Method and system for seamless steel tube quality evaluation based on mechanism model and machine learning

Abstract

The present application belongs to the technical field of straightening quality evaluation method, and relates to a seamless steel pipe quality evaluation method and system based on mechanism model and machine learning, which comprises the following steps: obtaining data samples and corresponding real outputs of a plurality of historical production tasks; constructing and training a mechanism-machine learning hybrid model; using the hybrid prediction model to perform real-time quality prediction and obtaining a prediction result; performing production guidance based on the prediction result to determine whether the straightening task precision requirement is met; evaluating the stability of the straightening process through evaluation indexes to determine whether the quality fluctuation is within the allowable range, and further determining whether the seamless steel pipe product meets the straightening task precision requirement. The beneficial effect is that the straightening effect of the final product is predicted by obtaining the straightening machine process parameters and the steel pipe parameters, and the coupling calculation method for the straightening process parameters and the steel pipe parameters is adopted to predict whether the product quality meets the standard by analyzing the change trend of the quality indexes in the straightening process.

Description

Technical Field

[0001] This invention relates to the field of straightening quality evaluation methods, and in particular to a seamless steel pipe quality evaluation method and system based on mechanism models and machine learning. Background Technology

[0002] Seamless steel pipes are core components in fields such as petrochemicals, nuclear power, and aerospace. Their straightness, dimensional accuracy, and surface quality directly affect the safety and reliability of products used in these fields. At the same time, with the upgrading of high-end manufacturing, industry standards have become more stringent in their requirements for steel pipe quality, necessitating further upgrades to quality prediction and evaluation technologies.

[0003] Traditional seamless steel pipe straightening mostly relies on manual experience to adjust parameters. Under this method, the straightening quality depends heavily on the operator's experience and judgment. Differences in operation between different personnel may lead to large fluctuations in straightness, resulting in rework. At the same time, more time is required for debugging when changing straightening specifications, which affects processing efficiency. Summary of the Invention

[0004] Technical problems to be solved

[0005] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides a seamless steel pipe quality evaluation method and system based on mechanism model and machine learning, which solves the technical problems of traditional seamless steel pipe straightening relying on manual experience to adjust parameters, resulting in the straightening quality being greatly affected by human experience, the straightness fluctuation being prone to rework, and the time-consuming specification change and debugging affecting processing efficiency.

[0006] Technical solution

[0007] To achieve the above objectives, the main technical solutions adopted by the present invention include:

[0008] In a first aspect, the present invention provides a method for evaluating the quality of seamless steel pipes based on mechanistic models and machine learning, comprising the following steps:

[0009] Step 1: Obtain data samples and corresponding actual outputs from multiple historical production tasks in the historical production database. ;

[0010] Step 2: Based on the characteristics of the seamless steel pipe straightening process, construct and train a mechanism-machine learning hybrid model;

[0011] Step 3: Use a hybrid prediction model to perform real-time quality prediction and obtain the prediction results;

[0012] Step 4: Provide production guidance based on the prediction results to determine whether the accuracy requirements of the straightening task are met;

[0013] The final quality prediction value Compared with the preset quality pass threshold Compare;

[0014] like If so, it is determined that the current straightening process parameters are set reasonably and production can proceed.

[0015] like If so, an early warning will be issued, and adjustments to the relevant straightening process parameters will be recommended;

[0016] Step 5: Evaluate the stability of the straightening process through quality indicators, determine whether the quality fluctuation is within the allowable range, and then determine whether the finished seamless steel pipe meets the accuracy requirements of the straightening task.

[0017] As a further improvement to the method of the present invention, each data sample includes:

[0018] Input feature set X: includes the material parameters, geometric parameters, straightening equipment parameters, and straightening process parameters of the steel pipe;

[0019] The material parameters include at least the yield strength. and elastic modulus The geometric parameters include at least the outer diameter. , inner diameter Steel pipe wall thickness The equipment parameters include at least the number of straightener rolls, the straightener roll diameter, the straightening speed, and the straightener roll spacing L; the process parameters include at least the theoretical reduction δ calculated based on the target plastic deformation rate η.

[0020] As a further improvement to the method of the present invention, step 2, the construction and training of the hybrid prediction model specifically includes the following methods:

[0021] Step 2.1: Construct a mathematical mechanism model for the straightening of seamless steel pipes;

[0022] Step 2.2: Prediction bias of the computer model ;

[0023] Step 2.3: Train the bias prediction machine learning model to form the bias prediction model;

[0024] Step 2.4: Using the input feature set X of all historical data samples as the training input bias prediction model, obtain the bias correction value. ;

[0025] Step 2.5: Use the deviation prediction model to predict the quality.

[0026] As a further improvement to the method of the present invention, in step 2.1, the mathematical mechanism modeling is based on elastoplastic mechanics and material deformation theory, constructs a basic parameter framework for the straightening process, describes the deformation law of the steel pipe under the action of the straightening roller, and establishes the basis for calculating the deformation amount.

[0027] As a further improvement to the method of the present invention, the core calculation formula of the mechanism model includes:

[0028] Moment of inertia of steel pipe section :

[0029] ;

[0030] In the formula, Moment of inertia of steel pipe section (unit: ); The outer diameter of the steel pipe. The inner diameter of the steel pipe, and , For the steel pipe wall thickness;

[0031] Relationship between the curvature and bending moment of steel pipe:

[0032] ;

[0033] In the formula, Indicates the bending radius (unit: m), reflecting the degree of bending of the steel pipe; Indicates the bending moment during the straightening process (unit: It is positively correlated with the reduction amount and the roller spacing; This represents the elastic modulus of the steel pipe (unit: Pa), an inherent property of the steel grade.

[0034] Used to calculate the theoretical reduction of the straightening machine The formula, The formula for calculating the plastic deformation rate is as follows:

[0035] ;

[0036] In the formula, Indicates the rate of plastic deformation; Indicates the distance between the straightening rollers; This indicates the maximum permissible bending radius, which is determined by the material and specifications of the steel pipe.

[0037] Iterative calculation model for residual deformation after multi-roll straightening: based on the original curvature, the inverse curvature applied by each straightening roll, and the material yield strength. Based on the cross-sectional properties, the elastic curvature and residual curvature are iteratively calculated to obtain the theoretical residual deformation distribution along the entire length of the steel pipe. The ratio of its maximum value to the length of the steel pipe is the theoretical straightness prediction value.

[0038] The total deformation curvature, derived from the superposition of the original curvature and the inverse curvature, reflects the overall degree of deformation during the straightening process. Its formula is:

[0039] ;

[0040] In the formula, The total deformation curvature; The original curvature of the steel pipe; The reverse bending rate applied to the straightening roller;

[0041] The curvature corresponding to the elastic recovery of the steel pipe after unloading is determined by the elastic properties of the material, and its formula is:

[0042] ;

[0043] In the formula, The yield strength of the steel pipe; This is the distance from the fiber with the largest cross-section to the neutral axis;

[0044] The residual curvature of the steel pipe after unloading is given by the following formula:

[0045] ;

[0046] In the formula, Residual curvature directly corresponds to straightness; the smaller the residual curvature, the better the straightness.

[0047] The critical curvature at which the surface fiber stress of the steel pipe reaches the yield strength, i.e., the yield curvature:

[0048] ;

[0049] The yield curvature is used to determine the starting point of plastic deformation; when the total deformation curvature... At this time, the steel pipe enters the plastic deformation stage.

[0050] As a further improvement to the method of the present invention, in step 2.2, the prediction bias of the computer model... ,include:

[0051] For each historical data sample, the input feature set X is input into the mechanism model to obtain the mechanism prediction value. , namely the residual deformation after straightening, is used to simulate and predict the straightness of the seamless steel pipe after straightening;

[0052] Residual deformation The ratio of the maximum value of the straightness to the length of the steel pipe is defined as straightness, thus completing the mapping from deformation to quality indicators. The mapped mechanistic model prediction values Used as a parameter for quality evaluation;

[0053] Deviation between computer-predicted values ​​and actual values :

[0054] .

[0055] As a further improvement to the method of the present invention, step 2.3, training the bias prediction machine learning model to form a bias prediction model, includes:

[0056] Using the input feature set X of all historical data samples as training input, and the corresponding prediction bias... The machine learning model is trained using the target variable as the training variable to obtain a well-trained bias prediction model.

[0057] Secondly, the present invention provides a seamless steel pipe quality evaluation system based on mechanism model and machine learning, including a data acquisition module, a prediction result calculation module, a first judgment module and a second judgment module;

[0058] The data acquisition module is used to acquire data samples and corresponding actual outputs from multiple historical production tasks from the historical production database. ;

[0059] The prediction result calculation module constructs and trains a mechanism-machine learning hybrid model based on the characteristics of the seamless steel pipe straightening process; and uses the hybrid prediction model to perform real-time quality prediction to obtain the prediction result.

[0060] The first judgment module provides production guidance based on the prediction results, determining whether the accuracy requirements of the straightening task are met; and sets the final quality prediction value. Compared with the preset quality pass threshold Compare; if If the current straightening process parameters are set correctly, production can proceed; otherwise... If so, an early warning will be issued, and adjustments to the relevant straightening process parameters will be recommended;

[0061] The second judgment module evaluates the stability of the straightening process through evaluation indicators, determines whether the quality fluctuation is within the allowable range, and then determines whether the finished seamless steel pipe meets the accuracy requirements of the straightening task.

[0062] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed, implements the seamless steel pipe quality evaluation method based on mechanism model and machine learning as described in any of the first aspects above.

[0063] Fourthly, the present invention provides a storage device, including a storage medium and a processor, wherein the storage medium stores a computer program, and when the program is executed by the processor, it implements the seamless steel pipe quality evaluation method based on mechanism model and machine learning as described in any one of the first aspects above.

[0064] Beneficial effects

[0065] The beneficial effects of this invention are:

[0066] A predictive model framework was constructed using a hybrid approach combining mathematical mechanistic models and machine learning models. This framework predicts the straightening effect of the final product by acquiring the straightening machine's process parameters and the steel pipe's parameters. It employs a coupled calculation method based on the straightening process parameters and steel pipe parameters, analyzing the changing trends of quality indicators during the straightening process to predict whether the product quality meets standards. This predictive method can anticipate potential quality deviations in the steel pipe through quality prediction, allowing for timely adjustments to the process before quality issues arise due to abnormal parameters. This reduces the defect rate and rework costs caused by quality problems, minimizes production interruptions due to quality issues, and ensures a smooth production process, effectively improving production efficiency. Attached Figure Description

[0067] Figure 1 A flowchart of a seamless steel pipe quality evaluation method based on mechanistic models and machine learning provided in an embodiment of the present invention;

[0068] Figure 2 This is a module structure diagram of a seamless steel pipe quality evaluation system based on mechanism model and machine learning provided in an embodiment of the present invention;

[0069] Figure 3 The diagram shows the results in an embodiment of the present invention. Detailed Implementation

[0070] To better understand the above technical solutions, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention can be understood more clearly and thoroughly, and that the scope of the present invention can be fully conveyed to those skilled in the art.

[0071] Firstly, such as Figure 1 and Figure 3 As shown, this embodiment of the invention provides a method for evaluating the quality of seamless steel pipes based on mechanistic models and machine learning, including the following steps:

[0072] Step 1: Obtain data samples and corresponding actual outputs from multiple historical production tasks in the historical production database. ;

[0073] Each data sample includes:

[0074] Input feature set X: includes the material parameters, geometric parameters, straightening equipment parameters, and straightening process parameters of the steel pipe;

[0075] The material parameters include at least the yield strength. and elastic modulus The geometric parameters include at least the outer diameter. , inner diameter Steel pipe wall thickness And length; the equipment parameters include at least the number of straightener rolls, the straightener roll diameter, the straightening speed, and the straightener roll spacing L; the process parameters include at least the theoretical reduction δ calculated based on the target plastic deformation rate η;

[0076] Step 2: Based on the characteristics of the seamless steel pipe straightening process, construct and train a mechanism-machine learning hybrid model;

[0077] The construction and training of the hybrid prediction model specifically includes the following methods:

[0078] Step 2.1: Construct a mathematical mechanism model for the straightening of seamless steel pipes;

[0079] The mathematical mechanism modeling is based on elastoplastic mechanics and material deformation theory, constructs a basic parameter framework for the straightening process, describes the deformation law of the steel pipe under the action of the straightening roller, and establishes the basis for calculating the deformation amount.

[0080] Specifically, based on the theory of elastoplastic bending, the steel pipe is assumed to be an ideal elastoplastic material, and the process of a single straightening roller action is regarded as a pure bending deformation, and a mathematical mechanism model is established.

[0081] The mathematical mechanism model calculates and predicts the theoretical residual deformation of the steel pipe after multi-roll straightening, i.e., the mechanism prediction value, based on the input feature set X. ;

[0082] The core calculation formulas of the mechanism model include:

[0083] Moment of inertia of steel pipe section :

[0084] ;

[0085] In the formula, Moment of inertia of steel pipe section (unit: ); The outer diameter of the steel pipe. The inner diameter of the steel pipe, and , For the steel pipe wall thickness;

[0086] Relationship between the curvature and bending moment of steel pipe:

[0087] ;

[0088] In the formula, Indicates the bending radius (unit: m), reflecting the degree of bending of the steel pipe; Indicates the bending moment during the straightening process (unit: It is positively correlated with the reduction amount and the roller spacing; This represents the elastic modulus of steel pipe (unit: Pa), an inherent property of the steel grade (e.g., P110 steel grade). (approximately 206 GPa)

[0089] Used to calculate the theoretical reduction of the straightening machine The formula, The formula is calculated based on the required plastic deformation rate (usually taken as 80%):

[0090] ;

[0091] In the formula, Indicates the rate of plastic deformation; Indicates the distance between the straightening rollers; This indicates the maximum permissible bending radius, which is determined by the material and specifications of the steel pipe.

[0092] Iterative calculation model for residual deformation after multi-roll straightening: based on the original curvature, the inverse curvature applied by each straightening roll, and the material yield strength. Based on the cross-sectional properties, the elastic curvature and residual curvature are iteratively calculated to obtain the theoretical residual deformation distribution along the entire length of the steel pipe. The ratio of its maximum value to the length of the steel pipe is the theoretical straightness prediction value.

[0093] The total deformation curvature, derived from the superposition of the original curvature and the inverse curvature, reflects the overall degree of deformation during the straightening process. Its formula is:

[0094] ;

[0095] In the formula, The total deformation curvature; The original curvature of the steel pipe; The reverse bending rate applied to the straightening roller.

[0096] The curvature corresponding to the elastic recovery of the steel pipe after unloading is determined by the elastic properties of the material, and its formula is:

[0097] ;

[0098] In the formula, The yield strength of the steel pipe; This is the distance from the fiber with the largest cross-section to the neutral axis.

[0099] The residual curvature of the steel pipe after unloading is given by the following formula:

[0100] ;

[0101] In the formula, Residual curvature directly corresponds to straightness; the smaller the residual curvature, the better the straightness.

[0102] The critical curvature at which the surface fiber stress of the steel pipe reaches the yield strength, i.e., the yield curvature:

[0103] ;

[0104] The yield curvature is used to determine the starting point of plastic deformation; when the total deformation curvature... At this time, the steel pipe enters the plastic deformation stage.

[0105] Step 2.2: Prediction bias of the computer model ,include:

[0106] For each historical data sample, the input feature set X is input into the mechanism model to obtain the mechanism prediction value. , namely the residual deformation after straightening, is used to simulate and predict the straightness of the seamless steel pipe after straightening;

[0107] Residual deformation The ratio of the maximum value of the straightness to the length of the steel pipe is defined as straightness, thus completing the mapping from deformation to quality indicators. The mapped mechanistic model prediction values Used as a parameter for quality evaluation.

[0108] Deviation between computer-predicted values ​​and actual values :

[0109] ;

[0110] Step 2.3: Train the bias prediction machine learning model to form the bias prediction model;

[0111] Using the input feature set X of all historical data samples as training input, and the corresponding prediction bias... The machine learning model is trained under supervised learning as the target variable for training, resulting in a well-trained deviation prediction model. This deviation prediction model is used to predict the error of the mechanism model based on the input process parameters.

[0112] Step 2.4: Using the input feature set X of all historical data samples as the training input bias prediction model, obtain the bias correction value. ;

[0113] In this embodiment, a linear regression model is selected as the machine learning model to initially fit the linear relationship between process parameters and quality indicators. The deviation correction value in the deviation prediction model... The calculation formula is:

[0114] ;

[0115] In the formula, For deviation correction values ​​(such as straightness, ellipticity); Indicates input features (such as reduction amount, roller speed, wall thickness, etc.); Represents the intercept term; Indicates the feature weights (obtained through training with historical data); This represents the error term, which the model needs to minimize.

[0116] Step 2.5: The expression for the deviation prediction model used for quality prediction is:

[0117] ;

[0118] Step 3: Use a hybrid prediction model to perform real-time quality prediction and obtain the prediction results;

[0119] Step 4: Provide production guidance based on the prediction results to determine whether the accuracy requirements of the straightening task are met;

[0120] The final quality prediction value Compared with the preset quality pass threshold Compare;

[0121] like If so, it is determined that the current straightening process parameters are set reasonably and production can proceed.

[0122] like If the error occurs, an early warning will be issued, and adjustments to the relevant straightening process parameters will be recommended.

[0123] In this embodiment, the following settings are provided: If the straightness prediction result for this batch of steel pipes is mm / m, then... If so, it is determined that the current straightening process parameters are set reasonably and production can proceed.

[0124] If the straightness prediction result of this batch of steel pipes If the error occurs, an early warning will be issued, and adjustments to the relevant straightening process parameters will be recommended.

[0125] Step 5: Evaluate the stability of the straightening process using the process capability index CPK to determine whether the quality fluctuation is within the allowable range, and thus determine whether the finished seamless steel pipe meets the accuracy requirements of the straightening task;

[0126] If the CPK evaluation meets the requirements, production can proceed; if the CPK evaluation does not meet the requirements, the model needs to be retrained. Furthermore, this indicator is related to quality prediction and serves as an evaluation standard for the stability of batch prediction results. Failure to meet the standard means that the quality fluctuates significantly, and the model needs to be retrained.

[0127] The formula for calculating CPK is:

[0128] ;

[0129] In the formula, This indicates the upper limit of the quality indicator. This indicates the lower limit of the quality indicator. This represents the average value of the quality indicators. Indicates the standard deviation of quality indicators;

[0130] Industrial requirements (Corresponding to a process non-conformity rate ≤ 0.006%), quality fluctuations are analyzed through statistical process analysis and the process capability index (CPK) is calculated. CPK is required to be ≥ 1.33 to ensure process stability.

[0131] The present invention provides a seamless steel pipe quality evaluation method based on mechanism model and machine learning. Through the hybrid modeling of mathematical mechanism model and machine learning model, it realizes the whole process from process parameter input to quality index prediction and evaluation.

[0132] Secondly, such as Figure 2 As shown, this embodiment of the invention provides a seamless steel pipe quality evaluation system based on mechanism model and machine learning, including a data acquisition module, a prediction result calculation module, a first judgment module and a second judgment module;

[0133] The data acquisition module is used to acquire data samples and corresponding actual outputs from multiple historical production tasks from the historical production database. ;

[0134] The prediction result calculation module constructs and trains a mechanism-machine learning hybrid model based on the characteristics of the seamless steel pipe straightening process; and uses the hybrid prediction model to perform real-time quality prediction to obtain the prediction result.

[0135] The first judgment module provides production guidance based on the prediction results, determining whether the accuracy requirements of the straightening task are met; and sets the final quality prediction value. Compared with the preset quality pass threshold Compare; if If the current straightening process parameters are set correctly, production can proceed; otherwise... If so, an early warning will be issued, and adjustments to the relevant straightening process parameters will be recommended;

[0136] The second judgment module evaluates the stability of the straightening process through evaluation indicators, determines whether the quality fluctuation is within the allowable range, and then determines whether the finished seamless steel pipe meets the accuracy requirements of the straightening task.

[0137] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed, implements the seamless steel pipe straightening quality prediction and evaluation method based on machine learning as described in any of the first aspects above.

[0138] Fourthly, embodiments of the present invention provide a storage device, including a storage medium and a processor, wherein the storage medium stores a computer program, and when the program is executed by the processor, it implements the seamless steel pipe straightening quality prediction and evaluation method based on machine learning as described in any one of the first aspects above.

[0139] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0140] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, then this invention should also include these modifications and variations.

[0141] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make modifications, alterations, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for evaluating the quality of seamless steel pipes based on mechanistic models and machine learning, characterized in that, Includes the following steps: Step 1: Obtain data samples and corresponding actual outputs from multiple historical production tasks in the historical production database. ; Step 2: Based on the characteristics of the seamless steel pipe straightening process, construct and train a mechanism-machine learning hybrid model; Step 3: Use a hybrid prediction model to perform real-time quality prediction and obtain the prediction results; Step 4: Provide production guidance based on the prediction results to determine whether the accuracy requirements of the straightening task are met; The final quality prediction value Compared with the preset quality pass threshold Compare; like If so, it is determined that the current straightening process parameters are set reasonably and production can proceed. like If so, an early warning will be issued, and adjustments to the relevant straightening process parameters will be recommended; Step 5: Evaluate the stability of the straightening process through quality indicators, determine whether the quality fluctuation is within the allowable range, and then determine whether the finished seamless steel pipe meets the accuracy requirements of the straightening task.

2. The seamless steel pipe quality evaluation method based on mechanism model and machine learning according to claim 1, characterized in that, Each data sample includes: Input feature set X: includes the material parameters, geometric parameters, straightening equipment parameters, and straightening process parameters of the steel pipe; The material parameters include at least the yield strength. and elastic modulus The geometric parameters include at least the outer diameter. , inner diameter Steel pipe wall thickness The equipment parameters include at least the number of straightener rolls, the straightener roll diameter, the straightening speed, and the straightener roll spacing L; the process parameters include at least the theoretical reduction δ calculated based on the target plastic deformation rate η.

3. The seamless steel pipe quality evaluation method based on mechanism model and machine learning according to claim 2, characterized in that, In step 2, the construction and training of the hybrid prediction model specifically includes the following methods: Step 2.1: Construct a mathematical mechanism model for the straightening of seamless steel pipes; Step 2.2: Prediction bias of the computer model ; Step 2.3: Train the bias prediction machine learning model to form the bias prediction model; Step 2.4: Use the input feature set X of all historical data samples as the training input bias prediction model to obtain the bias correction value. ; Step 2.5: Perform quality prediction using the aforementioned deviation prediction model.

4. The seamless steel pipe quality evaluation method based on mechanism model and machine learning according to claim 3, characterized in that, In step 2.1, the mathematical mechanism modeling is based on elastoplastic mechanics and material deformation theory, constructs the basic parameter framework of the straightening process, describes the deformation law of the steel pipe under the action of the straightening roller, and establishes the basis for calculating the deformation amount.

5. The seamless steel pipe quality evaluation method based on mechanism model and machine learning according to claim 4, characterized in that, The core calculation formulas of the mechanism model include: Moment of inertia of steel pipe section : ； In the formula, Moment of inertia of steel pipe section (unit: ); The outer diameter of the steel pipe. The inner diameter of the steel pipe, and , For the steel pipe wall thickness; Relationship between the curvature and bending moment of steel pipe: ； In the formula, Indicates the bending radius (unit: m), reflecting the degree of bending of the steel pipe; Indicates the bending moment during the straightening process (unit: It is positively correlated with the reduction amount and the roller spacing; This represents the elastic modulus of the steel pipe (unit: Pa), an inherent property of the steel grade. Used to calculate the theoretical reduction of the straightening machine The formula, The formula for calculating the plastic deformation rate is as follows: ； In the formula, Indicates the rate of plastic deformation; Indicates the distance between the straightening rollers; This indicates the maximum permissible bending radius, which is determined by the material and specifications of the steel pipe. Iterative calculation model for residual deformation after multi-roll straightening: based on the original curvature, the inverse curvature applied by each straightening roll, and the material yield strength. Based on the cross-sectional properties, the elastic curvature and residual curvature are calculated iteratively to obtain the theoretical residual deformation distribution along the entire length of the steel pipe. The ratio of its maximum value to the length of the steel pipe is the theoretical straightness prediction value. The total deformation curvature, derived from the superposition of the original curvature and the inverse curvature, reflects the overall degree of deformation during the straightening process. Its formula is: ； In the formula, The total deformation curvature; The original curvature of the steel pipe; The reverse bending rate applied to the straightening roller; The curvature corresponding to the elastic recovery of the steel pipe after unloading is determined by the elastic properties of the material, and its formula is: ； In the formula, The yield strength of the steel pipe; This is the distance from the fiber with the largest cross-section to the neutral axis; The residual curvature of the steel pipe after unloading is given by the following formula: ； In the formula, Residual curvature directly corresponds to straightness; the smaller the residual curvature, the better the straightness. The critical curvature at which the surface fiber stress of the steel pipe reaches the yield strength, i.e., the yield curvature: ； The yield curvature is used to determine the starting point of plastic deformation; when the total deformation curvature... At this time, the steel pipe enters the plastic deformation stage.

6. The seamless steel pipe quality evaluation method based on mechanism model and machine learning according to claim 5, characterized in that, In step 2.2, the prediction bias of the computer model. ,include: For each historical data sample, the input feature set X is input into the mechanism model to obtain the mechanism prediction value. , namely the residual deformation after straightening, is used to simulate and predict the straightness of the seamless steel pipe after straightening; Residual deformation The ratio of the maximum value of the straightness to the length of the steel pipe is defined as straightness, thus completing the mapping from deformation to quality indicators. The mapped mechanistic model prediction values Used as a parameter for quality evaluation; Deviation between computer-predicted values ​​and actual values : 。 7. The seamless steel pipe quality evaluation method based on mechanism model and machine learning according to claim 6, characterized in that, In step 2.3, training the bias prediction machine learning model to form the bias prediction model includes: Using the input feature set X of all historical data samples as training input, and the corresponding prediction bias... The machine learning model is trained using the target variable as the training variable to obtain a well-trained bias prediction model.

8. A seamless steel pipe quality evaluation system based on mechanistic models and machine learning, characterized in that, It includes a data acquisition module, a prediction result calculation module, a first judgment module, and a second judgment module; The data acquisition module is used to acquire data samples and corresponding actual outputs from multiple historical production tasks from the historical production database. ; The prediction result calculation module constructs and trains a mechanism-machine learning hybrid model based on the characteristics of the seamless steel pipe straightening process; and uses the hybrid prediction model to perform real-time quality prediction to obtain the prediction result. The first judgment module provides production guidance based on the prediction results, determining whether the accuracy requirements of the straightening task are met; and sets the final quality prediction value. Compared with the preset quality pass threshold Compare; if If the current straightening process parameters are set correctly, production can proceed; otherwise... If so, an early warning will be issued, and adjustments to the relevant straightening process parameters will be recommended; The second judgment module evaluates the stability of the straightening process through evaluation indicators, determines whether the quality fluctuation is within the allowable range, and then determines whether the finished seamless steel pipe meets the accuracy requirements of the straightening task.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the seamless steel pipe quality evaluation method based on mechanism model and machine learning as described in any one of claims 1 to 7.

10. A storage device comprising a storage medium and a processor, the storage medium storing a computer program, characterized in that, When the processor executes the computer program, it implements the seamless steel pipe quality evaluation method based on mechanism model and machine learning as described in any one of claims 1 to 7.

Images