Rod temperature field calculation method and device

Abstract

This invention provides a method and apparatus for calculating the temperature field of bar and wire rods, relating to the field of metallurgical engineering technology. The method includes: acquiring bar and wire rod production data and mechanistic model input parameters; calculating the mechanistic model input parameters based on a bar and wire rod heat exchange and conduction mechanism model to obtain calculated mechanistic model temperature values ​​corresponding to each discrete concentric ring of the bar and wire rod; predicting temperature residuals in the bar and wire rod production data according to a pre-trained preset residual prediction model to obtain calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar and wire rod; and correcting the calculated mechanistic model temperature values ​​at corresponding positions based on the calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar and wire rod to obtain the calculated temperature field of the bar and wire rod. The method and apparatus provided by this invention can improve the accuracy and applicability of temperature field calculation in the bar and wire rod production process.

Description

Technical Field

[0001] This invention relates to the field of metallurgical engineering technology, specifically to a method and apparatus for calculating the temperature field of bars and wires. Background Technology

[0002] Precise calculation of the temperature field during bar and wire rod production is crucial for determining rolling cooling regimes, microstructure, and properties, and is an important way to improve production efficiency, resource allocation efficiency, and reduce production costs. Bar and wire rod are the most important long products in the steel industry. Their rolling process involves a series of passes that periodically plastically deform and cool the high-temperature steel billet, gradually reducing its inner diameter to achieve the target size. During this process, the billet experiences deformation heat, radiative heat dissipation, conductive heat dissipation, and convective heat dissipation. The cooling process is the most critical stage determining the microstructure and mechanical properties, mainly controlling the phase transformation process of the rolled steel. When the temperature drops below 400-500℃, the phase transformation is complete, and the material enters the natural air-cooling stage. Establishing a temperature field model is essential for accurately simulating and predicting the temperature distribution of the rolled piece at any location and time during rolling and cooling, forming the basis for precise process control. Current technologies only study the temperature field of bar and wire rod within a specific rolling process, and the accuracy of comprehensive prediction and analysis of the cross-sectional temperature throughout the entire process is insufficient to meet actual production needs, resulting in inaccurate temperature field predictions and poor applicability. Summary of the Invention

[0003] To address the problems in the prior art, embodiments of the present invention provide a method and apparatus for calculating the temperature field of bars and wires, which can at least partially solve the problems existing in the prior art.

[0004] On the one hand, this invention proposes a method for calculating the temperature field of bars and wires, including: Obtain the production data and mechanism model input parameters of bar and wire rods, and calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of bar and wire rods to obtain the mechanism model temperature calculation values ​​corresponding to each discrete concentric ring of bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; Based on the pre-trained preset residual prediction model, the temperature residual is predicted for the production data of the bar and wire, and the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire is obtained. Based on the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire, the calculated temperature value of the mechanism model at the corresponding position is corrected and calculated to obtain the temperature field calculation result of the bar and wire. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0005] The work-heat factor of plastic deformation is calculated according to the following formula:

[0006] in, The work-heat factor for plastic deformation, For work-heat conversion efficiency, The flow stress of the material is the temperature. ,strain and strain rate The function.

[0007] The step of correcting the calculated temperature values ​​of the mechanism model at the corresponding positions based on the calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar / wire to obtain the calculated temperature field of the bar / wire includes: The sum of the calculated temperature residuals corresponding to each discrete bar and wire concentric ring and the calculated temperature of the mechanism model at the corresponding position is taken as the temperature field calculation result of the bar and wire.

[0008] The process of training a machine learning regression model based on preset residual prediction sample data to obtain the preset residual prediction model includes: The preset residual prediction model is obtained by using an objective function that includes a loss function and a regularization term, and training a machine learning regression model based on preset residual prediction sample data. The loss function is used to measure the difference between the actual value of the temperature residual and the predicted value of the temperature residual, and the regularization term is used to control the complexity of the preset residual prediction model.

[0009] Prior to the step of training a machine learning regression model based on preset residual prediction sample data to obtain the preset residual prediction model, the method for calculating the temperature field of bar and wire rods further includes: Preprocess the collected raw historical data on bar and wire production; The preprocessed data is used to construct and select features, and the constructed and selected feature variables are then standardized. The standardized feature variables and target variables are determined as the preset residual prediction sample data; the target variable is composed of the difference between the measured temperature corresponding to each discrete concentric ring of the bar and wire and the calculated temperature value of the mechanism model corresponding to the discrete point of the heat exchange and conduction mechanism model of the bar and wire.

[0010] The method for calculating the temperature field of bars and wires further includes: Monitor the accuracy index of the preset residual prediction model, and update the preset residual prediction model based on the monitoring results of the accuracy index.

[0011] On one hand, the present invention proposes a device for calculating the temperature field of bars and wires, comprising: The acquisition unit is used to acquire bar and wire production data and mechanism model input parameters, and calculate the mechanism model input parameters based on the pre-constructed bar and wire heat exchange and conduction mechanism model to obtain the mechanism model temperature calculation values ​​corresponding to each discrete bar and wire concentric ring. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; The calculation unit is used to predict the temperature residual of the bar and wire production data according to the preset residual prediction model obtained by pre-training, to obtain the temperature residual calculation value corresponding to each discrete bar and wire concentric ring, and to correct the temperature calculation value of the mechanism model at the corresponding position according to the temperature residual calculation value corresponding to each discrete bar and wire concentric ring, so as to obtain the bar and wire temperature field calculation result. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0012] In another aspect, embodiments of the present invention provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the following method: Obtain the production data and mechanism model input parameters of bar and wire rods, and calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of bar and wire rods to obtain the mechanism model temperature calculation values ​​corresponding to each discrete concentric ring of bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; Based on the pre-trained preset residual prediction model, the temperature residual is predicted for the production data of the bar and wire, and the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire is obtained. Based on the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire, the calculated temperature value of the mechanism model at the corresponding position is corrected and calculated to obtain the temperature field calculation result of the bar and wire. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0013] This invention provides a computer-readable storage medium, comprising: The computer-readable storage medium stores a computer program that, when executed by a processor, implements the following method: Obtain the production data and mechanism model input parameters of bar and wire rods, and calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of bar and wire rods to obtain the mechanism model temperature calculation values ​​corresponding to each discrete concentric ring of bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; Based on the pre-trained preset residual prediction model, the temperature residual is predicted for the production data of the bar and wire, and the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire is obtained. Based on the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire, the calculated temperature value of the mechanism model at the corresponding position is corrected and calculated to obtain the temperature field calculation result of the bar and wire. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0014] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the following method: Obtain the production data and mechanism model input parameters of bar and wire rods, and calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of bar and wire rods to obtain the mechanism model temperature calculation values ​​corresponding to each discrete concentric ring of bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; Based on the pre-trained preset residual prediction model, the temperature residual is predicted for the production data of the bar and wire, and the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire is obtained. Based on the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire, the calculated temperature value of the mechanism model at the corresponding position is corrected and calculated to obtain the temperature field calculation result of the bar and wire. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0015] The present invention provides a method and apparatus for calculating the temperature field of bar and wire rods. This method acquires bar and wire rod production data and mechanistic model input parameters. Based on a pre-constructed heat exchange and conduction mechanism model for bar and wire rods, it calculates the temperature values ​​corresponding to each discrete concentric ring of the bar and wire rod. The heat exchange and conduction mechanism model includes a plastic deformation work-heat factor in the rolling deformation zone. A pre-trained residual prediction model is used to predict the temperature residuals of the bar and wire rod production data, obtaining calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar and wire rod. The calculated temperature values ​​of the mechanistic model at corresponding positions are then corrected based on these calculated temperature residual values ​​to obtain the final temperature field calculation result for the bar and wire rod. The pre-trained residual prediction model, obtained by training a machine learning regression model using pre-trained residual prediction sample data, improves the accuracy and applicability of the temperature field calculation during the bar and wire rod production process. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a schematic flowchart of a method for calculating the temperature field of bars and wires according to an embodiment of the present invention.

[0017] Figure 2 This is a flowchart illustrating a method for calculating the temperature field of bars and wires according to another embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of the structure of a bar and wire temperature field calculation device provided in an embodiment of the present invention.

[0019] Figure 4 This is a schematic diagram of the physical structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments and descriptions of the present invention are used to explain the present invention, but are not intended to limit the present invention. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other.

[0021] Figure 1 This is a flowchart illustrating a method for calculating the temperature field of bars and wires according to an embodiment of the present invention, as shown below. Figure 1 As shown, the method for calculating the temperature field of bars and wires provided in this embodiment of the invention includes: Step S1: Obtain the production data and mechanism model input parameters of the bar and wire rods. Calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of the bar and wire rods to obtain the calculated temperature values ​​of the mechanism model corresponding to each discrete concentric ring of the bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone.

[0022] Step S2: Based on the pre-trained preset residual prediction model, perform temperature residual prediction on the bar and wire production data to obtain the calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar and wire, and correct the calculated temperature values ​​of the mechanism model at the corresponding positions based on the calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar and wire to obtain the calculated results of the bar and wire temperature field. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0023] In step S1 above, the device acquires the production data of bar and wire and the input parameters of the mechanism model, and calculates the input parameters of the mechanism model based on the pre-constructed heat exchange and conduction mechanism model of bar and wire to obtain the calculated temperature values ​​of the mechanism model corresponding to each discrete concentric ring of bar and wire. The heat exchange and conduction mechanism model for the bar and wire rod includes a heat factor related to the plastic deformation work in the rolling deformation zone. The device can be a computer, such as a server, that performs the method. The acquisition, storage, use, and processing of data in this application's technical solution all comply with relevant regulations. Figure 2 As shown, the heat exchange and conduction mechanism model of bar and wire rods is explained as follows: Based on the fundamental principles of heat transfer, a physical mechanism model is established to describe the entire heat transfer process of bar and wire rods from the heating furnace, through roughing, intermediate rolling, finishing, water cooling box, recovery section, and finally to the cooling bed. This model discretizes the rolled piece into multiple elements and considers the following heat conduction processes: Heat generated by deformation: The heat generated in each rolling pass is calculated based on the plastic deformation work.

[0024] Radiation heat loss: Calculate the radiative heat loss from the surface of the rolled piece to the environment based on the Stefan-Boltzmann law.

[0025] Conductive heat dissipation: Calculate the conductive heat loss when the workpiece comes into contact with the rolls and guide devices.

[0026] Convective heat dissipation: Calculates the heat loss of the rolled piece during natural cooling in air, descaling, and forced convection cooling. The input parameters for this mechanism model include: steel grade thermophysical properties (specific heat capacity, thermal conductivity, density, etc.), initial temperature, process parameters for each pass (rolling speed, reduction, interval time, etc.), and cooling parameters (water pressure, water volume, cooling length, etc.). Specifically, a one-dimensional unsteady-state heat transfer model is established, discretizing the circular cross-section bar into several concentric circular ring elements radially. The core governing equation is the cylindrical coordinate form of the Fourier thermal conductivity differential equation:

[0027] Where T is temperature (°C), t is time (s), and r is the radial coordinate (m). Density function with respect to temperature (kg / m³) 3 ), Let be the specific heat capacity function with respect to temperature (J / kg·℃), and k(T) be the thermal conductivity function with respect to temperature (W / m·℃). Internal heat element (W / m) 3 This mainly refers to the heat generated during deformation, specifically the heat factor of plastic deformation in the rolling deformation zone. The heat factor of plastic deformation is calculated using the following formula:

[0028] in, The work-heat factor for plastic deformation, For work-heat conversion efficiency, The flow stress of the material (MPa) is the temperature. ,strain and strain rate The function can be obtained using a thermal simulation testing machine. This machine rapidly heats the sample using resistance heating and can apply predetermined compressive or tensile deformations to the sample while precisely controlling the temperature. It can record load, displacement, temperature, and time in real time, thus obtaining the stress-strain curves of the material under different thermal conditions.

[0029] Establishing heat loss calculation models for the dephosphorization, water cooling, and controlled cooling stages is a standard technique in this field and will not be elaborated further. During the use of the preset residual prediction model, the input parameters of the mechanism model can be calculated based on the heat exchange and conduction mechanism model of the bar and wire rods. The output is the calculated temperature value of the mechanism model corresponding to each discrete concentric ring of the bar and wire rods. Each discrete concentric ring of the bar and wire rods can be understood as a concentric ring of the bar and wire rods at different discrete locations. Correspondingly, the calculated temperature value of the mechanism model corresponding to each discrete concentric ring of the bar and wire rods can be the calculated temperature value of the mechanism model at different time points corresponding to each discrete concentric ring of the bar and wire rods.

[0030] In step S2 above, the device predicts the temperature residual of the bar and wire production data according to the preset residual prediction model obtained by pre-training, obtains the temperature residual calculation value corresponding to each discrete concentric ring of the bar and wire, and corrects the temperature calculation value of the mechanism model at the corresponding position according to the temperature residual calculation value corresponding to each discrete concentric ring of the bar and wire, so as to obtain the bar and wire temperature field calculation result. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data. Since the bar and wire production data and the input parameters of the mechanism model are generated under the same bar and wire production conditions, the temperature residual calculation values ​​output by the preset residual prediction model from the bar and wire production data, corresponding to the discrete concentric rings of the bar and wire respectively, can correct the temperature calculation values ​​of the mechanism model output by the bar and wire heat exchange and conduction mechanism model corresponding to the discrete concentric rings of the bar and wire respectively.

[0031] The production data of bars and wires can be directly input into the preset residual prediction model, and the output of the preset residual prediction model can be used as the temperature residual calculation value corresponding to each discrete concentric ring of bars and wires.

[0032] The step of correcting the calculated temperature values ​​of the mechanism model at the corresponding positions based on the calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar / wire to obtain the calculated temperature field of the bar / wire includes: The sum of the calculated temperature residuals corresponding to each discrete concentric ring of the bar / wire and the calculated temperature from the corresponding mechanistic model is used as the temperature field calculation result for the bar / wire. This achieves fusion prediction of the temperature field during rolling and cooling processes. The results of the residual prediction model are combined with the results of the mechanistic calculation model to obtain the final high-precision temperature prediction value.

[0033] The rolling process is tracked, and the temperature T_mech for mechanistic calculation is calculated based on real-time process parameters. Then, the temperature is predicted using the XGBoost residual model based on all production data. T, ultimately T_final = T_mech + The T-formula is used to calculate the temperature field. While pure mechanistic models follow physical laws, they struggle to accurately describe complex and variable field boundary conditions. Pure data models, while fitting historical data, may produce physically unreliable predictions when process parameters exceed their training range. This invention, through fusion, retains the determinism of the physical model while incorporating the precise bias compensation of the data model. The data model portion, due to the fully connected layer architecture of the neural network, enables… The term T contains hidden information that cannot be observed in production data. It controls the influence of a specific environment in the hidden logic. Therefore, the mechanism and data can be integrated through a dynamic compensation chain.

[0034] The trained model can correct the deviations generated by the current mechanism model based on real-time process parameters.

[0035] The training process of the preset residual prediction model is explained as follows: Before the step of training a machine learning regression model based on preset residual prediction sample data to obtain the preset residual prediction model, the method for calculating the temperature field of bars and wires further includes: Preprocess the collected raw historical data on bar and wire production; The preprocessed data is used to construct and select features, and the constructed and selected feature variables are then standardized. The standardized feature variables and target variables are determined as the preset residual prediction sample data; the target variable is composed of the difference between the measured temperature corresponding to each discrete concentric ring of the bar and wire and the calculated temperature value of the mechanism model corresponding to the discrete point of the heat exchange and conduction mechanism model of the bar and wire.

[0036] The system collects six months of production data from the bar and wire rod production workshop, covering different batches, operating conditions, and steel grades. After data cleaning and time-series alignment, pre-defined residual prediction sample data is obtained. This pre-defined residual prediction sample data includes steel grade, carbon equivalent, and measurement temperature. (i.e., the measured temperatures corresponding to the concentric rings of each discrete bar / wire), and the temperature predicted by the mechanism model. (i.e., the calculated temperature values ​​of the mechanism model at each discrete point), rolling speed, total flow rate of the water-cooled box, tapping time, current time, estimated total cumulative deformation work, etc.

[0037] The residuals, i.e., the target variables, are calculated as follows:

[0038] in, and These are values ​​at the same location and at the same time. Therefore, each production batch corresponds to one sample at each same location, and all samples are aggregated to form a residual dataset, which is a preset residual prediction sample dataset composed of preset residual prediction sample data.

[0039] Preprocessing the collected raw historical data of bar and wire production can achieve the purpose of data cleaning. Specifically, this can include handling missing values, identifying and handling outliers using box plots, and aligning and resampling time-series data.

[0040] The construction and selection of features are explained below: Based on process knowledge, a feature set correlated with temperature prediction residuals is constructed from the cleaned data. This feature set includes material and specification features, process state features, time series and cumulative effect features, and environmental and boundary condition features.

[0041] The feature variables of different dimensions are standardized to meet the input requirements of the machine learning regression model.

[0042] The measured temperature corresponding to each discrete concentric ring of bar / wire is subtracted from the temperature calculated by the mechanistic model, and this subtraction is used as the target variable. Using the standardized feature variables and the target variable, a machine learning regression model is trained. Specifically, the machine learning regression model can be an XGBoost model, which approximates the target variable by iteratively constructing multiple decision trees. Its core lies in minimizing the following objective function:

[0043] in, It is the loss function, where n is the total number of preset residual prediction sample data. The loss function is used to measure the difference between the actual value of the temperature residual and the predicted value of the temperature residual.

[0044] This is a regularization term used to control the complexity of the preset residual prediction model and prevent overfitting. Let represent the prediction function of the k-th decision tree. In the gradient boosting framework, the final prediction is the cumulative sum of the predictions of all decision trees. k is the index of the decision tree. The total number of decision trees directly affects the capacity and fitting ability of the model. S is the number of leaf nodes, w is the weight, and γ and λ are both parameters.

[0045] The data was divided into training, parameter tuning, and test sets, with a ratio of 70%–15%–15%. The training set was used to fit the weight set of the XGBoost model, the parameter tuning set was used to learn and optimize γ and λ, and the test set was used to evaluate the model's fitting performance. The mean absolute error (MAE) was set to be within 5%.

[0046] Machine learning regression models can also be artificial neural network models, whose structure includes an input layer, at least two hidden layers, and an output layer. The network weights are adjusted through the backpropagation algorithm and gradient descent optimizer to minimize the prediction error.

[0047] This is achieved by constructing an Artificial Neural Network (ANN). ANNs can learn complex non-linear relationships between features and can also be used to process high-dimensional data. All parameters can be used as the input layer of the neural network model, and three fully connected layers can be constructed, with 128, 256, and 128 neurons per layer. The ReLU activation function is used to convert the residuals... T is used as the output layer. The Adam optimizer is used, and the number of training epochs is set to 200.

[0048] Table 1 shows the comparison of prediction accuracy for different steel grades: Table 1

[0049] The temperature field calculation results of bars and wires can also be sent to downstream systems or process control systems.

[0050] The method for calculating the temperature field of bars and wires also includes: Monitor the accuracy metrics of the preset residual prediction model and update the preset residual prediction model based on the monitoring results. Continuously monitor the accuracy metrics of the preset residual prediction model in practical applications, and issue an update alarm or initiate a model retraining step when the performance degrades to a preset threshold.

[0051] It can also display the predicted values ​​of the mechanism model, the corrected values ​​of the residuals, the final predicted values, and the key process parameters in the form of curves, charts or virtual diagrams, and provide historical prediction data and comparative analysis functions.

[0052] The method for calculating the temperature field of bars and wires provided in this invention has the following beneficial technical effects: Overcoming the limitations of traditional methods, this method achieves high-precision and highly adaptive prediction of the temperature field in the rolling and cooling processes of bar and wire rods, laying a technical foundation for the realization of digital and intelligent precise temperature control.

[0053] The method for calculating the temperature field of bar and wire rods provided in this invention involves acquiring bar and wire rod production data and mechanistic model input parameters, calculating the mechanistic model input parameters based on a pre-constructed bar and wire rod heat exchange and conduction mechanistic model, and obtaining mechanistic model temperature calculation values ​​corresponding to each discrete concentric ring of the bar and wire rod. The bar and wire rod heat exchange and conduction mechanistic model includes a plastic deformation work-heat factor in the rolling deformation zone. Temperature residuals are predicted based on a pre-trained preset residual prediction model of the bar and wire rod production data to obtain temperature residual calculation values ​​corresponding to each discrete concentric ring of the bar and wire rod. The mechanistic model temperature calculation values ​​at corresponding positions are then corrected based on the temperature residual calculation values ​​corresponding to each discrete concentric ring of the bar and wire rod to obtain the bar and wire rod temperature field calculation result. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data, which can improve the accuracy and applicability of temperature field calculation in the bar and wire rod production process.

[0054] In the above optional embodiments, the work-heat factor of plastic deformation is calculated according to the following formula:

[0055] in, The work-heat factor for plastic deformation, For work-heat conversion efficiency, The flow stress of the material is the temperature. ,strain and strain rate The function is described in the above examples and will not be repeated here.

[0056] In the above optional embodiments, the step of correcting the calculated temperature value of the mechanism model at the corresponding position based on the calculated temperature residual value corresponding to each discrete concentric ring of the bar / wire to obtain the calculated temperature field of the bar / wire includes: The sum of the calculated temperature residuals corresponding to each discrete bar / wire concentric ring and the calculated temperature from the mechanistic model at the corresponding position is taken as the temperature field calculation result of the bar / wire. This can be referred to the above embodiment for explanation, and will not be repeated here.

[0057] In the above optional embodiments, the preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data, including: The preset residual prediction model is obtained by using an objective function that includes a loss function and a regularization term, and training a machine learning regression model based on preset residual prediction sample data; the above embodiments can be referred to for explanation, and will not be repeated here.

[0058] The loss function measures the difference between the actual value and the predicted value of the temperature residual, while the regularization term controls the complexity of the preset residual prediction model. This can be further explained with reference to the above embodiments and will not be repeated here.

[0059] In the above optional embodiments, before the step of training a machine learning regression model based on preset residual prediction sample data to obtain the preset residual prediction model, the bar and wire temperature field calculation method further includes: The collected raw historical data on bar and wire production is preprocessed; this can be described with reference to the above embodiments and will not be repeated here.

[0060] The preprocessed data is then used to construct and select features, and the constructed and selected feature variables are then standardized. This can be referred to the above embodiments for explanation, and will not be repeated here.

[0061] The standardized feature variables and target variables are determined as the preset residual prediction sample data; the target variable is composed of the difference between the measured temperature corresponding to each discrete concentric ring of the bar / wire and the calculated temperature value of the mechanism model at the corresponding discrete point of the bar / wire heat exchange and conduction mechanism model. This can be referred to the above embodiment for further explanation, and will not be repeated here.

[0062] In the above optional embodiments, the method for calculating the temperature field of the bar and wire further includes: The accuracy index of the preset residual prediction model is monitored, and the preset residual prediction model is updated based on the monitoring results of the accuracy index. This can be referred to the above embodiments for explanation, and will not be repeated here.

[0063] Figure 3 This is a schematic diagram of the structure of a bar and wire temperature field calculation device provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the bar and wire temperature field calculation device provided in this embodiment of the invention includes an acquisition unit 301 and a calculation unit 302, wherein: The acquisition unit 301 is used to acquire bar and wire production data and mechanism model input parameters, and calculate the mechanism model input parameters based on the pre-constructed bar and wire heat exchange and conduction mechanism model to obtain the mechanism model temperature calculation values ​​corresponding to each discrete bar and wire concentric ring; wherein, the bar and wire heat exchange and conduction mechanism model includes the plastic deformation work heat factor in the rolling deformation zone; the calculation unit 302 is used to predict the temperature residual of the bar and wire production data according to the pre-trained preset residual prediction model to obtain the temperature residual calculation values ​​corresponding to each discrete bar and wire concentric ring, and correct the mechanism model temperature calculation values ​​at the corresponding positions according to the temperature residual calculation values ​​corresponding to each discrete bar and wire concentric ring to obtain the bar and wire temperature field calculation results; wherein, the preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0064] Specifically, the acquisition unit 301 in the device is used to acquire bar and wire production data and mechanism model input parameters, and calculates the mechanism model input parameters based on the pre-constructed bar and wire heat exchange and conduction mechanism model to obtain the mechanism model temperature calculation values ​​corresponding to each discrete bar and wire concentric ring; wherein, the bar and wire heat exchange and conduction mechanism model includes the plastic deformation work heat factor in the rolling deformation zone; the calculation unit 302 is used to predict the temperature residual of the bar and wire production data according to the pre-trained preset residual prediction model to obtain the temperature residual calculation values ​​corresponding to each discrete bar and wire concentric ring, and corrects the mechanism model temperature calculation values ​​at the corresponding positions according to the temperature residual calculation values ​​corresponding to each discrete bar and wire concentric ring to obtain the bar and wire temperature field calculation results; wherein, the preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0065] The present invention provides a method and apparatus for calculating the temperature field of bar and wire rods. This method acquires bar and wire rod production data and mechanistic model input parameters. Based on a pre-constructed heat exchange and conduction mechanism model for bar and wire rods, it calculates the temperature values ​​corresponding to each discrete concentric ring of the bar and wire rod. The heat exchange and conduction mechanism model includes a plastic deformation work-heat factor in the rolling deformation zone. A pre-trained residual prediction model is used to predict the temperature residuals of the bar and wire rod production data, obtaining calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar and wire rod. The calculated temperature values ​​of the mechanistic model at corresponding positions are then corrected based on these calculated temperature residual values ​​to obtain the final temperature field calculation result for the bar and wire rod. The pre-trained residual prediction model, obtained by training a machine learning regression model using pre-trained residual prediction sample data, improves the accuracy and applicability of the temperature field calculation during the bar and wire rod production process.

[0066] The embodiments of the present invention provide a bar and wire temperature field calculation device that can be used to execute the processing flow of the above-described method embodiments. Its functions will not be repeated here, but can be referred to the detailed description of the above-described method embodiments.

[0067] Figure 4 This is a schematic diagram of the physical structure of a computer device provided in an embodiment of the present invention, such as... Figure 4 As shown, the computer device includes: a memory 401, a processor 402, and a computer program stored in the memory 401 and executable on the processor 402. When the processor 402 executes the computer program, it implements the following method: Obtain the production data and mechanism model input parameters of bar and wire rods, and calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of bar and wire rods to obtain the mechanism model temperature calculation values ​​corresponding to each discrete concentric ring of bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; Based on the pre-trained preset residual prediction model, the temperature residual is predicted for the production data of the bar and wire, and the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire is obtained. Based on the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire, the calculated temperature value of the mechanism model at the corresponding position is corrected and calculated to obtain the temperature field calculation result of the bar and wire. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0068] This embodiment discloses a computer program product, which includes a computer program that, when executed by a processor, implements the following method: Obtain the production data and mechanism model input parameters of bar and wire rods, and calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of bar and wire rods to obtain the mechanism model temperature calculation values ​​corresponding to each discrete concentric ring of bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; Based on the pre-trained preset residual prediction model, the temperature residual is predicted for the production data of the bar and wire, and the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire is obtained. Based on the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire, the calculated temperature value of the mechanism model at the corresponding position is corrected and calculated to obtain the temperature field calculation result of the bar and wire. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0069] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the following method: Obtain the production data and mechanism model input parameters of bar and wire rods, and calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of bar and wire rods to obtain the mechanism model temperature calculation values ​​corresponding to each discrete concentric ring of bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; Based on the pre-trained preset residual prediction model, the temperature residual is predicted for the production data of the bar and wire, and the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire is obtained. Based on the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire, the calculated temperature value of the mechanism model at the corresponding position is corrected and calculated to obtain the temperature field calculation result of the bar and wire. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

[0070] Compared with existing technologies, the present invention provides a method for calculating the temperature field of bar and wire rods. This method acquires bar and wire rod production data and mechanistic model input parameters. Based on a pre-constructed heat exchange and conduction mechanism model for bar and wire rods, it calculates the temperature values ​​corresponding to each discrete concentric ring of the bar and wire rod. The heat exchange and conduction mechanism model includes a plastic deformation work-heat factor in the rolling deformation zone. A pre-trained residual prediction model is used to predict the temperature residuals of the bar and wire rod production data, obtaining the calculated temperature residuals corresponding to each discrete concentric ring of the bar and wire rod. The calculated temperature residuals are then used to correct the calculated temperature values ​​of the mechanistic model at the corresponding positions, resulting in the final temperature field calculation. The pre-trained residual prediction model, obtained by training a machine learning regression model based on pre-trained residual prediction sample data, improves the accuracy and applicability of the temperature field calculation during the bar and wire rod production process.

[0071] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0072] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0073] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0074] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0075] In the description of this specification, the references to terms such as "an embodiment," "a specific embodiment," "some embodiments," "for example," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0076] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. 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 method for calculating the temperature field of bars and wires, characterized in that, include: Obtain the production data and mechanism model input parameters of bar and wire rods, and calculate the mechanism model input parameters based on the pre-constructed heat exchange and conduction mechanism model of bar and wire rods to obtain the mechanism model temperature calculation values ​​corresponding to each discrete concentric ring of bar and wire rods. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; Based on the pre-trained preset residual prediction model, the temperature residual is predicted for the production data of the bar and wire, and the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire is obtained. Based on the calculated temperature residual corresponding to each discrete concentric ring of the bar and wire, the calculated temperature value of the mechanism model at the corresponding position is corrected and calculated to obtain the temperature field calculation result of the bar and wire. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

2. The method for calculating the temperature field of bars and wires according to claim 1, characterized in that, The work-heat factor of plastic deformation is calculated according to the following formula: in, The work-heat factor for plastic deformation, For work-heat conversion efficiency, The flow stress of the material is the temperature. ,strain and strain rate The function.

3. The method for calculating the temperature field of bars and wires according to claim 1, characterized in that, The step of correcting the calculated temperature values ​​of the mechanism model at the corresponding positions based on the calculated temperature residual values ​​corresponding to each discrete concentric ring of the bar / wire to obtain the calculated temperature field of the bar / wire includes: The sum of the calculated temperature residuals corresponding to each discrete bar and wire concentric ring and the calculated temperature of the mechanism model at the corresponding position is taken as the temperature field calculation result of the bar and wire.

4. The method for calculating the temperature field of bars and wires according to claim 1, characterized in that, The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data, including: The preset residual prediction model is obtained by using an objective function that includes a loss function and a regularization term, and training a machine learning regression model based on preset residual prediction sample data. The loss function is used to measure the difference between the actual value of the temperature residual and the predicted value of the temperature residual, and the regularization term is used to control the complexity of the preset residual prediction model.

5. The method for calculating the temperature field of bars and wires according to claim 4, characterized in that, Before the step of training a machine learning regression model based on preset residual prediction sample data to obtain the preset residual prediction model, the method for calculating the temperature field of bars and wires further includes: Preprocess the collected raw historical data on bar and wire production; The preprocessed data is used to construct and select features, and the constructed and selected feature variables are then standardized. The standardized feature variables and target variables are determined as the preset residual prediction sample data; the target variable is composed of the difference between the measured temperature corresponding to each discrete concentric ring of the bar and wire and the calculated temperature value of the mechanism model corresponding to the discrete point of the heat exchange and conduction mechanism model of the bar and wire.

6. The method for calculating the temperature field of bars and wires according to any one of claims 1 to 5, characterized in that, The method for calculating the temperature field of bars and wires also includes: Monitor the accuracy index of the preset residual prediction model, and update the preset residual prediction model based on the monitoring results of the accuracy index.

7. A device for calculating the temperature field of bars and wires, characterized in that, include: The acquisition unit is used to acquire bar and wire production data and mechanism model input parameters, and calculate the mechanism model input parameters based on the pre-constructed bar and wire heat exchange and conduction mechanism model to obtain the mechanism model temperature calculation values ​​corresponding to each discrete bar and wire concentric ring. The heat exchange and conduction mechanism model for bars and wires includes the heat factor of plastic deformation work in the rolling deformation zone; The calculation unit is used to predict the temperature residual of the bar and wire production data according to the preset residual prediction model obtained by pre-training, to obtain the temperature residual calculation value corresponding to each discrete bar and wire concentric ring, and to correct the temperature calculation value of the mechanism model at the corresponding position according to the temperature residual calculation value corresponding to each discrete bar and wire concentric ring, so as to obtain the bar and wire temperature field calculation result. The preset residual prediction model is obtained by training a machine learning regression model based on preset residual prediction sample data.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

Images