Anti-deformation self-adaptive bracket system for heat treatment of large steel plate

Abstract

The invention discloses an anti-deformation self-adaptive bracket system for heat treatment of a large steel plate, and belongs to the technical field of metal heat treatment. Steel plate basic parameters and heat treatment process data are acquired through a data acquisition module, a digital twinning construction module constructs and updates a digital twinning model in real time based on the steel plate basic parameters and the heat treatment process data, and a data analysis module analyzes deformation and form deviation of a steel plate in heating, heat preservation and quenching stages based on the digital twinning model. The strategy generation module generates a high-precision control strategy based on an analysis result of the data analysis module to drive the adaptive bracket to perform dynamic adjustment, the system has a self-learning optimization capability, digital twin model parameters are continuously calibrated through error feedback, the prediction accuracy is improved, and the prediction efficiency is improved. The auxiliary repairing module conducts multi-point-position high-precision correction and local heat treatment repairing on the deformed steel plate, the deformation problem in the heat treatment process is effectively restrained, and the product quality and the production stability are improved.

Description

Technical Field

[0001] This invention belongs to the field of metal heat treatment technology, and specifically discloses an anti-deformation adaptive bracket system for heat treatment of large steel plates. Background Technology

[0002] During heat treatment, large steel plates are prone to plastic deformation such as warping, denting, and twisting due to uneven heating, phase transformation stress, and self-weight. This seriously affects the flatness and dimensional accuracy of the product, and may even lead to the scrapping of the workpiece. Traditional heat treatment brackets mostly use rigid supports or simple adjustable structures, which lack the ability to monitor and dynamically respond to changes in the shape of the steel plate during heat treatment, and cannot effectively suppress or compensate for deformation.

[0003] Currently, the industry largely relies on experience to set process parameters and then corrects deformation through subsequent mechanical straightening. However, this method is inefficient, costly, and requires highly skilled operators. Although some studies have attempted to introduce sensors to monitor temperature or deformation, they still lack systematic digital modeling, real-time prediction, and closed-loop control capabilities, making it difficult to achieve precise deformation prevention and adaptive adjustment. Existing technological pain points:

[0004] Passive support: Traditional brackets are rigid fixed structures, which cannot cope with the complex internal stress changes of steel plates caused by thermal expansion and contraction and phase transformation during heat treatment, resulting in uncontrollable deformation;

[0005] Deformation is unpredictable: Deformation is the result of heat treatment and cannot be intervened during the process. It can only be corrected or scrapped afterward, which is costly.

[0006] Multi-factor coupling: Deformation is the result of the coupling effect of multiple factors such as steel plate composition, heating / cooling rate, and temperature uniformity, which is difficult to quantify and analyze using traditional experience;

[0007] Poor consistency: The deformation of steel plates in different batches and heats varies greatly, resulting in unstable product quality;

[0008] Therefore, it is necessary to invent an anti-deformation adaptive bracket system for heat treatment of large steel plates to solve the above problems. Summary of the Invention

[0009] To overcome the aforementioned deficiencies in the prior art, this invention provides an anti-deformation adaptive bracket system for heat treatment of large steel plates. The system utilizes a data acquisition module to collect basic steel plate parameters and heat treatment process data. A digital twin construction module builds and updates a digital twin model in real time based on these parameters. A data analysis module analyzes the deformation trends and morphological deviations of the steel plate during the heating, holding, and quenching stages using the digital twin model. A strategy generation module generates a high-precision control strategy based on the analysis results to drive the adaptive bracket to make dynamic adjustments. The system possesses self-learning optimization capabilities, continuously calibrating the digital twin model parameters through error feedback to improve prediction accuracy. An auxiliary repair module performs multi-point high-precision correction and local heat treatment repair on deformed steel plates, effectively solving the problems mentioned in the background art.

[0010] To achieve the above objectives, the present invention provides the following technical solution: a deformation-resistant adaptive bracket system for heat treatment of large steel plates, comprising a data acquisition module, a data analysis module, an adaptive array bracket, a digital twin construction module, a strategy generation module, a self-learning optimization module, and an auxiliary repair module;

[0011] Data acquisition module: Collects basic parameters of the steel plate and data on the heat treatment process;

[0012] Digital twin construction module: Based on the basic parameters of the steel plate, the preset parameters of the heat treatment process are integrated synchronously to generate a digital twin model that is dynamically mapped to the steel plate. During the heat treatment process, the heat treatment process data is received in real time to update the temperature distribution and morphology simulation results of the model. The digital twin model outputs the predicted morphology data of the steel plate based on the heat treatment process data.

[0013] Data Analysis Module: Based on digital twin model and heat treatment process data analysis, this module analyzes the deformation and morphological deviation of steel plates during heat treatment and outputs the analysis results.

[0014] Strategy generation module: Generates control strategies for adaptive array brackets based on the analysis results from the data analysis module;

[0015] Adaptive array bracket: Receives and executes control strategies generated by other modules;

[0016] Self-learning optimization module: Compares the morphological data of the steel plate after furnace exit predicted by the digital twin model with the actual morphological data of the steel plate after furnace exit, and calculates the error value; feeds back key process parameters and error values ​​to the digital twin construction module to automatically calibrate the model parameters in the digital twin model; accumulates heat treatment process data of similar steel plates, and iteratively optimizes the fitting accuracy of the function f(C, ΔT, G, t) in the deviation mathematical model;

[0017] Auxiliary Repair Module: Based on the heat treatment process data and the morphological deviation of the steel plate, the optimal correction control strategy for the deformed steel plate is generated; multi-point adjustment is performed on the deformed area through an adaptive array bracket, and repair is achieved by combining multiple local heat treatments.

[0018] Preferably, the basic parameters of the steel plate include the material composition, geometric dimensions, and initial flatness of the steel plate; the heat treatment process data includes temperature field distribution data, morphological data, and heating / cooling rate data; the morphological data includes three-dimensional point cloud data of the steel plate surface.

[0019] Preferably, the preset parameters of the heat treatment process include target temperature, holding time, heating rate, type and flow rate of cooling medium, and positioning data describing the initial position of the steel plate on the bracket.

[0020] Preferably, the method for analyzing the morphological deviation is as follows: by comparing the morphological data collected in real time by the data acquisition module with the expected morphology, the deviation of local bulges or depressions is calculated; the key point is the location of the electric servo cylinder of the adaptive array bracket.

[0021] Preferably, the deformation is analyzed by subtracting the morphological data collected in real time by the data acquisition module from the predicted morphological data output by the digital twin model to obtain the real-time deformation.

[0022] Preferably, the steps for generating the control strategy are as follows:

[0023] Calculate the feedforward control quantity based on the predicted deformation;

[0024] Calculate feedback control quantity based on morphological deviation;

[0025] The total control quantity is generated by combining the feedforward control quantity and the feedback control quantity.

[0026] Preferably, the feedforward control quantity is the predicted deformation quantity multiplied by the feedforward gain coefficient K. ff The feedback control quantity is obtained by multiplying the difference between the real-time shape deviation and the target shape deviation by the feedback gain coefficient K. fb get.

[0027] Preferably, the key process parameters include the carbon content of the steel plate and heat treatment process data; the function f in the deviation mathematical model is an empirical fitting function used to predict the maximum deformation of the steel plate under a specific process, and its mathematical expression is f(C, ΔT, G, t), where C represents the carbon content of the steel plate, ΔT represents the characteristic temperature difference, G represents the thickness-to-span ratio of the steel plate, and t represents the duration of a specific stage.

[0028] Preferably, the optimal correction control strategy is generated as follows: based on the morphological deviation data and the deviation mathematical model, the distribution of the support force to be applied and the heat treatment parameters are solved in reverse, and the adjustment command is generated with the goal of minimizing the residual stress. The heat treatment parameters include heating power and heating coordinates.

[0029] Preferably, the adaptive array bracket is a support matrix composed of m×n high-precision electric servo cylinders, and each cylinder can be independently controlled for lifting and lowering.

[0030] The technical effects and advantages of this invention are as follows:

[0031] 1. The system uses a digital twin model for real-time prediction and an adaptive array bracket to perform high-precision multi-point adjustments, achieving a revolutionary transformation from "passive support" to "active straightening". It can actively counteract the deformation caused by thermal stress and phase transformation during heat treatment. Through a feedforward-feedback composite control strategy and auxiliary repair function, it can effectively prevent irreversible excessive deformation of large steel plates during heat treatment, minimize scrap rate and subsequent straightening costs, and directly improve production yield and economic benefits.

[0032] 2. The constructed multi-physics coupled digital twin model can simulate the dynamic changes of temperature, stress and deformation during the heat treatment of steel plates with high fidelity, providing a scientific basis for prediction and control and reducing the reliance on traditional "trial and error" experience; based on comprehensive data acquisition and real-time morphological analysis, all control strategies and repair schemes are data-driven, making the decision-making process more objective, accurate and reliable.

[0033] 3. The self-learning optimization module can automatically reverse-calibrate key physical parameters in the digital twin model by utilizing the error between actual production data and predicted data, making the model more and more accurate as production data accumulates, and possessing the characteristic of "getting smarter with use"; through continuous iterative optimization of the deviation mathematical model f(C, ΔT, G, t), the system will continuously accumulate and solidify the process knowledge for processing similar steel plates, forming a unique and continuously optimized intelligent process database for the enterprise, thereby improving the overall process level;

[0034] 4. The system not only focuses on preventing deformation but also provides backup solutions for correcting it. For deviations that have already occurred, online repair can be performed through precise mechanical adjustments and controlled local heat treatment, maximizing product recovery and providing dual protection. From monitoring and analysis to control and optimization, the entire process is highly automated, reducing over-reliance on operator experience, minimizing quality fluctuations caused by human factors, and improving the stability and consistency of the production process. The online repair capability can reduce the time that steel plates need to be taken offline due to deformation, accelerating the production process and improving equipment utilization and capacity. Attached Figure Description

[0035] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0036] Figure 1 This is a schematic diagram of the overall structural modules and data interaction of the present invention.

[0037] Figure 2 This is a flowchart illustrating the overall steps of the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] This invention provides, for example Figure 1 The anti-deformation adaptive bracket system for heat treatment of large steel plates shown includes a data acquisition module, a data analysis module, an adaptive array bracket, a digital twin construction module, a strategy generation module, a self-learning optimization module, and an auxiliary repair module.

[0040] like Figure 2 As shown, the overall process flow of the present invention is as follows:

[0041] The data acquisition module collects basic parameters of the steel plate and data on the heat treatment process;

[0042] Furthermore, in the above technical solution, the basic parameters of the steel plate include the material composition, geometric dimensions, and initial flatness of the steel plate; the heat treatment process data includes temperature field distribution data, morphological data, and heating / cooling rate data; the morphological data includes three-dimensional point cloud data of the steel plate surface.

[0043] Furthermore, the material composition and geometric dimensions of the steel plate are obtained through the upstream production system interface. Specifically, the material composition of the steel plate can be analyzed by a spectrometer, and the geometric dimensions of the steel plate can be collected by a laser rangefinder. The morphological data is obtained by collecting point cloud data of the steel plate surface using a 3D laser scanner, and the data of each point includes its three-dimensional coordinates (xi, yi, zi). The initial flatness is calculated by a point cloud processing algorithm. The temperature field distribution data is obtained by deploying an infrared thermal imager to monitor the surface temperature of the steel plate in real time. The heating / cooling rate data, i.e., the rate of change of temperature over time, is recorded by a high-frequency temperature sensor.

[0044] Furthermore, the calculation process for the initial flatness is as follows:

[0045] The least squares method is used to fit the preprocessed point cloud data to a plane to obtain the "best fitting plane" that best represents the overall tilt state of the steel plate. The equation of this reference plane can be expressed as: Ax + By + Cz + D = 0, where A, B, and C are normal vectors and D is a constant.

[0046] Calculate the distance from each point in the point cloud to the reference plane, i.e., the deviation value di of that point. The calculation formula is: ;

[0047] Using the formula: Calculate the standard deviation of all point deviation values ​​di, which is used as a flatness index to represent flatness. In the formula, μ is the average value of all di, and n is the total number of points.

[0048] The digital twin construction module is based on the basic parameters of the steel plate and synchronously integrates the preset parameters of the heat treatment process to generate a digital twin model that is dynamically mapped to the steel plate. During the heat treatment process, the module receives heat treatment process data in real time and updates the temperature distribution and morphology simulation results of the model. The digital twin model outputs the predicted morphology data of the steel plate based on the heat treatment process data.

[0049] Furthermore, in the above technical solution, the preset parameters of the heat treatment process include target temperature, holding time, heating rate, cooling medium type and flow rate, and positioning data describing the initial position of the steel plate on the bracket.

[0050] Furthermore, the construction process of the digital twin model is as follows:

[0051] Based on the basic parameters of the steel plate and the preset process parameters, a multi-physics coupling simulation technology, such as finite element analysis (FEA), is used to construct an initial digital twin model. The model covers physical processes such as heat conduction, phase transformation, thermal stress, and elastic-plastic deformation.

[0052] During the heat treatment process, temperature field and morphology data are received in real time; the temperature distribution and predicted morphology of the model are dynamically updated to keep it synchronized with the physical entity; and the predicted morphology data of the steel plate are output in real time to provide data support for subsequent analysis and control modules.

[0053] The data analysis module analyzes the deformation and morphological deviation of the steel plate during heat treatment based on the digital twin model and heat treatment process data, and outputs the analysis results.

[0054] Furthermore, in the above technical solution, the method for analyzing the morphological deviation is as follows: by comparing the morphological data collected in real time by the data acquisition module with the expected morphology, the deviation of local bulges or depressions is calculated; the key point is the location of the electric servo cylinder of the adaptive array bracket, and the expected morphology is the ideal morphology of the steel plate at the current moment during the standard heat treatment process.

[0055] Furthermore, the specific analysis process of the morphological deviation is as follows:

[0056] Find the points corresponding to the key points in both the actual point cloud and the predicted point cloud;

[0057] For each keypoint, calculate the difference between its actual height and expected height in the normal direction. The calculation formula is δi=zia−ziw; δi>0 indicates that the actual position of the point is higher than the predicted position, and δi<0 indicates that the actual position of the point is lower than the predicted position; in the formula, zia represents the actual height of keypoint i in the z-axis direction, and ziw represents the expected height of keypoint i in the z-axis direction.

[0058] It should be further explained that zia, ziw, and zip all originate from morphological data, that is, the z-coordinate of the point cloud data corresponding to the key points;

[0059] Furthermore, in the above technical solution, the deformation amount is analyzed by subtracting the morphological data collected in real time by the data acquisition module from the predicted morphological data output by the digital twin model to obtain the real-time deformation amount.

[0060] Furthermore, the specific analysis process for the deformation amount is as follows:

[0061] The real-time acquired morphological data is compared point by point with the predicted morphological data, and the deformation value ΔS of each key point is calculated to obtain a deformation matrix. The deformation amount is obtained by summing the deformation values ​​ΔS in the deformation matrix. The calculation formula is S=∑|ΔS|; where ΔS=zia-zip, and zip represents the predicted height of key point i in the z-axis direction.

[0062] The strategy generation module generates an adaptive array bracket control strategy based on the analysis results of the data analysis module;

[0063] Furthermore, in the above technical solution, the steps for generating the control strategy are as follows:

[0064] Calculate the feedforward control quantity based on the predicted deformation;

[0065] Furthermore, the formula for calculating the feedforward control quantity Uff at the key point is: Uff=Kff×ΔS, where Kff is the feedforward gain coefficient, which is pre-calibrated based on material properties and process parameters;

[0066] Calculate feedback control quantity based on morphological deviation;

[0067] Furthermore, the formula for calculating the feedback control quantity Ufb at the key point is: Ufb=Kfb(δa−δt), where Kfb is the feedback gain coefficient, determined by the fuzzy control algorithm, δa is the actual shape deviation, and δt is the target shape deviation, which is usually 0, i.e., no deviation.

[0068] The total control quantity is generated by combining the feedforward control quantity and the feedback control quantity.

[0069] The formula for calculating the total control quantity Ut at the next critical point is: Ut = Uff + Ufb;

[0070] The adaptive array bracket receives and executes control strategies generated by other modules;

[0071] The self-learning optimization module compares the morphological data of the steel plate after it comes out of the furnace predicted by the digital twin model with the actual morphological data of the steel plate after it comes out of the furnace, and calculates the error value; it feeds back the key process parameters and error value to the digital twin construction module, automatically calibrates the model parameters in the digital twin model; it accumulates the heat treatment process data of similar steel plates, iteratively optimizes the fitting accuracy of the function f in the deviation mathematical model, and reduces the subsequent prediction deviation;

[0072] Furthermore, in the above technical solution, the key process parameters include: heat treatment process data and carbon content of the steel plate during the heat treatment process.

[0073] Furthermore, the function f in the aforementioned deviation mathematical model is an empirical fitting function used to predict the maximum deformation of a steel plate under a specific process. Its mathematical expression is f(C, ΔT, G, t), where C represents the carbon content of the steel plate, a key factor affecting phase transformation behavior and hardenability; ΔT represents the characteristic temperature difference, which can be defined as the maximum temperature difference between the surface and core of the steel plate during heating or cooling, or the maximum temperature difference between the two ends along its length, and is the main driving source of thermal stress; G represents the thickness-to-span ratio of the steel plate, i.e., the ratio of thickness to the maximum support span, used to quantify the bending stiffness of the steel plate under its own weight and thermal stress; t represents the duration of a specific stage, such as the heating stage, holding stage, or quenching stage. The initial form of this function can be preset as a linear polynomial, for example: f = w C ·C+w b ·ΔT+w G ·G+w d ·t+e, where w C w b w G w d Here, e represents the initial weight coefficients. The goal of the self-learning optimization module is to continuously adjust these coefficients through regression analysis, and may introduce cross terms, such as C*ΔT or higher-order terms, to improve the function's prediction accuracy for the final deformation.

[0074] Furthermore, the definition of a similar steel plate is a carbon content deviation ∈ [-0.02, 0.02]% and a geometric dimension deviation ∈ [-5, 5]%; the geometric dimensions include length, width and thickness.

[0075] Furthermore, the model parameters in the digital twin model are calibrated using the least squares method or gradient descent method, employing an error backpropagation mechanism to gradually reduce the prediction error; the optimization process of the model parameters is as follows:

[0076] Error Calculation and Quantification: The shape data of the steel plate after being unloaded from the furnace predicted by the digital twin model and the actual shape data of the steel plate after being unloaded from the furnace are compared. The height deviation is calculated point by point at the same key points. The root mean square error is used as the quantitative index of the overall error, i.e., the error value.

[0077] Determine the parameters to be calibrated: Select key physical parameters that are sensitive to the prediction results and have high uncertainty as the parameters to be calibrated, which usually include: thermal expansion coefficient, phase transformation plasticity coefficient and phase transformation expansion coefficient;

[0078] Example:

[0079] If the system finds that the predicted deformation is generally less than the actual deformation, it may indicate that the thermal expansion coefficient or phase transformation expansion coefficient in the model is underestimated. These two parameters should be prioritized as parameters to be calibrated.

[0080] Constructing the loss function and optimization problem: The root mean square error is usually used directly as the loss function. The optimization objective is to minimize the error between the model's predicted output and the actual data. The optimal value of the parameter to be calibrated is found by using the Bayesian optimization algorithm.

[0081] Furthermore, the optimization of the deviation mathematical model is achieved through a data-driven iterative loop: the system continuously collects heat treatment process parameters of similar steel plates, such as carbon content, heating rate, geometric dimensions, and actual deformation results. After cleaning and feature engineering, machine learning algorithms, such as random forests or gradient boosting trees, are used to train and fine-tune the prediction function f with the goal of minimizing the error between the predicted deformation and the actual deformation. The newly trained model is deployed for real-time prediction, and its performance is continuously monitored. A new round of training is periodically triggered using data of similar steel plates newly accumulated in production, thereby continuously evolving the model's prediction accuracy and ultimately forming a self-learning intelligent system that encapsulates the company's core process knowledge.

[0082] The auxiliary repair module generates the optimal correction control strategy for the deformed steel plate based on the heat treatment process data and the morphological deviation of the steel plate; it performs multi-point adjustment with a precision of ±0.01mm on the deformed area through an adaptive array bracket, and achieves repair by combining multiple local heat treatments.

[0083] Furthermore, in the above technical solution, the optimal correction control strategy is generated as follows: based on the morphological deviation data and the deviation mathematical model, the distribution of the support force to be applied and the heat treatment parameters are solved in reverse, and the adjustment command is generated with the goal of minimizing the residual stress. The heat treatment parameters include heating power and heating coordinates.

[0084] Furthermore, the inverse solution is a mathematical optimization process aimed at "minimizing residual stress," the core of which is to perform inverse calculations based on an established digital twin model or biased mathematical model.

[0085] The specific method is as follows:

[0086] Optimization variables: the forces to be applied at each support point and the parameters of local heat treatment;

[0087] Objective function: The primary objective is to minimize the overall residual stress of the corrected workpiece, while also taking into account the minimization of morphological deviations;

[0088] Constraints: The variables must be within the physical limits of the system, and the corrected steel plate shape must meet the flatness tolerance requirements;

[0089] Solution process: The system uses a digital twin model as a surrogate model and runs iterative optimization algorithms, such as gradient descent, genetic algorithm, or sequential quadratic programming, to solve the problem. Within the constraints of the variables, the algorithm automatically generates multiple sets of candidate combinations of support force distribution and heat treatment parameters. Each candidate combination is input into the digital twin model to quickly simulate and calculate the residual stress and final shape of the steel plate after applying the set of parameters. The objective function value is calculated based on the simulation results. The algorithm continuously updates and generates better parameter combinations until it finds the optimal solution that minimizes the objective function.

[0090] Furthermore, the local heat treatment refers to a non-uniform, precisely controllable heating and cooling process performed on a specific area that has already deformed.

[0091] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A deformation prevention self-adaptive bracket system for heat treatment of large steel plates, characterized in that, The system comprises a data acquisition module, a data analysis module, an adaptive array carrier, a digital twin construction module, a strategy generation module, a self-learning optimization module, and an auxiliary repair module. The data acquisition module acquires steel plate basic parameters and heat treatment process data. The digital twin construction module generates a digital twin model dynamically mapped to the steel plate based on the steel plate basic parameters and the preset heat treatment process parameters, and updates the temperature distribution and shape simulation results of the model in real time during the heat treatment process according to the heat treatment process data. The data analysis module analyzes the deformation and shape deviation of the steel plate during the heat treatment process based on the digital twin model and the heat treatment process data, and outputs the analysis results. The strategy generation module generates a control strategy for the adaptive array carrier based on the analysis results of the data analysis module. The adaptive array carrier receives and executes the control strategy generated by other modules. The self-learning optimization module compares the predicted shape data of the steel plate after being discharged from the heat treatment process with the actual shape data of the steel plate after being discharged, calculates the error value, feeds back the key process parameters and the error value to the digital twin construction module, automatically calibrates the model parameters in the digital twin model, accumulates the heat treatment process data of the same type of steel plate, and iteratively optimizes the fitting precision of the function f(C, ΔT, G, t) in the deviation mathematical model. The auxiliary repair module generates an optimal correction control strategy for the deformed steel plate based on the heat treatment process data and the shape deviation of the steel plate, and adjusts the deformed area through the adaptive array carrier and realizes repair through multiple local heat treatments.

2. A self-adaptive deformation-prevention cradle system for heat treatment of large steel plates as claimed in claim 1, characterized in that: The steel plate basic parameters include the material composition, geometric size, and initial flatness of the steel plate.

3. A self-adaptive cradle system for preventing deformation of a large steel plate during heat treatment according to claim 1, wherein: The heat treatment process preset parameters include the target temperature, holding time, heating rate, cooling medium type and its flow rate, and positioning data for describing the initial position of the steel plate on the carrier.

4. A self-adaptive cradle system for preventing deformation of a large steel plate during heat treatment according to claim 1, wherein: The analysis method of the shape deviation is to calculate the local bulge or depression deviation by comparing the real-time shape data collected by the data acquisition module with the expected shape.

5. A self-adaptive cradle system for preventing deformation of large steel plates during heat treatment as claimed in claim 1, wherein: The analysis method of the deformation amount is to calculate the real-time deformation amount by subtracting the predicted shape data output by the digital twin model from the real-time shape data collected by the data acquisition module.

6. A self-adapting cradle system for preventing deformation of large steel plates during heat treatment as claimed in claim 1, wherein: The generation steps of the control strategy are as follows: Calculate the feedforward control amount based on the predicted deformation amount. Calculate the feedback control amount based on the shape deviation. Combine the feedforward control amount and the feedback control amount to generate the total control amount.

7. A self-adapting cradle system for preventing deformation of large steel plates during heat treatment, as claimed in claim 6, wherein: The feedforward control amount is obtained by multiplying the predicted deformation amount by a feedforward gain coefficient K ff The feedback control amount is obtained by multiplying the difference between the real-time shape deviation and the target shape deviation by a feedback gain coefficient K fb .

8. A self-adapting cradle system for preventing deformation of large steel plates during heat treatment as claimed in claim 1, wherein: The key process parameters include the carbon content of the steel plate and the heat treatment process data. The function f in the deviation mathematical model is an empirical fitting function for predicting the maximum deformation amount of the steel plate under a specific process, and its mathematical expression is f(C, ΔT, G, t), where C represents the carbon content of the steel plate, ΔT represents the characteristic temperature difference, G represents the thickness-span ratio of the steel plate, and t represents the duration of a specific stage.

9. A self-adapting cradle system for preventing deformation of large steel plates during heat treatment as claimed in claim 1, wherein: The optimal correction control strategy is generated in the following manner: based on the shape deviation data and the deviation mathematical model, the support force distribution and the heat treatment parameters to be applied are reversely solved, adjustment instructions are generated with the objective of minimizing residual stress, and the heat treatment parameters include heating power and heating coordinates.

10. A self-adapting cradle system for preventing deformation of large steel plates during heat treatment as claimed in claim 1, wherein: The adaptive array bracket is composed of a support matrix of m*n high-precision electric servo cylinders, and each cylinder body can be independently controlled to rise and fall.

Images