition stage in real time according to the dynamic adjustment curve. The dynamic adjustment curve refers to the curve showing the change of furnace temperature and steel strip running speed in the incineration heating section over time.

[0049] In one embodiment of this disclosure, the operation point (OP) variables of the steady-state stage before the transition stage and the operation point variables of the steady-state stage after the transition stage can be obtained. The operation point variables include: steel strip specifications, steel strip running speed and furnace temperature of open flame heating section. The obtained operation point variables are used as input to the second prediction model to obtain the output dynamic adjustment curve.

[0050] In actual production, the process mainly includes the two production stages mentioned above: the steady-state stage and the transition stage. The steady-state stage refers to the stage where the specifications of the steel strip do not change, and the annealing furnace operates stably during this stage. The transition stage refers to the stage where the specifications of the steel strip change. For example, if the specifications of the previous steel strip are different from those of the next steel strip, the two steel strips can be connected by welding. Theoretically, the process / stage from the weld point entering the annealing furnace to the weld point leaving the annealing furnace can be considered as the transition stage. However, in practice, the moment when the distance between the weld point and the inlet of the annealing furnace reaches a predetermined value (such as 12 meters) is usually taken as the start of the transition stage, and the moment when the distance between the weld point and the annealing furnace reaches the predetermined value is taken as the end of the transition stage.

[0051] In actual production, the process is mostly in a steady state, with only a few periods in a transitional state. The process is in a steady state before and after the transitional state.

[0052] When it is determined that the steady-state stage will be entered, the steel strip specifications of the steady-state stage can be used as the input of the first prediction model corresponding to the steady-state stage to obtain the parameter values ​​of the output steady-state stage, namely the furnace temperature and steel strip running speed of the open flame heating section in the predicted steady-state stage. Furthermore, assuming that the steady-state stage lasts for 200 seconds, the furnace temperature and steel strip running speed of the open flame heating section can always use the predicted parameter values ​​during these 200 seconds.

[0053] Steel strip specifications may include the width and thickness of the steel strip.

[0054] The above treatment ensures the accuracy of the furnace temperature and steel strip running speed in the open flame heating section during the steady-state stage, thereby improving the heating effect during the steady-state stage.

[0055] When it is determined that the transition phase is about to begin, the operating point variables of the steady-state phase before the transition phase and the operating point variables of the steady-state phase after the transition phase can be obtained. The operating point variables may include: steel strip specifications, steel strip running speed, and furnace temperature of the open flame heating section. The obtained operating point variables can be used as input to the second prediction model corresponding to the transition phase to obtain the output dynamic adjustment curve, that is, to obtain the dynamic adjustment curve of the two parameters of furnace temperature of open flame heating section and steel strip running speed. Furthermore, assuming that the transition phase lasts for 50 seconds, the furnace temperature of open flame heating section and steel strip running speed can be adjusted in real time according to the dynamic adjustment curve during these 50 seconds.

[0056] The steady-state stages before and after the transition stage are two adjacent steady-state stages. The steel strip specifications in the operating point variables of the steady-state stages before and after the transition stage are known. In addition, the steel strip running speed and the furnace temperature of the open flame heating section in the operating point variables of the steady-state stages before and after the transition stage can be obtained through the first prediction model.

[0057] Since the specifications of the steel strip will change during the transition phase, and different specifications of steel strip have different heating requirements for the annealing furnace, the furnace temperature of the open flame heating section and the running speed of the steel strip need to be dynamically adjusted in real time during the transition phase so that steel strips of different specifications can reach the target temperature at the outlet of the annealing furnace.

[0058] In other words, the above treatment can meet the requirements for real-time optimized control of the annealing furnace.

[0059] In practical applications, annealing furnaces typically include a preheating section and an open flame heating section, such as one preheating section and three open flame heating sections. Accordingly, the furnace temperature of the open flame heating section can include the furnace temperatures of the three open flame heating sections, namely the furnace temperature of the first open flame heating section, the furnace temperature of the second open flame heating section, and the furnace temperature of the third open flame heating section.

[0060] In actual production, the steel strip passes sequentially through a preheating section and three open-flame heating sections. The open-flame heating sections are equipped with burners, which directly heat the furnace and steel strip using natural gas combustion. The heat exchange between the steel strip and the flue gas and furnace includes radiation and convection. The preheating section does not contain burners. Flue gas from the open-flame heating sections flows into the preheating section, utilizing waste heat to heat the steel strip. After its temperature rises in the preheating section, the steel strip enters the open-flame heating sections. The heat exchange between the steel strip and the flue gas and furnace in the preheating section also includes radiation and convection. The furnace temperature in the preheating section is uncontrolled and can only be predicted; therefore, the scheme described in this disclosure only predicts the furnace temperature of the open-flame heating sections.

[0061] As can be seen from the above introduction, the implementation of the parameter adjustment method described in this disclosure depends on the pre-obtained prediction model. The following is an explanation of the method for obtaining the prediction model.

[0062] Figure 2 This is a flowchart illustrating an embodiment of the prediction model acquisition method described in this disclosure. Figure 2 As shown, the specific implementation methods are as follows.

[0063] In step 201, for a predetermined production stage in the process of heating the object using a heating device, training data corresponding to the production stage is obtained. The training data is training data generated based on real historical production data.

[0064] In step 202, a prediction model corresponding to the production stage is trained using the training data. The prediction model is used to predict the parameter adjustment method of the production stage during the heating process so that the parameters of the production stage can be adjusted according to the parameter adjustment method. The parameters are predetermined heating-related parameters.

[0065] By employing the scheme described in the above-described method embodiments, a prediction model can be trained using real historical production data. This model can then be used to predict parameter adjustment methods, allowing for corresponding parameter adjustments. This improves the accuracy of the adjustment results, eliminates reliance on human experience, and enhances product qualification rates and product performance stability.

[0066] In one embodiment of this disclosure, the heating device may be an annealing furnace, the object to be heated may be a steel strip, and the parameters may be the furnace temperature of the open flame heating section of the annealing furnace and the running speed of the steel strip.

[0067] In one embodiment of this disclosure, the production stage may include a steady-state stage in which the steel strip specifications do not change, and correspondingly, a first prediction model corresponding to the steady-state stage can be trained using the training data.

[0068] In one embodiment of this disclosure, the production stage may further include a transition stage in which the specifications of the steel strip change, and correspondingly, a second prediction model corresponding to the transition stage can be trained using the training data.

[0069] In other words, corresponding prediction models can be obtained for different production stages, making the prediction results more targeted and further improving the accuracy of the prediction results.

[0070] The methods for obtaining the first and second prediction models are explained below.

[0071] 1) First prediction model

[0072] In one embodiment of this disclosure, for the current training data, the following first process can be performed: obtaining a first output result of the first prediction model corresponding to the training data; determining predetermined intermediate parameters based on the training data and the first output result; determining a loss based on the first output result and the intermediate parameters, and updating the first prediction model using the loss; in response to determining that the first prediction model has converged, using the latest obtained first prediction model as the first prediction model corresponding to the steady-state stage; otherwise, for the next training data, the first process is repeated.

[0073] In one embodiment of this disclosure, the steel strip specifications and spatial coordinates in the training data can be used as input to a first prediction model to obtain a first output result. The first output result may include: steel strip temperature, steel strip running speed, and furnace temperature of the open flame heating section. The spatial coordinates are the coordinates of any position from the inlet to the outlet of the annealing furnace.

[0074] For example, if the length of the annealing furnace from inlet to outlet is 52 meters, then a spatial coordinate can be taken every 0.5 meters, and a corresponding training data point can be generated. The origin of the spatial coordinates can be the annealing furnace inlet, i.e., the inlet of the preheating section, and the endpoint can be the annealing furnace outlet, i.e., the outlet of the last open flame heating section. The training data can include spatial coordinates as well as the actual width and thickness of the steel strip, etc.

[0075] Based on the input, the first output result of the first prediction model can be obtained, including the steel strip temperature, the steel strip running speed, and the furnace temperature of the open flame heating section, such as the furnace temperature of the three open flame heating sections.

[0076] In another embodiment of this disclosure, the steel strip specifications in the training data, the steel strip running speed in the first output result, and the furnace temperature of the open flame heating section in the first output result can be used as inputs to a pre-trained preheating section furnace temperature prediction model to obtain the output preheating section furnace temperature. The annealing furnace includes a preheating section and an open flame heating section. The steel strip specifications in the training data, the steel strip running speed in the first output result, and the furnace temperature of the open flame heating section in the first output result can be used as inputs to a pre-trained annealing furnace thermal parameter prediction model to obtain the output preheating section composite convective heat transfer coefficient and the open flame heating section composite convective heat transfer coefficient.

[0077] The preheating section furnace temperature prediction model and the annealing furnace thermal parameter prediction model can be pre-trained. For example, training data can be generated using real historical production data, and then the preheating section furnace temperature prediction model and the annealing furnace thermal parameter prediction model can be trained.

[0078] Among them, the preheating section furnace temperature prediction model can be a pure data-driven deep neural network model, such as a regression model. Taking a model with three open flame heating sections as an example, the input of the model can be six features: the width and thickness of the steel strip, the running speed of the steel strip, and the furnace temperature of the three open flame heating sections. These six features can be collectively referred to as operating point variables, and the output of the model is the furnace temperature of the preheating section.

[0079] The thermal parameter prediction model for the annealing furnace can be a physical information-informed neural network (PINN) model, such as a regression model. The inputs to the model are the width, thickness, running speed of the steel strip, and the furnace temperature of the three open flame heating sections. The output of the model is the combined convective heat transfer coefficient of the preheating section and the heating section.

[0080] Based on the first output result and intermediate parameters, the loss can be determined, and the first prediction model can be updated using the loss.

[0081] In one embodiment of this disclosure, the first prediction model may be a neural network model based on physical information. In addition, the loss that satisfies the heat transfer mechanism constraint may be determined based on the first output result and intermediate parameters.

[0082] In the actual production process, the steel strip is heated sequentially through a preheating section and an open flame heating section. Taking the steel strip temperature Ts as the research object, the energy differential equation of the steel strip can be obtained as follows:

[0083]

[0084] It can be seen that the strip temperature Ts is a function of the spatial coordinate x and the time coordinate τ. The origin of the spatial coordinate can be the inlet of the annealing furnace, and the endpoint can be the outlet of the annealing furnace. The origin of the time coordinate can be arbitrarily chosen. ρ represents the strip density, and cp δ represents the specific heat capacity of the steel strip, v represents the running speed of the steel strip, w represents the width of the steel strip, δ represents the thickness of the steel strip, and Ф represents the heat exchange capacity of the steel strip.

[0085] When the annealing furnace is in a steady state, the temperature of the steel strip does not change with time, therefore equation (1) can be simplified to:

[0086]

[0087] Based on the above formula Figure 3 This is a schematic diagram of the physical information neural network architecture of the first prediction model of this disclosure.

[0088] like Figure 3 As shown, the inputs of the first prediction model are the width, thickness and spatial coordinates of the steel strip, and the outputs are the steel strip temperature, the steel strip running speed and the furnace temperature of the three open flame heating sections.

[0089] like Figure 3 As shown, the width, thickness, running speed of the steel strip, and furnace temperature of the three open flame heating sections can be used as inputs to the preheating section furnace temperature prediction model and the annealing furnace thermal parameter prediction model, and the furnace temperature of the preheating section, the composite convective heat transfer coefficient of the preheating section, and the composite convective heat transfer coefficient of the heating section can be obtained from the outputs of the two models respectively.

[0090] like Figure 3 As shown, further, the model loss can be calculated, such as differential equation loss, target temperature loss at the annealing furnace outlet, and boundary condition loss at the annealing furnace inlet.

[0091] Generally speaking, differential equation loss is present for all different training data, but target temperature loss at the annealing furnace outlet and boundary condition loss at the annealing furnace inlet are only present for some training data, such as training data with spatial coordinates of 0 and 52 (assuming the length of the annealing furnace is 52 meters).

[0092] Among them, the differential equation loss can be obtained based on formula (2). For example, the value of the left side of formula (2) “=" can be calculated first, where Ts is the steel strip temperature in the first output result, x is the spatial coordinate in the training data, v is the steel strip running speed in the first output result, and the value of the right side of formula (2) “=" can be calculated, where w and δ are the width and thickness of the steel strip in the training data, respectively, and Ф can be calculated by the furnace temperature of the preheating section, the composite convection heat transfer coefficient of the preheating section, the composite convection heat transfer coefficient of the heating section, and the furnace temperature of the three open flame heating sections in the first output result.

[0093] The target strip temperature loss at the annealing furnace outlet can be obtained by subtracting the strip temperature from the first output result and the actual strip outlet temperature (i.e., the target temperature) corresponding to the training data. The boundary condition loss at the annealing furnace inlet can be obtained by subtracting the strip temperature from the first output result and the actual strip inlet temperature corresponding to the training data. The inlet temperature can be the ambient temperature, i.e., the workshop temperature.

[0094] Once the first prediction model has converged, training can be stopped, and it can be deployed to the actual production line.

[0095] It can be seen that the neural network features of the first prediction model and other models in the scheme described in this disclosure are selected based on physical mechanisms. The input and output variables of the neural network have a clear qualitative relationship based on physical mechanisms, thereby enhancing the interpretability of the model.

[0096] Furthermore, in the scheme described in this disclosure, all losses used are losses that satisfy the heat transfer mechanism constraints, thereby ensuring that the regression results of the neural network satisfy the physical mechanism constraints and that the prediction results of the neural network conform to physical reality.

[0097] Furthermore, the modeling and training process in the scheme described in this disclosure couples physical mechanisms with historical big data, thereby further improving the model's accuracy, interpretability, and generalization.

[0098] 2) Second prediction model

[0099] In one embodiment of this disclosure, for the current training data, the following second process may be performed: obtaining a second output result of the second prediction model corresponding to the training data; determining predetermined intermediate parameters based on the training data and the second output result; determining a loss based on the second output result and the intermediate parameters, and updating the second prediction model using the loss; in response to determining that the second prediction model has converged, using the latest obtained second prediction model as the second prediction model corresponding to the transition phase; otherwise, for the next training data, repeating the second process.

[0100] In one embodiment of this disclosure, spatial coordinates, temporal coordinates, operating point variables of the steady-state stage before the transition stage, and operating point variables of the steady-state stage after the transition stage in the training data can be used as inputs to a second prediction model to obtain a second output result. The operating point variables may include: the width and thickness of the steel strip, the running speed of the steel strip, and the furnace temperature of the open flame heating section. The second output result may include: the steel strip temperature, the moving distance of the weld point, the running speed of the steel strip, and the furnace temperature of the open flame heating section. The weld point is used to connect steel strips of different specifications. The spatial coordinates are the coordinates of any position from the inlet to the outlet of the annealing furnace, and the temporal coordinates are the coordinates of any time point in the transition stage, with the temporal coordinates being relative to a selected time origin.

[0101] For example, the moment when the weld point has not entered the annealing furnace and the distance between it and the furnace inlet reaches a predetermined value (such as 12 meters) can be used as the time origin.

[0102] Based on the input, a second output result of the second prediction model can be obtained, which may include the strip temperature, the moving distance of the weld point, the strip running speed, and the furnace temperature of the open flame heating section, such as the furnace temperature of three open flame heating sections. The moving distance may refer to the distance the weld point moves relative to its position at the aforementioned time origin.

[0103] In another embodiment of this disclosure, the comprehensive steel strip specification can be determined based on the steel strip specification in the operating point variables of the steady-state stage before the transition stage and the steel strip specification in the operating point variables of the steady-state stage after the transition stage. The comprehensive steel strip specification, the steel strip running speed in the second output result, and the furnace temperature of the open flame heating section in the second output result can be used as inputs to the preheating section furnace temperature prediction model obtained through pre-training to obtain the output furnace temperature of the preheating section. The annealing furnace includes a preheating section and an open flame heating section. In addition, the comprehensive steel strip specification, the steel strip running speed in the second output result, and the furnace temperature of the open flame heating section in the second output result can be used as inputs to the preheating furnace thermal parameter prediction model obtained through pre-training to obtain the output composite convective heat transfer coefficient of the preheating section and the composite convective heat transfer coefficient of the open flame heating section.

[0104] There are no restrictions on how to obtain the comprehensive steel strip specifications. For example, for the width of the steel strip in the operating point variables of the steady-state stage before the transition stage and the width of the steel strip in the operating point variables of the steady-state stage after the transition stage, the comprehensive width can be obtained by weighted averaging the lengths of steel strips of different specifications in the annealing furnace. For the thickness of the steel strip in the operating point variables of the steady-state stage before the transition stage and the thickness of the steel strip in the operating point variables of the steady-state stage after the transition stage, the weighted cross-sectional area of ​​the steel strip can be obtained (e.g., by weighted averaging the lengths of steel strips of different specifications in the annealing furnace). This cross-sectional area can then be divided by the comprehensive width to obtain the comprehensive thickness. The obtained comprehensive width and comprehensive thickness are used as the comprehensive steel strip specifications.

[0105] In addition, preheating furnace temperature prediction models and annealing furnace thermal parameter prediction models can be pre-trained.

[0106] Among them, the preheating section furnace temperature prediction model can be a pure data-driven deep neural network model, such as a regression model. Taking a model with three open flame heating sections as an example, the input of the model can be the width and thickness of the steel strip, the running speed of the steel strip, and the furnace temperature of the three open flame heating sections. The output of the model is the furnace temperature of the preheating section.

[0107] The thermal parameter prediction model for the annealing furnace can be a neural network model based on physical information, such as a regression model. The inputs to the model are the width, thickness, running speed of the steel strip, and the furnace temperature of the three open flame heating sections. The output of the model is the combined convective heat transfer coefficient of the preheating section and the heating section.

[0108] Based on the second output and intermediate parameters, the loss can be determined, and the loss can be used to update the second prediction model.

[0109] In one embodiment of this disclosure, the second prediction model may be a neural network model based on physical information. In addition, the loss that satisfies the heat transfer mechanism constraint may be determined based on the second output result and intermediate parameters.

[0110] Based on the above introduction, Figure 4 This is a schematic diagram of the physical information neural network architecture of the second prediction model of this disclosure.

[0111] like Figure 4 As shown, the inputs of the second prediction model are spatial coordinates, time coordinates, and the operating point variables of the steady-state stage before the transition stage (i.e., the first operating point variables) and the operating point variables of the steady-state stage after the transition stage (i.e., the second operating point variables). The outputs are the steel strip temperature, the moving distance of the weld point, the running speed of the steel strip, and the furnace temperature of the three open flame heating sections.

[0112] like Figure 4 As shown, the comprehensive steel strip specifications, steel strip running speed, and furnace temperatures of the three open flame heating sections can be used as inputs to the preheating section furnace temperature prediction model and the annealing furnace thermal parameter prediction model, and the furnace temperature of the preheating section, the composite convective heat transfer coefficient of the preheating section, and the composite convective heat transfer coefficient of the heating section can be obtained from the outputs of the two models respectively.

[0113] like Figure 4 As shown, further, the model loss can be calculated, which may include: differential equation loss, target temperature loss at the annealing furnace outlet, boundary condition loss at the annealing furnace inlet, initial condition loss at the annealing furnace, displacement derivative velocity loss, initial displacement condition loss, steady-state velocity loss, steady-state furnace temperature loss, etc.

[0114] Generally speaking, differential equation loss is available for all different training data, but some losses are only available for a portion of the training data, depending on the specific circumstances.

[0115] The differential equation loss, the target temperature loss at the annealing furnace outlet, and the boundary condition loss at the annealing furnace inlet are the same as those mentioned above.

[0116] The initial condition loss of the annealing furnace refers to the temperature loss at the inlet and outlet of the annealing furnace when the time coordinate is the origin, i.e., when τ = 0. The displacement derivative velocity loss can be expressed as dl / dτ - v, where l represents the distance the weld point moves, τ represents the time coordinate, and v represents the predicted strip speed. The initial condition loss is used to assess whether the distance the weld point moves when τ = 0 is correct. The steady-state velocity loss is used to assess whether the predicted strip speed at τ = 0 is equal to the strip speed in the steady-state stage before the transition phase, and whether the predicted strip speed at the end of the transition phase (e.g., τ = 50, assuming the transition phase lasts 50 seconds) is equal to the strip speed in the steady-state stage after the transition phase. The steady-state furnace temperature loss is similar to the steady-state velocity loss, except that the strip speed is replaced by the furnace temperature.

[0117] Once the second prediction model has converged, training can be stopped, and it can be deployed to the actual production line.

[0118] It can be seen that the neural network features of the second prediction model and other models in the scheme described in this disclosure are selected based on physical mechanisms. The input and output variables of the neural network have a clear qualitative relationship based on physical mechanisms, thereby enhancing the interpretability of the model.

[0119] Furthermore, in the scheme described in this disclosure, all losses used are losses that satisfy the heat transfer mechanism constraints, thereby ensuring that the regression results of the neural network satisfy the physical mechanism constraints and that the prediction results of the neural network conform to physical reality.

[0120] Furthermore, the modeling and training process in the scheme described in this disclosure couples physical mechanisms with historical big data, thereby further improving the model's accuracy, interpretability, and generalization.

[0121] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this disclosure is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this disclosure. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this disclosure. Furthermore, for parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0122] In summary, the scheme described in the embodiments of this disclosure can achieve real-time optimization and recommendation of annealing furnace temperature and steel strip running speed during steady-state and transitional stages, thereby realizing real-time optimization and control of annealing furnace temperature and steel strip running speed. It also eliminates the reliance on manual experience, improves product qualification rate and product performance stability. In addition, the model modeling and training process couples physical mechanisms with historical production big data, improving the accuracy, interpretability and generalization of the model.

[0123] The above is an introduction to the method embodiments. The following describes the solution described in this disclosure further through device embodiments.

[0124] Figure 5 This is a schematic diagram of the structural composition of Embodiment 500 of the parameter adjustment device described in this disclosure. Figure 5 As shown, it includes: a first acquisition module 501, a prediction module 502, and an adjustment module 503.

[0125] The first acquisition module 501 is used to determine the production stage to be entered during the process of heating the object using a heating device, and to acquire the prediction model corresponding to the production stage.

[0126] The prediction module 502 is used to predict the parameter adjustment method of the production stage using the prediction model, wherein the parameters are predetermined heating-related parameters.

[0127] The adjustment module 503 is used to adjust the parameters of the production stage according to the parameter adjustment method.

[0128] In one embodiment of this disclosure, the heating device may be an annealing furnace, the object to be heated may be a steel strip, and the parameters may be the furnace temperature of the open flame heating section of the annealing furnace and the running speed of the steel strip.

[0129] By adopting the scheme described in the above-mentioned device embodiment, the parameter adjustment method can be predicted using the prediction model, such as the adjustment method of the furnace temperature of the open flame heating section and the running speed of the steel strip. Then, the furnace temperature of the open flame heating section and the running speed of the steel strip can be adjusted accordingly, thereby improving the accuracy of the adjustment results, eliminating the reliance on manual experience, and improving the product qualification rate and product performance stability.

[0130] In one embodiment of this disclosure, the production stage may include a steady-state stage in which the steel strip specifications have not changed. Accordingly, the prediction module 502 may use the first prediction model corresponding to the steady-state stage to predict the parameter values ​​of the parameters in the steady-state stage, and the adjustment module 503 may adjust the parameters to the parameter values.

[0131] In one embodiment of this disclosure, the prediction module 502 can use the steel strip specifications in the steady-state stage as input to the first prediction model to obtain the output parameter values.

[0132] In one embodiment of this disclosure, the production stage may further include a transition stage in which the specifications of the steel strip change. Accordingly, the prediction module 502 may use the second prediction model corresponding to the transition stage to predict the dynamic adjustment curve of the parameters in the transition stage, and the adjustment module 503 may adjust the parameters in the transition stage in real time according to the dynamic adjustment curve.

[0133] In one embodiment of this disclosure, the prediction module 502 can obtain the operating point variables of the steady-state stage before the transition stage and the operating point variables of the steady-state stage after the transition stage. The operating point variables include: steel strip specifications, steel strip running speed and furnace temperature of the open flame heating section. The obtained operating point variables are used as inputs to the second prediction model to obtain the output dynamic adjustment curve.

[0134] Figure 6 This is a schematic diagram of the composition of Embodiment 600 of the prediction model acquisition device described in this disclosure. Figure 6 As shown, it includes: a second acquisition module 601 and a training module 602.

[0135] The second acquisition module 601 is used to acquire training data corresponding to a predetermined production stage in the process of heating an object using a heating device. The training data is training data generated based on real historical production data.

[0136] The training module 602 is used to train a prediction model corresponding to the production stage using the training data. The prediction model is used to predict the parameter adjustment method of the production stage during the heating process so that the parameters of the production stage can be adjusted according to the parameter adjustment method. The parameters are predetermined heating-related parameters.

[0137] By adopting the scheme described in the above-mentioned device embodiment, a prediction model can be trained using real historical production data. The prediction model can then be used to predict the parameter adjustment method, and the corresponding parameter adjustments can be made. This improves the accuracy of the adjustment results, eliminates the reliance on human experience, and enhances the product qualification rate and product performance stability.

[0138] In one embodiment of this disclosure, the heating device may be an annealing furnace, the object to be heated may be a steel strip, and the parameters may be the furnace temperature of the open flame heating section of the annealing furnace and the running speed of the steel strip.

[0139] In one embodiment of this disclosure, the production stage may include a steady-state stage in which the steel strip specifications do not change. Accordingly, the training module 602 may use the training data to train a first prediction model corresponding to the steady-state stage.

[0140] In one embodiment of this disclosure, the training module 602 may perform the following first processing for the current training data: obtaining a first output result of the first prediction model corresponding to the training data; determining predetermined intermediate parameters based on the training data and the first output result; determining a loss based on the first output result and the intermediate parameters, and updating the first prediction model using the loss; in response to determining that the first prediction model has converged, using the latest obtained first prediction model as the first prediction model corresponding to the steady-state stage; otherwise, repeating the first processing for the next training data.

[0141] Specifically, in one embodiment of this disclosure, the training module 602 can use the steel strip specifications and spatial coordinates in the training data as input to the first prediction model to obtain a first output result. The first output result includes: steel strip temperature, steel strip running speed and furnace temperature of the open flame heating section. The spatial coordinates are the coordinates of any position from the inlet to the outlet of the annealing furnace.

[0142] In another embodiment of this disclosure, the training module 602 can use the steel strip specifications in the training data, the steel strip running speed in the first output result, and the furnace temperature of the open flame heating section in the first output result as inputs to the preheating section furnace temperature prediction model obtained through pre-training, and obtain the output furnace temperature of the preheating section. The annealing furnace includes a preheating section and an open flame heating section, and can use the steel strip specifications in the training data, the steel strip running speed in the first output result, and the furnace temperature of the open flame heating section in the first output result as inputs to the preheating furnace thermal parameter prediction model obtained through pre-training, and obtain the output composite convective heat transfer coefficient of the preheating section and the composite convective heat transfer coefficient of the open flame heating section.

[0143] In one embodiment of this disclosure, both the first prediction model and the annealing furnace thermal parameter prediction model are neural network models based on physical information. The training module 602 can determine the loss that satisfies the heat transfer mechanism constraint based on the first output result and intermediate parameters.

[0144] The production stage may also include a transition stage where the steel strip specifications change. Accordingly, the training module 602 can use the training data to train a second prediction model corresponding to the transition stage.

[0145] In one embodiment of this disclosure, the training module 602 may perform the following second processing for the current training data: obtaining a second output result of the second prediction model corresponding to the training data; determining predetermined intermediate parameters based on the training data and the second output result; determining a loss based on the second output result and the intermediate parameters, and updating the second prediction model using the loss; in response to determining that the second prediction model has converged, using the latest obtained second prediction model as the second prediction model corresponding to the transition phase; otherwise, repeating the second processing for the next training data.

[0146] Specifically, the training module 602 can use the spatial coordinates, temporal coordinates, and steady-state operation point variables before and after the transition phase from the training data as inputs to the second prediction model to obtain the second output results. The operation point variables include: steel strip specifications, steel strip running speed, and furnace temperature of the open flame heating section. The second output results include: steel strip temperature, welding point moving distance, steel strip running speed, and furnace temperature of the open flame heating section. The welding point is used to connect steel strips of different specifications. The spatial coordinates are the coordinates of any position from the inlet to the outlet of the annealing furnace, and the temporal coordinates are the coordinates of any time point in the transition phase.

[0147] In another embodiment of this disclosure, the training module 602 can determine the comprehensive steel strip specification based on the steel strip specification in the operating point variables of the steady-state stage before the transition stage and the steel strip specification in the operating point variables of the steady-state stage after the transition stage. The comprehensive steel strip specification, the steel strip running speed in the second output result, and the furnace temperature of the open flame heating section in the second output result can be used as inputs to the pre-trained preheating section furnace temperature prediction model to obtain the output preheating section furnace temperature. The annealing furnace includes a preheating section and an open flame heating section. In addition, the comprehensive steel strip specification, the steel strip running speed in the second output result, and the furnace temperature of the open flame heating section in the second output result can also be used as inputs to the pre-trained annealing furnace thermal parameter prediction model to obtain the output preheating section composite convective heat transfer coefficient and the open flame heating section composite convective heat transfer coefficient.

[0148] In one embodiment of this disclosure, both the second prediction model and the annealing furnace thermal parameter prediction model are neural network models based on physical information. The training module 602 can determine the loss that satisfies the heat transfer mechanism constraint based on the second output result and intermediate parameters.

[0149] Figure 5 and Figure 6 The specific workflow of the device embodiment shown can be found in the relevant descriptions in the foregoing method embodiments, and will not be repeated here.

[0150] In summary, by adopting the scheme described in the embodiments of this disclosure, real-time optimization and recommendation of annealing furnace temperature and steel strip running speed can be achieved during the steady-state and transition stages, thereby realizing real-time optimization and control of annealing furnace temperature and steel strip running speed. This eliminates the reliance on human experience, improves product qualification rate and product performance stability. In addition, the model modeling and training process couples physical mechanisms with historical production big data, improving the model's accuracy, interpretability and generalization.

[0151] The solutions described in this disclosure can be applied to the field of artificial intelligence, particularly deep learning and big data processing. Artificial intelligence is the study of enabling computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, and planning). It involves both hardware and software technologies. Artificial intelligence hardware technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, and big data processing. Artificial intelligence software technologies mainly include computer vision, speech recognition, natural language processing, machine learning / deep learning, big data processing, and knowledge graph technologies.

[0152] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0153] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0154] Figure 7 A schematic block diagram of an electronic device 700 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0155] like Figure 7As shown, device 700 includes a computing unit 701, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 702 or a computer program loaded from storage unit 708 into random access memory (RAM) 703. RAM 703 may also store various programs and data required for the operation of device 700. The computing unit 701, ROM 702, and RAM 703 are interconnected via bus 704. Input / output (I / O) interface 705 is also connected to bus 704.

[0156] Multiple components in device 700 are connected to I / O interface 705, including: input unit 706, such as keyboard, mouse, etc.; output unit 707, such as various types of monitors, speakers, etc.; storage unit 708, such as disk, optical disk, etc.; and communication unit 709, such as network card, modem, wireless transceiver, etc. Communication unit 709 allows device 700 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0157] The computing unit 701 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 701 performs the various methods and processes described above, such as those described in this disclosure. For example, in some embodiments, the methods described in this disclosure can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 708. In some embodiments, part or all of the computer program can be loaded and / or installed on device 700 via ROM 702 and / or communication unit 709. When the computer program is loaded into RAM 703 and executed by the computing unit 701, one or more steps of the methods described in this disclosure can be performed. Alternatively, in other embodiments, the computing unit 701 can be configured to perform the methods described in this disclosure by any other suitable means (e.g., by means of firmware).

[0158] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0159] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0160] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0161] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0162] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0163] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0164] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0165] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for obtaining a prediction model, comprising: For a predetermined production stage in the process of heating an object using a heating device, training data corresponding to the production stage is obtained, and the training data is training data generated based on real historical production data. The heating equipment includes an annealing furnace, and the object being heated includes a steel strip. The prediction model corresponding to the production stage is trained using the training data, including: in response to the production stage being a steady-state stage where the steel strip specifications have not changed, the following first processing is performed on the current training data: the steel strip specifications and spatial coordinates in the training data are used as inputs to the first prediction model corresponding to the steady-state stage, resulting in a first output result including the following information: the steel strip running speed and the furnace temperature of the open flame heating section, where the spatial coordinates are the coordinates of any position from the inlet to the outlet of the annealing furnace; the steel strip specifications in the training data, the steel strip running speed in the first output result, and the furnace temperature of the open flame heating section are inputs to the pre-trained prediction model, resulting in the following intermediate parameters: the furnace temperature of the preheating section, the composite convective heat transfer coefficient of the preheating section, and the composite convective heat transfer coefficient of the open flame heating section; the loss satisfying the heat transfer mechanism constraint is determined based on the first output result and the intermediate parameters, and the first prediction model is updated using the loss; the prediction model is used to predict the parameter adjustment method of the production stage during the heating process so that the parameters of the production stage can be adjusted according to the parameter adjustment method, where the parameters are predetermined heating-related parameters.

2. The method according to claim 1, wherein, The parameters include: the furnace temperature of the open flame heating section of the annealing furnace and the running speed of the steel strip.

3. The method according to claim 1, further comprising: In response to determining that the first prediction model has converged, the latest obtained first prediction model is used as the first prediction model corresponding to the steady-state stage; otherwise, the first process is repeated for the next training data.

4. The method according to claim 1, wherein, The first output result also includes: steel strip temperature.

5. The method according to claim 1, wherein, The following intermediate parameters are obtained by inputting the steel strip specifications from the training data, the steel strip running speed from the first output result, and the furnace temperature of the open flame heating section into the pre-trained prediction model: The steel strip specifications in the training data, the steel strip running speed in the first output result, and the furnace temperature of the open flame heating section in the first output result are used as inputs to the preheating section furnace temperature prediction model obtained through pre-training, and the output furnace temperature of the preheating section is obtained. The annealing furnace includes the preheating section and the open flame heating section. The steel strip specifications in the training data, the steel strip running speed in the first output result, and the furnace temperature of the open flame heating section in the first output result are used as inputs to the pre-trained annealing furnace thermal parameter prediction model to obtain the output composite convective heat transfer coefficient of the preheating section and the composite convective heat transfer coefficient of the open flame heating section.

6. The method according to claim 5, wherein, Both the first prediction model and the annealing furnace thermal parameter prediction model are neural network models based on physical information.

7. The method according to claim 1, wherein, The production stage also includes a transitional stage where the steel strip specifications change; The step of training a prediction model corresponding to the production stage using the training data includes: training a second prediction model corresponding to the transition stage using the training data.

8. The method according to claim 7, wherein, The step of training the second prediction model corresponding to the transition phase using the training data includes: For the current training data, perform the following second processing: Obtain the second output result of the second prediction model corresponding to the training data. The second output result includes: steel strip temperature, welding point moving distance, steel strip running speed and furnace temperature of open flame heating section. The welding point is used to connect steel strips of different specifications. Based on the training data and the second output result, predetermined intermediate parameters are determined, including: furnace temperature of the preheating section, composite convective heat transfer coefficient of the preheating section, and composite convective heat transfer coefficient of the open flame heating section. The loss is determined based on the second output result and the intermediate parameters, and the second prediction model is updated using the loss. In response to determining that the second prediction model has converged, the latest obtained second prediction model is used as the second prediction model corresponding to the transition phase; otherwise, the second processing is repeated for the next training data.

9. The method according to claim 8, wherein, The step of obtaining the second output result of the second prediction model corresponding to the training data includes: The spatial coordinates, temporal coordinates, the operation point variables of the steady-state phase before the transition phase, and the operation point variables of the steady-state phase after the transition phase in the training data are used as inputs to the second prediction model to obtain the second output result. The operating point variables include: steel strip specifications, steel strip running speed, and furnace temperature of the open flame heating section. The spatial coordinates are the coordinates of any position from the inlet to the outlet of the annealing furnace, and the time coordinates are the coordinates of any time point in the transition stage.

10. The method according to claim 9, wherein, The step of determining the predetermined intermediate parameters based on the training data and the second output result includes: The comprehensive steel strip specification is determined based on the steel strip specifications in the operating point variables of the steady-state stage before the transition stage and the steel strip specifications in the operating point variables of the steady-state stage after the transition stage. The comprehensive steel strip specifications, the steel strip running speed in the second output result, and the furnace temperature of the open flame heating section in the second output result are used as inputs to the preheating section furnace temperature prediction model obtained through pre-training, and the output furnace temperature of the preheating section is obtained. The annealing furnace includes the preheating section and the open flame heating section. The comprehensive steel strip specifications, the steel strip running speed in the second output result, and the furnace temperature of the open flame heating section in the second output result are used as inputs to the pre-trained annealing furnace thermal parameter prediction model to obtain the output composite convective heat transfer coefficient of the preheating section and the composite convective heat transfer coefficient of the open flame heating section.

11. The method according to claim 10, wherein, Both the second prediction model and the annealing furnace thermal parameter prediction model are neural network models based on physical information; The step of determining the loss based on the second output result and the intermediate parameters includes: determining the loss that satisfies the heat transfer mechanism constraint based on the second output result and the intermediate parameters.

12. A parameter adjustment method, comprising: During the process of heating an object using a heating device, the production stage to be entered is determined, and a prediction model corresponding to the production stage is obtained, wherein the prediction model is obtained according to the method of any one of claims 1-11; The prediction model is used to predict the parameter adjustment method for the production stage, where the parameters are predetermined heating-related parameters; The parameters for the production stage are adjusted according to the parameter adjustment method described above.

13. The method according to claim 12, wherein, The heating equipment includes: an annealing furnace; The object to be heated includes: a steel strip; The parameters include: the furnace temperature of the open flame heating section of the annealing furnace and the running speed of the steel strip.

14. The method according to claim 13, wherein, The production stage includes: a steady-state stage in which the steel strip specifications do not change; The method of predicting the parameter adjustment of the production stage using the prediction model includes: using the first prediction model corresponding to the steady-state stage to predict the parameter value of the steady-state stage; The adjustment of the parameters in the production stage includes: adjusting the parameters to the parameter values.

15. The method according to claim 14, wherein, The step of using the first prediction model corresponding to the steady-state stage to predict the parameter values ​​of the steady-state stage includes: The steel strip specifications during the steady-state phase are used as input to the first prediction model to obtain the output parameter values.

16. The method according to claim 13, 14 or 15, wherein, The production stage includes: a transitional stage in which the specifications of the steel strip change; The method of predicting the parameter adjustment of the production stage using the prediction model includes: using the second prediction model corresponding to the transition stage to predict the dynamic adjustment curve of the parameter of the transition stage; The adjustment of the parameters in the production stage includes: adjusting the parameters in the transition stage in real time according to the dynamic adjustment curve.

17. The method according to claim 16, wherein, The step of using the second prediction model corresponding to the transition phase to predict the dynamic adjustment curve of the parameters in the transition phase includes: Obtain the operating point variables of the steady-state stage before the transition stage and the operating point variables of the steady-state stage after the transition stage. The operating point variables include: steel strip specifications, steel strip running speed, and furnace temperature of the open flame heating section. The obtained operation point variables are used as input to the second prediction model to obtain the output dynamic adjustment curve.

18. A parameter adjustment device, comprising: The module consists of an acquisition module, a prediction module, and an adjustment module. The first acquisition module is used to determine the production stage to be entered during the process of heating the object using a heating device, and to acquire a prediction model corresponding to the production stage, wherein the prediction model is obtained by the method according to any one of claims 1-11; The prediction module is used to predict the parameter adjustment method of the production stage using the prediction model, wherein the parameters are predetermined heating-related parameters; The adjustment module is used to adjust the parameters in the production stage according to the parameter adjustment method.

19. The apparatus according to claim 18, wherein, The heating equipment includes: an annealing furnace; The object to be heated includes: a steel strip; The parameters include: the furnace temperature of the open flame heating section of the annealing furnace and the running speed of the steel strip.

20. The apparatus according to claim 19, wherein, The production stage includes: a steady-state stage in which the steel strip specifications do not change; The prediction module uses the first prediction model corresponding to the steady-state stage to predict the parameter values ​​of the parameters in the steady-state stage. The adjustment module adjusts the parameter to the parameter value.

21. The apparatus according to claim 20, wherein, The prediction module takes the steel strip specifications in the steady-state stage as input to the first prediction model and obtains the output parameter values.

22. The apparatus according to claim 19, 20 or 21, wherein, The production stage includes: a transitional stage in which the specifications of the steel strip change; The prediction module uses the second prediction model corresponding to the transition phase to predict the dynamic adjustment curve of the parameters in the transition phase. The adjustment module adjusts the parameters of the transition phase in real time according to the dynamic adjustment curve.

23. The apparatus according to claim 22, wherein, The prediction module acquires the operating point variables of the steady-state stage before the transition stage and the operating point variables of the steady-state stage after the transition stage. The operating point variables include: steel strip specifications, steel strip running speed and furnace temperature of the open flame heating section. The acquired operating point variables are used as input to the second prediction model to obtain the output dynamic adjustment curve.

24. A predictive model acquisition device, comprising: The second acquisition module and the training module; The second acquisition module is used to acquire training data corresponding to a predetermined production stage in the process of heating an object using a heating device. The training data is training data generated based on real historical production data. The heating equipment includes an annealing furnace, and the object being heated includes a steel strip. The training module is used to train a prediction model corresponding to the production stage using the training data, including: in response to the production stage being a steady-state stage where the steel strip specifications have not changed, performing the following first processing on the current training data: using the steel strip specifications and spatial coordinates in the training data as input to the first prediction model corresponding to the steady-state stage, obtaining a first output result including the following information: steel strip running speed and furnace temperature of the open flame heating section, wherein the spatial coordinates are the coordinates of any position from the inlet to the outlet of the annealing furnace; and using the steel strip specifications and the first output result in the training data... The prediction model, pre-trained based on the steel strip running speed and furnace temperature of the open flame heating section, outputs the following intermediate parameters: furnace temperature of the preheating section, composite convective heat transfer coefficient of the preheating section, and composite convective heat transfer coefficient of the open flame heating section. Based on the first output and the intermediate parameters, the loss satisfying the heat transfer mechanism constraint is determined, and the first prediction model is updated using the loss. The prediction model is used to predict the parameter adjustment method of the production stage during the heating process so that the parameters of the production stage can be adjusted according to the parameter adjustment method. The parameters are predetermined heating-related parameters.

25. The apparatus according to claim 24, wherein, The parameters include: the furnace temperature of the open flame heating section of the annealing furnace and the running speed of the steel strip.

26. The apparatus according to claim 24, wherein, In response to determining that the first prediction model has converged, the training module uses the latest obtained first prediction model as the first prediction model corresponding to the steady-state stage; otherwise, it repeats the first processing for the next training data.

27. The apparatus according to claim 24, wherein, The first output result also includes: steel strip temperature.

28. The apparatus according to claim 24, wherein, The training module takes the steel strip specifications, steel strip running speed, and open flame heating section furnace temperature from the training data as inputs to the preheating section furnace temperature prediction model obtained through pre-training, and obtains the output furnace temperature of the preheating section. The annealing furnace includes the preheating section and the open flame heating section. The training module takes the steel strip specifications, steel strip running speed, and open flame heating section furnace temperature from the training data as inputs to the preheating furnace thermal parameter prediction model obtained through pre-training, and obtains the output composite convective heat transfer coefficient of the preheating section and the composite convective heat transfer coefficient of the open flame heating section.

29. The apparatus according to claim 28, wherein, Both the first prediction model and the annealing furnace thermal parameter prediction model are neural network models based on physical information.

30. The apparatus according to claim 24, wherein, The production stage also includes a transitional stage where the steel strip specifications change; The training module uses the training data to train a second prediction model corresponding to the transition phase.

31. The apparatus according to claim 30, wherein, The training module performs the following second processing on the current training data: It obtains a second output result of the second prediction model corresponding to the training data, the second output result including: steel strip temperature, weld point movement distance, steel strip running speed, and furnace temperature of the open flame heating section, the weld point being used to connect steel strips of different specifications; it determines predetermined intermediate parameters based on the training data and the second output result, the intermediate parameters including: furnace temperature of the preheating section, composite convective heat transfer coefficient of the preheating section, and composite convective heat transfer coefficient of the open flame heating section; it determines a loss based on the second output result and the intermediate parameters, and updates the second prediction model using the loss; in response to determining that the second prediction model has converged, it uses the latest obtained second prediction model as the second prediction model corresponding to the transition stage; otherwise, it repeats the second processing for the next training data.

32. The apparatus according to claim 31, wherein, The training module takes the spatial coordinates, time coordinates, the operation point variables of the steady-state stage before the transition stage and the operation point variables of the steady-state stage after the transition stage from the training data as inputs to the second prediction model to obtain the second output result. The operating point variables include: steel strip specifications, steel strip running speed, and furnace temperature of the open flame heating section. The spatial coordinates are the coordinates of any position from the inlet to the outlet of the annealing furnace, and the time coordinates are the coordinates of any time point in the transition stage.

33. The apparatus according to claim 32, wherein, The training module determines the comprehensive steel strip specification based on the steel strip specifications in the operating point variables of the steady-state stage before the transition stage and the steel strip specifications in the operating point variables of the steady-state stage after the transition stage. It then uses the comprehensive steel strip specification, the steel strip running speed in the second output result, and the furnace temperature of the open flame heating section in the second output result as inputs to a pre-trained preheating section furnace temperature prediction model to obtain the output furnace temperature of the preheating section. The annealing furnace includes the preheating section and the open flame heating section. The module also uses the comprehensive steel strip specification, the steel strip running speed in the second output result, and the furnace temperature of the open flame heating section in the second output result as inputs to a pre-trained annealing furnace thermal parameter prediction model to obtain the output composite convective heat transfer coefficient of the preheating section and the composite convective heat transfer coefficient of the open flame heating section.

34. The apparatus according to claim 33, wherein, Both the second prediction model and the annealing furnace thermal parameter prediction model are neural network models based on physical information; The training module determines the loss that satisfies the heat transfer mechanism constraint based on the second output result and the intermediate parameters.

35. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-17.

36. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-17.

37. A computer program product comprising a computer program / instructions that, when executed by a processor, implement the method of any one of claims 1-17.

Images