Wide and thick steel plate quality detection method and device based on data analysis and optimization

Abstract

The invention discloses a wide and thick steel plate quality detection method and device based on data analysis and optimization. The method comprises the following steps: acquiring real-time production process data of a wide and thick steel plate to be detected, and producing design standard data of the wide and thick steel plate to be detected; predicting the real-time production process data by adopting preset point prediction models corresponding to different mechanical property indexes to obtain point prediction values of the mechanical property indexes corresponding to the real-time production process data; performing conformal prediction on the real-time production process data by adopting a preset mechanical property interval prediction model based on the predicted value of each point to obtain a prediction interval of each mechanical property index corresponding to the real-time production process data; and performing quality detection on the to-be-detected wide and thick steel plate based on the prediction interval corresponding to the real-time production process data and the design standard data to obtain a quality detection result of the to-be-detected wide and thick steel plate. According to the method, the quality detection efficiency can be improved, and the quality detection accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of quality inspection, and in particular to a method and device for inspecting the quality of wide and thick steel plates based on data analysis and optimization. Background Art

[0002] As a key steel material, heavy plate is widely used in shipbuilding, bridge construction, building structures, pressure vessels, energy pipelines, and machinery manufacturing. Its high strength, excellent toughness, and ductility make it an irreplaceable force in withstanding heavy loads, complex environments, and high-stress conditions. Wide and heavy plate products are widely used in critical infrastructure and high-end equipment, often facing complex mechanical forces and harsh external conditions. Therefore, to ensure the safety, reliability, and long-term performance of these products in practical applications, downstream companies have imposed stricter requirements for the stability and consistency of heavy plate quality, meeting diverse performance requirements and high industry standards. Furthermore, as the application areas of heavy plate products continue to expand and customer demands become increasingly personalized and diversified, heavy plate manufacturers are shifting from large-scale standardized production to large-scale customized production. However, the production process for heavy plate is typically multi-stage, encompassing key steps such as steelmaking, continuous casting, multiple rolling passes, and heat treatment, resulting in a complex production process. This large-scale customized production model can lead to frequent switching of production routes, increasing the risk of errors and variability in product quality. At the same time, the small-batch, high-variety production process makes it difficult for a unified quality control system to cover all customized products, placing higher demands on process control and product quality assessment. Therefore, accurate and efficient online assessment of the quality of wide and heavy plate products is a key link in achieving process optimization and stable product delivery.

[0003] Currently, the mechanical quality assessment of thick and heavy plate products primarily relies on sampling inspection. This involves sampling representative plates and conducting laboratory tests and assays to determine their mechanical performance indicators, thereby inferring the quality level of the entire batch. However, traditional laboratory testing methods suffer from high costs and time delays, which can delay delivery. Furthermore, small-scale samples are difficult to characterize the full-scale mechanical properties of thick and heavy plate and fail to reflect fluctuations during the production process. Physical metallurgical principles indicate that the mechanical properties of thick and heavy plate products are directly determined by their microstructure, which is primarily determined by the composition ratio and rolling process. Based on this, researchers have utilized expert knowledge and data-driven approaches to predict the mechanical properties of thick and heavy plate. Expert knowledge approaches leverage domain knowledge—accumulated associations between material composition, processing, and properties—to establish a rule base and design specific inference mechanisms to infer the mechanical properties of metal materials. However, for multi-specification products targeted for mass customization, these approaches suffer from limitations such as insufficient knowledge acquisition and unstable inference results. The data-driven approach is to train a regression model on a large amount of historical inspection and testing data, using machine learning methods to characterize the mapping relationship between composition, process, structure, and performance, and thus achieve regression modeling of the mechanical properties of thick and wide plates. However, for thick and wide plate products with uneven mechanical property distribution, existing data-driven methods have the following problems: they only provide a single point prediction value, lack an assessment of the credibility of the prediction value and quantify the uncertainty of the model. If the point prediction result is directly used as a criterion for quality judgment, it is easy to misjudge unqualified products; such models often lack the guidance of physical mechanisms, and their generalization ability and interpretability are limited, making it difficult to effectively capture the inherent changes in the mechanical properties of materials under complex processing conditions; existing data-driven models of the mechanical properties of thick and wide plates are usually based on fixed mapping relationships constructed offline, lack online adaptive capabilities, and cannot update the model in real time to reflect new mechanical behavior characteristics when dynamic environments or complex process conditions change. Therefore, although existing technologies can predict the mechanical properties of thick and heavy plate products, they cannot provide the uncertainty information predicted by the model to reduce the risk of quality judgment. Moreover, the model lacks an online adaptive mechanism, making it difficult to achieve real-time correction and prediction. It is also unable to cope with the frequent changes in actual production process conditions under large-scale customization and cannot directly participate in the quality judgment process of thick and heavy plate products. Quality assurance during the delivery of thick and heavy plate products still relies on sampling inspection based on laboratory testing. Summary of the Invention

[0004] In view of this, the present invention provides a method and device for wide and thick steel plate quality detection based on data analysis and optimization, the main purpose of which is to solve the current problem of inaccurate quality detection of wide and thick steel plates.

[0005] To solve the above problems, the present application provides a method for detecting the quality of wide and thick steel plates based on data analysis and optimization, comprising:

[0006] Acquire real-time production process data of the thick and heavy steel plate to be tested and design standard data for producing the thick and heavy steel plate to be tested;

[0007] Predicting the real-time production process data using preset point prediction models corresponding to different mechanical performance indicators to obtain point prediction values ​​of each mechanical performance indicator corresponding to the real-time production process data;

[0008] Based on the point prediction values, a preset mechanical property interval prediction model is used to perform conformal prediction on the real-time production process data to obtain a prediction interval of each mechanical property indicator corresponding to the real-time production process data;

[0009] The quality inspection of the wide and thick steel plate to be inspected is performed based on the prediction interval corresponding to the real-time production process data and the design standard data to obtain a quality inspection result of the wide and thick steel plate to be inspected.

[0010] Optionally, before predicting the real-time production process data using the preset point prediction models corresponding to different mechanical performance indicators, the method further includes: constructing a preset point prediction model;

[0011] The construction of the preset point prediction model specifically includes:

[0012] Acquiring historical experimental test data of the wide and thick steel plate to be tested;

[0013] Performing feature extraction on the historical experimental detection data to obtain feature data;

[0014] Arrhenius rate equation is used to convert temperature characteristic data in the characteristic data to obtain a target characteristic data set;

[0015] Annotating the target feature dataset to obtain a label dataset;

[0016] The labeled data set is used as a training sample to construct a model using the Lasso method to obtain a preset point prediction model corresponding to different mechanical performance indicators;

[0017] The mechanical performance indicators include yield strength, tensile strength, elongation and average impact energy.

[0018] Optionally, the label data set is used as a training sample to construct a model using the Lasso method to obtain a preset point prediction model corresponding to different mechanical performance indicators, specifically including:

[0019] Divide the labeled data set into sample sets according to a preset ratio to obtain a training set, a calibration set, and a test set;

[0020] Normalizing the first label data in the training set to obtain a normalized first feature data set;

[0021] Based on the first feature data set and the label data set, model training is performed respectively from different mechanical performance index dimensions using the Lasso method to obtain the preset point prediction model corresponding to each of the mechanical performance indicators.

[0022] Optionally, the method further includes:

[0023] Acquire experimental test data of the thick and wide steel plate to be tested for a preset time period before the current moment;

[0024] Based on the experimental test data as training samples, the model is retrained using the Lasso method to obtain current point prediction models corresponding to different mechanical performance indicators;

[0025] The preset point prediction model is updated based on the current point prediction model.

[0026] Optionally, before performing conformal prediction on the real-time production process data using a preset mechanical property interval prediction model based on each of the point prediction values, the method further comprises: constructing a preset mechanical property interval prediction model;

[0027] The construction of the preset mechanical property interval prediction model specifically includes:

[0028] Normalizing the second label data in the calibration set to obtain normalized second feature data;

[0029] Using the preset point prediction model to predict the second feature data, to obtain first point prediction values ​​of different mechanical performance indicators corresponding to the second label data;

[0030] Performing calculations based on the predicted values ​​of each of the first points and the true values ​​annotated by the second label data to obtain first inconsistency scores for different mechanical performance indicators corresponding to the second label data;

[0031] Performing calculations based on each of the first inconsistency scores and a predetermined error coverage level to obtain first empirical quantiles of different mechanical performance indicators corresponding to the second label data;

[0032] A model is constructed based on each of the first empirical quantiles and each of the first point prediction values ​​to obtain the preset mechanical property interval prediction model.

[0033] Optionally, the method further includes:

[0034] Performing calculations based on the experimental test data and point prediction values ​​corresponding to the experimental test data obtained by using the preset point prediction model to obtain second inconsistency scores corresponding to different mechanical performance indicators;

[0035] Performing calculation based on the second inconsistency score using a preset objective function to obtain a second empirical quantile at the current moment;

[0036] Calculating the second empirical quantile using a sub-gradient descent optimization method to obtain a third empirical quantile at the next moment;

[0037] The preset mechanical property interval prediction model is updated based on the third empirical quantile.

[0038] Optionally, the method further includes:

[0039] Normalizing the third label data in the test set to obtain normalized third feature data;

[0040] Using the preset point prediction model to predict the third characteristic data, to obtain a second point prediction value of a different mechanical performance index corresponding to the third label data;

[0041] Performing conformal prediction using the updated preset mechanical property interval prediction model based on the second-point prediction value to obtain a prediction interval for each of the mechanical property indicators corresponding to the third characteristic data;

[0042] Performing calculations based on the second-point predicted value corresponding to each of the mechanical performance indicators, the true value corresponding to each of the mechanical performance indicators, and the prediction interval corresponding to each of the mechanical performance indicators to obtain an interval coverage index and an interval average width index;

[0043] The preset mechanical property interval prediction model is tested and adjusted based on the interval coverage index and the interval average width index to obtain the preset mechanical property interval prediction model that meets the preset accuracy requirements.

[0044] Optionally, the predicting of the real-time production process data using preset point prediction models corresponding to different mechanical performance indicators to obtain point prediction values ​​of the mechanical performance indicators corresponding to the real-time production process data specifically includes:

[0045] Normalizing the real-time production process data to obtain normalized production process characteristic data;

[0046] Based on the production process characteristic data, the preset point prediction model corresponding to different mechanical performance indicators is used to perform predictions to obtain point prediction values ​​of each mechanical performance indicator corresponding to the real-time production process data.

[0047] Optionally, performing quality inspection on the wide and thick steel plate to be inspected based on the prediction interval corresponding to the real-time production process data and the design standard data to obtain a quality inspection result of the wide and thick steel plate to be inspected specifically includes:

[0048] Extracting the upper bound value and the lower bound value of the prediction interval of each different mechanical performance index based on the prediction interval corresponding to the real-time production process data;

[0049] Comparing the upper limit value and the lower limit value of the prediction interval of the same mechanical property index with the design standard data to obtain a comparison result;

[0050] When the lower limit value of the prediction interval of the average impact energy mechanical performance index is greater than or equal to the minimum value of the standard impact energy, the lower limit value of the prediction interval of the yield strength mechanical performance index is greater than or equal to the minimum value of the standard yield strength, the lower limit value of the prediction interval of the tensile strength performance index is greater than or equal to the minimum value of the standard tensile strength and the upper limit value of the prediction interval of the tensile strength performance index is less than or equal to the maximum value of the standard tensile strength, and the lower limit value of the prediction interval of the elongation mechanical performance index is greater than or equal to the minimum value of the standard elongation, it is determined that the quality of the wide and thick steel plate to be tested is qualified.

[0051] To solve the above problems, this application provides a wide and thick steel plate quality detection device based on data analysis and optimization, comprising:

[0052] An acquisition module is used to acquire real-time production process data of the wide and thick steel plate to be tested and design standard data for producing the wide and thick plate to be tested;

[0053] a point prediction module, configured to predict the real-time production process data using preset point prediction models corresponding to different mechanical performance indicators, and obtain a point prediction value of each mechanical performance indicator corresponding to the real-time production process data;

[0054] a conformal prediction module, configured to perform conformal prediction on the real-time production process data using a preset mechanical property interval prediction model based on each of the point prediction values, and obtain a prediction interval for each of the mechanical property indicators corresponding to the real-time production process data;

[0055] The quality inspection module is used to perform quality inspection on the wide and thick steel plates to be inspected based on the prediction interval corresponding to the real-time production process data and the design standard data, and obtain quality inspection results of the wide and thick steel plates to be inspected.

[0056] The beneficial effects of this application are as follows: This application integrates the design standards of wide and thick plate products with the production process variables in the same database, and constructs an online real-time quality judgment system to provide confidence intervals for the mechanical properties of wide and thick plate products under a given confidence level, thereby realizing a dynamic comparison between the design standards of wide and thick plate products and the predicted performance, thereby accelerating the efficiency of product quality judgment and reducing the possibility of misjudgment of product quality.

[0057] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0059] Figure 1 A schematic flow chart of a method for detecting quality of wide and thick steel plates based on data analysis and optimization provided in an embodiment of the present application is shown;

[0060] Figure 2 A schematic flow chart of a method for detecting quality of wide and thick steel plates based on data analysis and optimization provided in an embodiment of the present application is shown;

[0061] Figure 3 A schematic diagram of point prediction and interval prediction results using the tensile strength mechanical performance indicator as an example is shown in the embodiments of the present application;

[0062] Figure 4 A schematic diagram showing the comparison results between the predicted range of mechanical performance indicators of thick and heavy plate products provided in the embodiments of the present application and the design standards;

[0063] Figure 5 A structural block diagram of a wide and thick steel plate quality detection device based on data analysis and optimization provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0064] Various aspects and features of the present application are described herein with reference to the accompanying drawings.

[0065] It should be understood that various modifications may be made to the embodiments of the present application. Therefore, the above description should not be considered as limiting, but merely as an example of an embodiment. Other modifications within the scope and spirit of the present application will occur to those skilled in the art.

[0066] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, serve to explain the principles of the present application.

[0067] These and other characteristics of the present application will become apparent from the following description of a preferred form of embodiment given as a non-limiting example with reference to the accompanying drawings.

[0068] It should also be understood that although the present application has been described with reference to certain specific examples, those skilled in the art will readily be able to implement many other equivalent forms of the present application.

[0069] The above and other aspects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings.

[0070] Specific embodiments of the present application will be described hereinafter with reference to the accompanying drawings; however, it should be understood that the embodiments described are merely examples of the present application and may be implemented in a variety of ways. Familiar and / or repetitive functions and structures are not described in detail to avoid obscuring the present application with unnecessary or redundant details. Therefore, the specific structural and functional details described herein are not intended to be limiting, but rather serve merely as a basis and representative basis for the claims to teach those skilled in the art to variously utilize the present application with substantially any suitable detailed structure.

[0071] This specification may use the phrases "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which may all refer to one or more of the same or different embodiments according to the present application.

[0072] The present application provides a method for detecting the quality of thick steel plates based on data analysis and optimization. Figure 1 Shown, including:

[0073] Step S101: acquiring real-time production process data of the thick and heavy steel plate to be inspected and design standard data for producing the thick and heavy steel plate to be inspected;

[0074] During the specific implementation process, the real-time production process data includes the full-process production and manufacturing data of wide and thick steel, including: process variable data in steelmaking, rolling, heat treatment and other processes; the design standard data includes steel brand, internal steel type, steel plate thickness, maximum and minimum values ​​of the weight percentage of alloy elements in continuous casting billets and maximum and minimum values ​​of carbon equivalent, minimum thickness of the first controlled rolling point, target temperature of the first controlled rolling point, target value of the start cooling temperature, target value of the final cooling temperature, minimum impact energy, minimum yield strength, minimum tensile strength, maximum tensile strength, minimum elongation and other data.

[0075] Step S102: using preset point prediction models corresponding to different mechanical performance indicators to predict the real-time production process data, to obtain point prediction values ​​of the mechanical performance indicators corresponding to the real-time production process data;

[0076] In a specific implementation, the real-time production process data is normalized to obtain normalized production process characteristic data. Based on the production process characteristic data, the preset point prediction models corresponding to different mechanical performance indicators are used to perform predictions to obtain point prediction values ​​for each mechanical performance indicator corresponding to the real-time production process data. The mechanical performance indicators include yield strength, tensile strength, elongation, and average impact energy.

[0077] Step S103: performing conformal prediction on the real-time production process data using a preset mechanical property interval prediction model based on each of the point prediction values ​​to obtain a prediction interval for each of the mechanical property indicators corresponding to the real-time production process data;

[0078] During the specific implementation process, the preset mechanical property interval prediction model at the current moment is obtained; based on the prediction value of each point, the preset mechanical property interval prediction model is used to perform conformal prediction on the real-time production process data to obtain the prediction interval of each mechanical property indicator corresponding to the real-time production process data.

[0079] Step S104: performing quality inspection on the wide and thick steel plate to be inspected based on the prediction interval corresponding to the real-time production process data and the design standard data to obtain a quality inspection result of the wide and thick steel plate to be inspected.

[0080] In a specific implementation process, based on the prediction interval corresponding to the real-time production process data, the upper limit value and the lower limit value of the prediction interval of each different mechanical performance index are extracted; the upper limit value and the lower limit value of the prediction interval of the same mechanical performance index are compared with the design standard data to obtain a comparison result; when the lower limit value of the prediction interval of the average impact energy mechanical performance index is greater than or equal to the minimum value of the standard impact energy, the lower limit value of the prediction interval of the yield strength mechanical performance index is greater than or equal to the minimum value of the standard yield strength, the lower limit value of the prediction interval of the tensile strength performance index is greater than or equal to the minimum value of the standard tensile strength and the upper limit value of the prediction interval of the tensile strength performance index is less than or equal to the maximum value of the standard tensile strength, and the lower limit value of the prediction interval of the elongation mechanical performance index is greater than or equal to the minimum value of the standard elongation, it is determined that the quality of the wide and thick steel plate to be tested is qualified. Otherwise, it is determined that the quality of the wide and thick steel plate to be tested is qualified.

[0081] This application integrates the design standards of wide and thick plate products with the production process variables in the same database, and builds an online real-time quality judgment system to provide confidence intervals for the mechanical properties of wide and thick plate products under a given confidence level, thereby realizing a dynamic comparison between the design standards of wide and thick plate products and the predicted performance, accelerating the efficiency of product quality judgment and reducing the possibility of misjudgment of product quality.

[0082] Another embodiment of the present application provides another method for detecting the quality of thick steel plates based on data analysis and optimization, such as Figure 2 Shown, including:

[0083] Step S201: Constructing a full-process manufacturing database and a process design database for thick and heavy plates;

[0084] In the specific implementation process of this step, the construction of the full-process production and manufacturing database of thick and heavy plates includes the following steps:

[0085] Step 1. Determine the acquisition variable fields and data types according to the full production process of wide and thick plates, including the full production process data collected in real time and the experimental detection data of the representative steel plate production process obtained through experimental detection. Including process variable data in steelmaking, rolling, heat treatment and other processes. The process variables in the steelmaking process include but are not limited to continuous casting billet number, furnace number, continuous casting billet width, continuous casting billet thickness, continuous casting billet length, weight percentage (%) of alloy elements in the continuous casting billet and carbon equivalent (%), wherein the alloy elements include but are not limited to carbon, silicon, manganese, phosphorus, sulfur, chromium, nickel, copper, molybdenum, vanadium, aluminum, titanium and niobium; the process variables in the rolling process include but are not limited to steel grade, internal steel type, steel plate number, steel plate width, steel plate length, steel plate thickness, continuous casting billet temperature before charging, slab first heating section temperature, slab second heating section temperature, slab uniformity temperature, slab tapping temperature, controlled rolling code, starting thickness, rough rolling rolling start temperature, rough rolling and finishing temperature, finishing rolling start temperature, finishing rolling and finishing temperature, cooling start temperature, finishing cooling temperature, temperature before hot straightening, temperature after hot straightening, etc.; process variables in the heat treatment process include but are not limited to heat treatment furnace number, heat treatment material number, heat treatment furnace holding temperature, heat treatment furnace holding time, plate temperature, water quenching temperature, quenching speed, tempering temperature, etc.; performance variables included in the testing process include but are not limited to tensile test batch number, elongation, sample width, sample thickness, tensile strength, yield strength, impact test batch number, impact test temperature, average impact energy, impact energy single value 1, impact energy single value 2, impact energy single value 3;

[0086] Step 2: Determine the process design standard fields and data types according to the full production process of wide and thick plates, including but not limited to steel grade, internal steel grade, steel plate thickness, maximum and minimum values ​​of alloy element weight percentage and carbon equivalent in continuous casting billet, minimum thickness of the first controlled rolling point, target temperature of the first controlled rolling point, target value of the initial cooling temperature, target value of the final cooling temperature, minimum impact energy, minimum yield strength, minimum tensile strength, maximum tensile strength, minimum elongation, etc.;

[0087] Step 3: For each process variable field and data type determined in Step 1, design data tables for each process, including slab composition table, slab rolling table, steel plate heat treatment table, steel plate tensile property table, and steel plate impact property table; establish an entity-relationship model based on the production relationship, and define the main entities in each process table, including furnace number, steel plate number, heat treatment material number, tensile test batch number, impact test batch number, etc.;

[0088] Step 4: Design the process standard table structure based on the process standard fields and data types determined in step 2;

[0089] Step 5. Establish a connection with the Manufacturing Execution System (MES) database, collect corresponding field data from the MES database through a direct database connection, and implement automated data collection and update through a data synchronization mechanism. Specifically, the initial collection adopts a full collection method, collecting all historical data in batches according to timestamps and performing data integrity verification through random sampling. After the historical data collection is completed, starting from the current time τ, configure the scheduling tool Crontab to perform regular incremental collection, implement an online synchronization update mechanism between the local database and the MES system database, and perform data integrity verification through random sampling.

[0090] Step 6: Achieve data alignment and integration. Based on the coupling and correlation between production processes, determine the logical relationship between the subjects in the data table. Design the entire production process in the order of slab information - composition - rolling process - heat treatment process - testing information. Integrate each process data table designed in step 1.3 into the steel plate product level.

[0091] Step 7: Clean and preprocess the data. During the collection process, clean and correct data quality issues (such as missing values, outliers, and noisy data). At the same time, assign a unified timestamp to all records to ensure the temporal consistency and traceability of the data.

[0092] Step S202: extracting the real-time production process data of the thick and heavy steel plate to be tested from the manufacturing database and extracting the design standard data for producing the thick and heavy steel plate to be tested from the process design database;

[0093] During the specific implementation of this step, the real-time production process data includes the full-process production and manufacturing data of wide and thick steel, including: process variable data in steelmaking, rolling, heat treatment and other processes; the design standard data includes steel brand, internal steel type, steel plate thickness, maximum and minimum values ​​of the weight percentage of alloy elements in the continuous casting billet and the maximum and minimum values ​​of the carbon equivalent, minimum thickness of the first controlled rolling point, target temperature of the first controlled rolling point, target value of the start cooling temperature, target value of the final cooling temperature, minimum impact energy, minimum yield strength, minimum tensile strength, maximum tensile strength, minimum elongation and other data.

[0094] Step S203: constructing a preset point prediction model corresponding to different mechanical performance indicators;

[0095] In the specific implementation process of this step, the historical experimental test data H = {(X, Y)} of the thick steel plate to be tested is obtained; the feature data is extracted from the historical experimental test data to obtain the feature data N represents the number of historical data samples, x i ∈R d represents the characteristic data composed of all process variables in step S101, d represents the number of characteristics; y i ∈R m represents the label data composed of all mechanical performance indicators, and m=4 represents the number of labels; different mechanical performance indicators include yield strength, tensile strength, elongation and average impact energy; the temperature characteristic data in the characteristic data are converted using the Arrhenius rate equation to obtain the target characteristic data set; the mathematical expression of the Arrhenius rate equation can be expressed as follows:

[0096]

[0097] Among them, k is the process rate, A is a constant, Q is the activation energy, R is the gas constant, and T is the absolute temperature. The Arrhenius rate equation is an equation proposed by Swedish chemist Svante Arrhenius in 1889 to describe the relationship between chemical reaction rate and temperature. The above formula shows that the temperature and process rate are logarithmic, and the temperature usually appears in the empirical formula of the mechanism in logarithmic form. At the same time, in the high temperature zone, the effect of temperature on the performance index tends to be flat. After taking the logarithm, the data of high and low temperature zones can be processed more evenly to avoid weight imbalance. Therefore, x is selected i The characteristic variables related to temperature are logarithmically transformed. To avoid confusion, the transformed data are still represented by x iRepresentation; the target feature dataset is annotated to obtain a label dataset; the label dataset is used as a training sample to construct a model using the Lasso method (Least Absolute Shrinkage and Selection Operator) to obtain a preset point prediction model corresponding to different mechanical performance indicators; wherein each mechanical performance indicator includes a yield strength indicator, a tensile strength indicator, an elongation indicator, and an average impact energy indicator. Specifically, the label dataset is divided into sample sets according to a preset ratio to obtain a training set H train ={(X train ,Y train )}, calibration set H cal ={(X cal ,Y cal )} and the test set H test ={(X test ,Y test )}; the historical data set can be divided into a training set, a calibration set, and a test set in a ratio of 5:4:1; the first label data in the training set is normalized to obtain a normalized first feature data set; the mathematical formula for performing MinMax normalization on the obtained training set data can be expressed as the following formula (2):

[0098]

[0099] in, Represents the normalized data, n train is the number of training set samples; X min ∈R d represents the minimum value vector of the data, where each element represents the minimum value of the jth feature; X max ∈R d Represents the maximum value vector of the data, where each element Represents the maximum value of the jth feature; based on the first feature data set X' train and the labeled dataset y train ={y i , i=1,…,m} respectively from different mechanical performance index dimensions, using the Lasso method to train the model, and obtain the preset point prediction model corresponding to each mechanical performance index. The objective function of the model training using the Lasso method can be expressed as the following formula (3):

[0100]

[0101] Among them, f i represents the point prediction model of the i-th label, Y trainiRepresents the label data of the i-th label, i = 1, ..., m corresponds to the m mechanical performance indicators of the wide and thick plate products; β i ∈R d is the regression coefficient of the ith model, λ i is a regularization coefficient used to control the sparsity of the regression coefficients. The Lasso point prediction model uses the L1 norm as regularization to achieve automated feature selection, increase model interpretability, and narrow the scope for subsequent root cause analysis of product quality anomalies.

[0102] Step S204: using preset point prediction models corresponding to different mechanical performance indicators to predict the real-time production process data, to obtain point prediction values ​​of the mechanical performance indicators corresponding to the real-time production process data;

[0103] During the specific implementation of this step, the real-time production process data is normalized to obtain normalized production process characteristic data; based on the production process characteristic data, the preset point prediction model corresponding to different mechanical performance indicators is used to perform prediction to obtain the point prediction value of each mechanical performance indicator corresponding to the real-time production process data.

[0104] During the continuous testing of the thick and wide steel plates to be tested, experimental test data for the thick and wide steel plates to be tested for a preset period of time before the current moment is obtained; based on the experimental test data as training samples, the model is retrained using the Lasso method to obtain current-point prediction models corresponding to different mechanical property indicators; and the preset-point prediction model is updated based on the current-point prediction model. Regularly updating the model can improve prediction accuracy, adapt to data changes, and reduce cumulative errors.

[0105] Step S205: constructing a preset mechanical property interval prediction model;

[0106] In the specific implementation process of this step, the second label data in the calibration set is normalized to obtain the normalized second feature data; the mathematical expression for normalizing the second label data in the calibration set can be expressed by the following formula (4):

[0107]

[0108] in, Represents the normalized validation set data; n cal Represents the number of samples in the calibration set. The second feature data is predicted using the preset point prediction model to obtain the first point prediction value of different mechanical performance indicators corresponding to the second label data; the mathematical expression can be expressed as follows:

[0109]

[0110] in, Represents the vector of the first point prediction value of the i-th label in the calibration set;

[0111] Based on the predicted values ​​of each first point and the true values ​​annotated by the second label data, a calculation process is performed to obtain a first inconsistency score of different mechanical performance indicators corresponding to the second label data; the mathematical expression of the first inconsistency score can be expressed as the following formula (6):

[0112]

[0113] Among them, a i represents the inconsistency score vector of the Lasso point prediction model in the i-th label of the calibration set; based on each of the first inconsistency scores and the predetermined error coverage level, a calculation is performed to obtain the first empirical quantiles of different mechanical performance indicators corresponding to the second label data; the mathematical expression of the first empirical quantile can be expressed as follows:

[0114]

[0115] Among them, s i represents the first empirical quantile of the first inconsistency score vector in the i-th label of the calibration set; a model is constructed based on each of the first empirical quantiles and each of the first point prediction values ​​to obtain the preset mechanical property interval prediction model. The mathematical expression of the interval prediction model can be expressed as follows:

[0116]

[0117] in, Represents the prediction interval of the interval prediction model at the i-th label at the test point.

[0118] After constructing the preset mechanical property interval prediction model, the preset mechanical property interval prediction model is tested using the test set data; specifically, the third label data in the test set is normalized to obtain the normalized third feature data; the preset point prediction model is used to predict the third feature data to obtain the second point prediction value of the different mechanical property indicators corresponding to the third label data; based on the second point prediction value, the updated preset mechanical property interval prediction model is used to perform conformal prediction to obtain the prediction interval of each mechanical property indicator corresponding to the third feature data; based on the second point prediction value corresponding to each mechanical property indicator, the true value corresponding to each mechanical property indicator and the prediction interval corresponding to each mechanical property indicator, calculation processing is performed to obtain the interval coverage index and the interval average width index; based on the interval coverage index and the interval average width index, the preset mechanical property interval prediction model is tested and adjusted to obtain the preset mechanical property interval prediction model that meets the preset accuracy requirements.

[0119] Step S206: performing conformal prediction on the real-time production process data using a preset mechanical property interval prediction model based on each of the point prediction values ​​to obtain a prediction interval for each of the mechanical property indicators corresponding to the real-time production process data;

[0120] During the specific implementation of this step, a preset mechanical property interval prediction model is used to perform conformal prediction on the real-time production process data based on the predicted values ​​of each point, so as to obtain the prediction interval of each mechanical property index corresponding to the real-time production process data.

[0121] In the process of continuously testing the wide and thick steel plates to be tested, calculations are performed based on the experimental test data and the point prediction values ​​corresponding to the experimental test data obtained by using the preset point prediction model to obtain second inconsistency scores corresponding to different mechanical performance indicators; the mathematical expression of the second inconsistency score can be expressed as the following formula (9):

[0122]

[0123] in, Represents the second inconsistency score of the i-th label of the current data point, that is, the absolute error of the i-th label; based on the second inconsistency score, a preset objective function is used to calculate and process to obtain the second empirical quantile at the current moment The mathematical expression of the preset objective function can be expressed as follows:

[0124]

[0125] The preset objective function can be expressed as the following formula (11) through the pinball loss function:

[0126]

[0127] The loss function is a convex function, and its differential equation is as shown in the following formula (12):

[0128]

[0129] The second empirical quantile is calculated using the sub-gradient descent optimization method to obtain the third empirical quantile at the next moment. The mathematical expression can be expressed as follows:

[0130]

[0131] in, Indicates the radius length of the prediction interval of the i-th label at the next moment; η i represents the learning rate corresponding to the i-th label, which is used to control the step size during each iterative parameter update. The preset mechanical property interval prediction model is updated based on the third empirical quantile. This adaptively updates the prediction interval radius to achieve online calibration of the conformal prediction model.

[0132] Step S207: performing quality inspection on the wide and thick steel plate to be inspected based on the prediction interval corresponding to the real-time production process data and the design standard data to obtain a quality inspection result of the wide and thick steel plate to be inspected.

[0133] During the specific implementation of this step, based on the prediction interval corresponding to the real-time production process data, the upper limit value and the lower limit value of the prediction interval of each different mechanical performance index are extracted;

[0134] The upper limit value and the lower limit value of the prediction interval of the same mechanical performance index are compared with the design standard data to obtain a comparison result; when the lower limit value of the prediction interval of the average impact energy mechanical performance index is greater than or equal to the minimum value of the standard impact energy, the lower limit value of the prediction interval of the yield strength mechanical performance index is greater than or equal to the minimum value of the standard yield strength, the lower limit value of the prediction interval of the tensile strength performance index is greater than or equal to the minimum value of the standard tensile strength and the upper limit value of the prediction interval of the tensile strength performance index is less than or equal to the maximum value of the standard impact tensile strength, and the lower limit value of the prediction interval of the elongation mechanical performance index is greater than or equal to the minimum value of the standard elongation, it is determined that the quality of the wide and thick steel plate to be tested is qualified. For the wide and thick plate products whose quality is judged to be unqualified, according to the characteristic attributes automatically selected by the Lasso point prediction model in step S204, the process data of the data point is read and compared with the key process node standards to achieve cause analysis. According to the three specifications of the current wide and thick plate product, namely the steel grade, internal steel type, and steel plate thickness, the quality design standards for products of the same specifications are retrieved in the wide and thick plate process design library, represented by IMPACT_ENERGY_MIN, YS_MIN, TS_MIN, TS_MAX, and EL_MIN, which correspond to the minimum impact energy, minimum yield strength, minimum tensile strength, maximum tensile strength, and minimum elongation, respectively; the predicted upper and lower bounds of each performance indicator obtained in step S206 are compared with the product quality design standards, and the product quality grade is automatically determined based on the comparison results, and the quality of each wide and thick plate is automatically determined according to the preset grading rules.

[0135] This application pre-constructs a full-process production and manufacturing database and a process design database for wide and thick plates, and ensures the temporal consistency and traceability of data by performing data processing methods such as alignment and integration, cleaning and pre-processing on the data; by constructing preset point prediction models corresponding to different mechanical performance indicators, it can predict the real-time production process data collected in real time, and obtain point prediction values ​​corresponding to mechanical performance parameters such as yield strength, elongation, tensile strength, and average impact energy corresponding to the real-time production process data, and conformally predict each of the point prediction values ​​through the constructed preset mechanical performance interval prediction model, and obtain the prediction interval of each mechanical performance indicator corresponding to the real-time production process data; it effectively avoids the disadvantage of traditional point prediction models that cannot quantify the uncertainty of the model itself, and can adjust the confidence level according to demand to adapt to the quality inspection requirements of different wide and thick plate products. This application models the online calibration problem of conformal prediction as a convex optimization problem, and dynamically adapts to changes in data distribution through online optimization methods. After verification by actual production data, it can effectively improve the accuracy and stability of model predictions.

[0136] The following is an explanation of the wide and thick steel plate quality inspection method of this application in combination with specific application scenarios:

[0137] A method for detecting thick steel plates based on data analysis and optimization according to an embodiment of the present application is applied to a thick steel plate manufacturing plant of a certain steel plant to automatically determine product quality. For the four mechanical performance indicators of yield strength, tensile strength, elongation, and average impact energy, data from January 2023 to May 2024 are selected as historical data to train the Lasso point prediction model and the interval prediction model. Starting from June 1, 2024, the online quality determination and interval prediction model are adaptively updated. As of August 31, 2024, the RMSE and R 2 To evaluate the Lasso point prediction model, the interval coverage and average interval width were selected to evaluate the conformal interval prediction model. The model performance was calculated using representative plate data as shown in Table 1:

[0138] Table 1: Results of Lasso point prediction model and conformal prediction model

[0139] tensile strength Yield strength Elongation Average impact energy RMSE 14.37 19.77 1.35 28.33 <![CDATA[R 2 ]]> 0.905 0.837 0.775 0.61 Interval coverage 92.16％ 92.01％ 90.32％ 91.79％ Interval width 46.01 66.19 4.58 98.26

[0140] As can be seen from Table 1, the interval coverage of the conformal prediction model always remains above the set value; at the same time, taking the tensile strength performance index as an example, the point prediction and interval prediction results are as follows Figure 3 As shown in the figure, it can be seen that the point prediction curve is very close to the true value curve, and the true value curve falls into the prediction interval with a high probability, which proves that the model established in the example of the present invention has a high prediction accuracy. Figure 4 As shown, it can be seen that product quality judgment based on the prediction interval is more robust, which proves the effectiveness of the method proposed in this invention.

[0141] Another embodiment of the present application provides a device for detecting the quality of thick steel plates based on data analysis and optimization, such as Figure 5 Shown, including:

[0142] Acquisition module 1, used to acquire real-time production process data of the thick and heavy steel plate to be tested and design standard data for producing the thick and heavy steel plate to be tested;

[0143] Point prediction module 2, used to predict the real-time production process data using preset point prediction models corresponding to different mechanical performance indicators, and obtain point prediction values ​​of each mechanical performance indicator corresponding to the real-time production process data;

[0144] A conformal prediction module 3 is configured to perform conformal prediction on the real-time production process data using a preset mechanical property interval prediction model based on each of the point prediction values, and obtain a prediction interval for each of the mechanical property indicators corresponding to the real-time production process data;

[0145] The quality inspection module 4 is used to perform quality inspection on the wide and thick steel plate to be inspected based on the prediction interval corresponding to the real-time production process data and the design standard data, and obtain the quality inspection result of the wide and thick steel plate to be inspected.

[0146] During the specific implementation process, the device also includes a preset point prediction model construction module, which is specifically used to obtain historical experimental detection data of the wide and thick steel plate to be tested; perform feature extraction on the historical experimental detection data to obtain feature data; use the Arrhenius rate equation to convert the temperature feature data in the feature data to obtain a target feature data set; label the target feature data set to obtain a label data set; use the label data set as a training sample to construct a model using the Lasso method to obtain preset point prediction models corresponding to different mechanical performance indicators; wherein each of the mechanical performance indicators includes a yield strength indicator, a tensile strength indicator, an elongation indicator and an average impact energy indicator.

[0147] During the specific implementation process, the preset point prediction model construction module is also used to divide the label data set into sample sets according to a preset ratio to obtain a training set, a calibration set and a test set; normalize the first label data in the training set to obtain a normalized first feature data set; based on the first feature data set and the label data set, model training is performed using the Lasso method from different mechanical performance indicator dimensions to obtain the preset point prediction model corresponding to each of the mechanical performance indicators.

[0148] During the specific implementation process, the device also includes a first update module, which is specifically used to obtain experimental test data of the wide and thick steel plate to be tested for a preset time length before the current moment; based on the experimental test data as a training sample, the Lasso method is used to re-train the model to obtain the current point prediction model corresponding to different mechanical performance indicators; based on the current point prediction model, the preset point prediction model is updated.

[0149] During the specific implementation process, the device also includes a preset mechanical property interval prediction model construction module, which is specifically used to normalize the second label data in the calibration set to obtain normalized second feature data; use the preset point prediction model to predict the second feature data to obtain the first point prediction value of the different mechanical property indicators corresponding to the second label data; perform calculation processing based on each first point prediction value and the true value marked by the second label data to obtain the first inconsistency score of the different mechanical property indicators corresponding to the second label data; perform calculation processing based on each first inconsistency score and a predetermined error coverage level to obtain the first empirical quantile of the different mechanical property indicators corresponding to the second label data; perform model construction based on each first empirical quantile and each first point prediction value to obtain the preset mechanical property interval prediction model.

[0150] During the specific implementation process, the device also includes a second update module, which is specifically used to perform calculations based on the experimental test data and the point prediction value corresponding to the experimental test data obtained by using the preset point prediction model to obtain a second inconsistency score corresponding to different mechanical performance indicators; based on the second inconsistency score, a preset objective function is used to perform calculations to obtain a second empirical quantile at the current moment; based on the second empirical quantile, a subgradient descent optimization method is used to perform calculations to obtain a third empirical quantile at the next moment; and based on the third empirical quantile, the preset mechanical performance interval prediction model is updated.

[0151] During the specific implementation process, the device also includes a model testing module, which is specifically used to normalize the third label data in the test set to obtain normalized third feature data; use the preset point prediction model to predict the third feature data to obtain a second point prediction value of different mechanical performance indicators corresponding to the third label data; based on the second point prediction value, use the updated preset mechanical performance interval prediction model to perform conformal prediction to obtain the prediction interval of each mechanical performance indicator corresponding to the third feature data; based on the second point prediction value corresponding to each mechanical performance indicator, the true value corresponding to each mechanical performance indicator and the prediction interval corresponding to each mechanical performance indicator, calculate and process to obtain an interval coverage index and an interval average width index; test and adjust the preset mechanical performance interval prediction model based on the interval coverage index and the interval average width index to obtain the preset mechanical performance interval prediction model that meets the preset accuracy requirements.

[0152] During the specific implementation process, the point prediction module 2 is specifically used to: normalize the real-time production process data to obtain normalized production process characteristic data; based on the production process characteristic data, use the preset point prediction model corresponding to different mechanical performance indicators to perform predictions to obtain point prediction values ​​of each mechanical performance indicator corresponding to the real-time production process data.

[0153] In a specific implementation process, the quality detection module 4 is specifically used to extract the upper limit value and the lower limit value of the prediction interval of each different mechanical performance index based on the prediction interval corresponding to the real-time production process data;

[0154] The upper limit value and the lower limit value of the prediction interval of the same mechanical property index are compared with the design standard data to obtain a comparison result; when the lower limit value of the prediction interval of the average impact energy mechanical property index is greater than or equal to the standard minimum impact energy, the lower limit value of the prediction interval of the yield strength mechanical property index is greater than or equal to the standard minimum yield strength, the lower limit value of the prediction interval of the tensile strength performance index is greater than or equal to the standard minimum tensile strength and the upper limit value of the prediction interval of the tensile strength performance index is less than or equal to the standard maximum tensile strength, and the lower limit value of the prediction interval of the elongation mechanical property index is greater than or equal to the standard minimum elongation, it is determined that the quality of the wide and thick steel plate to be tested is qualified.

[0155] This application integrates the design standards of wide and thick plate products with the production process variables in the same database, and builds an online real-time quality judgment system to provide confidence intervals for the mechanical properties of wide and thick plate products under a given confidence level, thereby realizing a dynamic comparison between the design standards of wide and thick plate products and the predicted performance, accelerating the efficiency of product quality judgment and reducing the possibility of misjudgment of product quality.

[0156] The above embodiments are merely exemplary embodiments of the present application and are not intended to limit the scope of the present application. The scope of protection of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and scope of protection of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the scope of protection of the present application.

Claims

1. A method for detecting the quality of wide and thick steel plates based on data analysis and optimization, characterized in that: include: Acquire real-time production process data of the thick and heavy steel plate to be tested and design standard data for producing the thick and heavy steel plate to be tested; Predicting the real-time production process data using preset point prediction models corresponding to different mechanical performance indicators to obtain point prediction values ​​of each mechanical performance indicator corresponding to the real-time production process data; Based on the point prediction values, a preset mechanical property interval prediction model is used to perform conformal prediction on the real-time production process data to obtain a prediction interval of each mechanical property indicator corresponding to the real-time production process data; The quality inspection of the wide and thick steel plate to be inspected is performed based on the prediction interval corresponding to the real-time production process data and the design standard data to obtain a quality inspection result of the wide and thick steel plate to be inspected.

2. The method according to claim 1, wherein Before predicting the real-time production process data using the preset point prediction models corresponding to different mechanical performance indicators, the method further includes: constructing the preset point prediction models; The construction of the preset point prediction model specifically includes: Acquiring historical experimental test data of the wide and thick steel plate to be tested; Performing feature extraction on the historical experimental detection data to obtain feature data; Arrhenius rate equation is used to convert temperature characteristic data in the characteristic data to obtain a target characteristic data set; Annotating the target feature dataset to obtain a label dataset; The labeled data set is used as a training sample to construct a model using the Lasso method to obtain a preset point prediction model corresponding to different mechanical performance indicators; The mechanical performance indicators include yield strength, tensile strength, elongation and average impact energy.

3. The method according to claim 2, wherein The label data set is used as a training sample to construct a model using the Lasso method to obtain a preset point prediction model corresponding to different mechanical performance indicators, specifically including: Divide the labeled data set into sample sets according to a preset ratio to obtain a training set, a calibration set, and a test set; Normalizing the first label data in the training set to obtain a normalized first feature data set; Based on the first feature data set and the label data set, model training is performed respectively from different mechanical performance index dimensions using the Lasso method to obtain the preset point prediction model corresponding to each of the mechanical performance indicators.

4. The method according to claim 3, wherein The method further comprises: Acquire experimental test data of the thick and wide steel plate to be tested for a preset time period before the current moment; Based on the experimental test data as training samples, the model is retrained using the Lasso method to obtain current point prediction models corresponding to different mechanical performance indicators; The preset point prediction model is updated based on the current point prediction model.

5. The method according to claim 3, wherein Before conformal prediction of the real-time production process data is performed using a preset mechanical property interval prediction model based on each of the point prediction values, the method further comprises: constructing a preset mechanical property interval prediction model; The construction of the preset mechanical property interval prediction model specifically includes: Normalizing the second label data in the calibration set to obtain normalized second feature data; Using the preset point prediction model to predict the second feature data, to obtain first point prediction values ​​of different mechanical performance indicators corresponding to the second label data; Performing calculations based on the predicted values ​​of each of the first points and the true values ​​annotated by the second label data to obtain first inconsistency scores for different mechanical performance indicators corresponding to the second label data; Performing calculations based on each of the first inconsistency scores and a predetermined error coverage level to obtain first empirical quantiles of different mechanical performance indicators corresponding to the second label data; A model is constructed based on each of the first empirical quantiles and each of the first point prediction values ​​to obtain the preset mechanical property interval prediction model.

6. The method according to claim 4, wherein The method further comprises: Performing calculations based on the experimental test data and point prediction values ​​corresponding to the experimental test data obtained by using the preset point prediction model to obtain second inconsistency scores corresponding to different mechanical performance indicators; Performing calculation based on the second inconsistency score using a preset objective function to obtain a second empirical quantile at the current moment; Calculating the second empirical quantile using a sub-gradient descent optimization method to obtain a third empirical quantile at the next moment; The preset mechanical property interval prediction model is updated based on the third empirical quantile.

7. The method according to claim 6, wherein The method further comprises: Normalizing the third label data in the test set to obtain normalized third feature data; Using the preset point prediction model to predict the third characteristic data, to obtain a second point prediction value of a different mechanical performance index corresponding to the third label data; Performing conformal prediction using the updated preset mechanical property interval prediction model based on the second-point prediction value to obtain a prediction interval for each of the mechanical property indicators corresponding to the third characteristic data; Performing calculations based on the second-point predicted value corresponding to each of the mechanical performance indicators, the true value corresponding to each of the mechanical performance indicators, and the prediction interval corresponding to each of the mechanical performance indicators to obtain an interval coverage index and an interval average width index; The preset mechanical property interval prediction model is tested and adjusted based on the interval coverage index and the interval average width index to obtain the preset mechanical property interval prediction model that meets the preset accuracy requirements.

8. The method according to claim 1, wherein The method of using the preset point prediction models corresponding to different mechanical performance indicators to predict the real-time production process data to obtain the point prediction values ​​of the mechanical performance indicators corresponding to the real-time production process data specifically includes: Normalizing the real-time production process data to obtain normalized production process characteristic data; Based on the production process characteristic data, the preset point prediction model corresponding to different mechanical performance indicators is used to perform predictions to obtain point prediction values ​​of each mechanical performance indicator corresponding to the real-time production process data.

9. The method according to claim 1, wherein The method further comprises: performing a quality inspection on the wide and thick steel plate to be inspected based on the prediction interval corresponding to the real-time production process data and the design standard data to obtain a quality inspection result of the wide and thick steel plate to be inspected. Extracting the upper bound value and the lower bound value of the prediction interval of each different mechanical performance index based on the prediction interval corresponding to the real-time production process data; Comparing the upper limit value and the lower limit value of the prediction interval of the same mechanical property index with the design standard data to obtain a comparison result; When the lower limit value of the prediction interval of the average impact energy mechanical performance index is greater than or equal to the minimum value of the standard impact energy, the lower limit value of the prediction interval of the yield strength mechanical performance index is greater than or equal to the minimum value of the standard yield strength, the lower limit value of the prediction interval of the tensile strength performance index is greater than or equal to the minimum value of the standard tensile strength and the upper limit value of the prediction interval of the tensile strength performance index is less than or equal to the maximum value of the standard impact tensile strength, and the lower limit value of the prediction interval of the elongation mechanical performance index is greater than or equal to the minimum value of the standard elongation, it is determined that the quality of the wide and thick steel plate to be tested is qualified.

10. A device for detecting the quality of wide and thick steel plates based on data analysis and optimization, characterized in that: include: An acquisition module is used to acquire real-time production process data of the wide and thick steel plate to be tested and design standard data for producing the wide and thick plate to be tested; a point prediction module, configured to predict the real-time production process data using preset point prediction models corresponding to different mechanical performance indicators, and obtain a point prediction value of each mechanical performance indicator corresponding to the real-time production process data; a conformal prediction module, configured to perform conformal prediction on the real-time production process data using a preset mechanical property interval prediction model based on each of the point prediction values, and obtain a prediction interval for each of the mechanical property indicators corresponding to the real-time production process data; The quality inspection module is used to perform quality inspection on the wide and thick steel plates to be inspected based on the prediction interval corresponding to the real-time production process data and the design standard data, and obtain quality inspection results of the wide and thick steel plates to be inspected.

Images