Quality detection method and system for seamless steel tube finished product

Abstract

The invention relates to the related technical field of quality detection, in particular to a quality detection method and system for a seamless steel tube finished product, and the method comprises the steps: collecting a tube blank material and size information of a seamless steel tube, and collecting heat treatment and surface treatment parameters; configuring a heat treatment analysis network and a surface treatment analysis network, and establishing a heat treatment performance analysis hyperplane and a surface treatment performance analysis hyperplane; structural point clouds are obtained through a phased array ultrasonic module, the seamless steel tube quality is evaluated through performance analysis hyperplane, and the technical problems that conventional nondestructive testing has limitation on sensitivity of tiny defects, destructive testing means are needed, and seamless steel tube finished product quality evaluation precision is low are solved. The technical effects that the heat treatment analysis network and the surface treatment analysis network are configured, the phased array ultrasonic module is combined for multi-angle monitoring, the quality of the seamless steel pipe is comprehensively evaluated, the quality of the seamless steel pipe finished product is efficiently and accurately detected, and the quality evaluation precision of the seamless steel pipe finished product is improved are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field related to quality inspection, and particularly relates to a quality inspection method and system for seamless steel pipe finished products. Background Art

[0002] Seamless steel pipes are widely used as basic industrial materials in multiple fields such as construction, machinery, petroleum, and chemical industry. The quality of finished products directly affects the safe operation and service life of related equipment. Conventional quality inspection methods for seamless steel pipes mainly rely on mechanical property tests and ultrasonic non-destructive testing. Although they can evaluate the mechanical properties and internal defects of steel pipes to a certain extent, there are limitations in terms of inspection efficiency, accuracy, and sensitivity to micro defects. It is easy to miss micro flaws, resulting in lax quality control and affecting the overall performance of products. With the continuous development of the manufacturing industry and the increasing requirements for product quality, high-efficiency and accurate quality inspection of seamless steel pipe finished products has become an urgent need in the industry.

[0003] In summary, there are technical problems in the prior art that conventional non-destructive testing has limitations in terms of sensitivity to micro defects, destructive testing means need to be relied on, and the accuracy of quality evaluation of seamless steel pipe finished products is low. Summary of the Invention

[0004] The present application provides a quality inspection system for seamless steel pipe finished products, aiming to solve the technical problems in the prior art that conventional non-destructive testing has limitations in terms of sensitivity to micro defects, destructive testing means need to be relied on, and the accuracy of quality evaluation of seamless steel pipe finished products is low.

[0005] In view of the above problems, the technical solution of the present application is as follows: On the one hand, this application provides a quality inspection method for seamless steel pipe finished products. Among them, the method includes: obtaining the basic production information of seamless steel pipe finished products, where the basic production information includes billet material information and billet size information; collecting a set of heat treatment parameters based on the steel pipe heat treatment process, where the set of heat treatment parameters includes subsets of heat treatment parameters corresponding to M heat treatment stages, and the heat treatment stages include normalizing stage, annealing stage, quenching stage, and tempering stage, M≥4 and M is a positive integer; collecting a set of surface treatment parameters based on the steel pipe surface treatment process, where the set of surface treatment parameters includes subsets of surface treatment parameters corresponding to N surface treatment stages, and the surface treatment stages include pickling stage, shot peening stage, phosphating treatment stage, and coating treatment stage, N≥4 and N is a positive integer; configuring a heat treatment analysis network through the steel pipe heat treatment process and the set of heat treatment parameters, and establishing a heat treatment performance analysis hyperplane; configuring a surface treatment analysis network through the steel pipe surface treatment process and the set of surface treatment parameters, and establishing a surface treatment performance analysis hyperplane; based on the basic production information, using a phased array ultrasonic module to perform multi-angle monitoring, obtaining the structural point cloud of seamless steel pipe finished products, and performing quality evaluation of seamless steel pipe finished products through the heat treatment performance analysis hyperplane and the surface treatment performance analysis hyperplane.

[0006] On the other hand, this application provides a quality inspection system for seamless steel pipe finished products. Among them, the system includes: a production information acquisition module for obtaining the basic production information of seamless steel pipe finished products, where the basic production information includes billet material information and billet size information; a heat treatment parameter collection module for collecting a set of heat treatment parameters based on the steel pipe heat treatment process, where the set of heat treatment parameters includes subsets of heat treatment parameters corresponding to M heat treatment stages, and the heat treatment stages include normalizing stage, annealing stage, quenching stage, and tempering stage, M≥4 and M is a positive integer; a surface treatment parameter collection module for collecting a set of surface treatment parameters based on the steel pipe surface treatment process, where the set of surface treatment parameters includes subsets of surface treatment parameters corresponding to N surface treatment stages, and the surface treatment stages include pickling stage, shot peening stage, phosphating treatment stage, and coating treatment stage, N≥4 and N is a positive integer; a heat treatment analysis module for configuring a heat treatment analysis network through the steel pipe heat treatment process and the set of heat treatment parameters, and establishing a heat treatment performance analysis hyperplane; a surface treatment analysis module for configuring a surface treatment analysis network through the steel pipe surface treatment process and the set of surface treatment parameters, and establishing a surface treatment performance analysis hyperplane; a quality evaluation module for performing multi-angle monitoring using a phased array ultrasonic module based on the basic production information, obtaining the structural point cloud of seamless steel pipe finished products, and performing quality evaluation of seamless steel pipe finished products through the heat treatment performance analysis hyperplane and the surface treatment performance analysis hyperplane.

[0007] In summary, one or more technical solutions provided in this application solve the technical problems that conventional non-destructive testing has limitations in terms of sensitivity to micro-defects, destructive testing means need to be used, and the accuracy of the quality assessment of seamless steel pipe finished products is low. By configuring a heat treatment analysis network and a surface treatment analysis network, combined with a phased array ultrasonic module for multi-angle monitoring, the quality of seamless steel pipes is comprehensively evaluated, and efficient and accurate quality inspection of seamless steel pipe finished products is carried out, achieving the technical effect of improving the accuracy of the quality assessment of seamless steel pipe finished products. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 FIG. is a schematic flow chart of a quality inspection method for seamless steel pipe finished products provided by this application; Figure 2 FIG. is a schematic structural diagram of a quality inspection system for seamless steel pipe finished products provided by this application.

[0009] Description of reference numerals: Production information acquisition module M100, heat treatment parameter acquisition module M200, surface treatment parameter acquisition module M300, heat treatment analysis module M400, surface treatment analysis module M500, quality assessment module M600. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0010] Embodiment 1 The following specifically describes this application with reference to the drawings. As Figure 1 shown, this application provides a quality inspection method for seamless steel pipe finished products, wherein the method includes: S1: Obtain the basic production information of the seamless steel pipe finished product, where the basic production information includes billet material information and billet size information; S2: Based on the steel pipe heat treatment process, collect a set of heat treatment parameters, where the set of heat treatment parameters includes subsets of heat treatment parameters corresponding to M heat treatment stages, where the heat treatment stages include normalizing stage, annealing stage, quenching stage, and tempering stage, M≥4 and M is a positive integer; S3: Based on the steel pipe surface treatment process, collect a set of surface treatment parameters, where the set of surface treatment parameters includes subsets of surface treatment parameters corresponding to N surface treatment stages, where the surface treatment stages include pickling stage, shot peening stage, phosphating treatment stage, and coating treatment stage, N≥4 and N is a positive integer.

[0011] Specifically, collect billet material information; determine the material type of the billet (such as carbon steel, alloy steel, stainless steel, etc.), understand its chemical composition and physical property indexes; record the material source and batch number for traceability. Record the billet size information: measure and record the original diameter, wall thickness, length and other dimensions of the billet, and confirm whether the dimensional tolerance meets the requirements of subsequent processing.

[0012] Collect the heat treatment parameter set and define the heat treatment stages: clarify the four basic stages of normalizing, annealing, quenching, and tempering. According to the actual process requirements, it also includes other special heat treatment stages. Further, for each heat treatment stage, record specific parameters such as temperature, holding time, cooling rate, etc., to ensure that the parameter subsets corresponding to M heat treatment stages are complete, and each subset contains at least the necessary key parameters, satisfying the condition of M≥4.

[0013] Collect the surface treatment parameter set and define the surface treatment stages: list the four basic steps of pickling, shot peening, phosphating treatment, and coating treatment. Similarly, add more treatment stages according to the actual situation. Further, for each surface treatment stage, collect relevant process parameters such as pickling solution concentration, shot peening intensity, phosphating solution composition, coating type and thickness, etc., to ensure that the parameter subsets of N surface treatment stages are detailed, satisfying the condition of N≥4.

[0014] S4: Configure the heat treatment analysis network and establish the heat treatment performance analysis hyperplane through the steel pipe heat treatment process and the heat treatment parameter set; S5: Configure the surface treatment analysis network and establish the surface treatment performance analysis hyperplane through the steel pipe surface treatment process and the surface treatment parameter set; S6: Based on the basic production information, use the phased array ultrasonic module for multi-angle monitoring to obtain the point cloud of the seamless steel pipe finished product structure, and conduct quality assessment of the seamless steel pipe finished product through the heat treatment performance analysis hyperplane and the surface treatment performance analysis hyperplane.

[0015] Specifically, collect and integrate the detailed heat treatment parameters of the steel pipe in each stage such as normalizing, annealing, quenching, and tempering, including temperature, time, cooling rate, etc., and confirm the heat treatment parameter subsets corresponding to each heat treatment stage in M heat treatment stages; then, through the steel pipe heat treatment process and the heat treatment parameter set, use the graph model to represent the parameter interaction in the heat treatment process, further analyze the correlation between parameters in the network, and identify the key parameter combinations that have a significant impact on the material properties; apply machine learning algorithms such as support vector machine (SVM) or neural network, and use the parameter subsets and material property data to train the model; through the model learning results, establish the heat treatment performance analysis hyperplane, and the heat treatment performance analysis hyperplane can distinguish the material property levels under different heat treatment effects.

[0016] Sort out the data of the steel pipe surface treatment process, covering the specific parameters of each stage such as pickling, shot peening, phosphating treatment, and coating treatment, and confirm the subset of heat treatment parameters corresponding to each heat treatment stage in the M heat treatment stages; similarly, through the steel pipe surface treatment process and the surface treatment parameter set, use a graph model to represent the parameter interaction in the surface treatment process, further analyze the correlation between parameters in the network, and identify the key parameter combinations that have a significant impact on the material properties; utilize the relationship between the surface treatment parameters and the expected surface properties, and establish a surface treatment performance analysis hyperplane through machine learning methods. The surface treatment performance analysis hyperplane is used to predict and evaluate the surface treatment effect.

[0017] Use phased array ultrasonic technology to conduct a full-range inspection of seamless steel pipes, collect data from multiple angles, and generate a three-dimensional point cloud of the internal structure and surface state of the steel pipes; map the obtained structural point cloud data to the previously established heat treatment performance analysis hyperplane and surface treatment performance analysis hyperplane; by comparing the positions of the point cloud data on the hyperplane, evaluate the internal quality (such as grain size, hardness distribution) after heat treatment and the quality after surface treatment (such as roughness, coating uniformity) of the steel pipes; comprehensively analyze the results and give an overall quality assessment report of the seamless steel pipe finished products, including existing defects, quality grade classification, etc., ensuring the high efficiency and accuracy of the seamless steel pipe quality inspection.

[0018] Furthermore, through the steel pipe heat treatment process and the heat treatment parameter set, configure a heat treatment analysis network and establish a heat treatment performance analysis hyperplane. The method of this application includes: Take the steel pipe heat treatment process as a line and the heat treatment parameter set as points to configure a heat treatment analysis network; through the heat treatment analysis network, analyze the relationship between the heat treatment parameter set and the material properties, and establish a heat treatment performance analysis hyperplane.

[0019] Specifically, record the heat treatment process, detail the heat treatment process of seamless steel pipes, clarify all included stages, namely normalizing, annealing, quenching, tempering, etc., to ensure that the operation process of each stage is clear and without omission; collect the specific parameters of each heat treatment stage, such as temperature, holding time, cooling rate, etc., to form M subsets of heat treatment parameters, ensuring that the parameters are complete and accurate; take the heat treatment process as the time line, regard each heat treatment stage as a node in the network, and the heat treatment parameters as the characteristic points on these nodes; utilize graph theory or neural network theory to construct an analysis network that reflects the interaction between the heat treatment process and parameters.

[0020] In the heat treatment analysis network, the potential impacts between various stages and of various parameters on the final material properties are represented by connections or weights, forming a complex analysis framework; test data of the material properties of seamless steel pipes under different combinations of heat treatment parameters are collected, including key indicators such as hardness, strength, and toughness; then, using historical data, the correlation between the heat treatment parameter set and the material properties is analyzed through machine learning algorithms (such as regression analysis, support vector machines, neural networks, etc.). During training, the heat treatment parameters serve as input variables and the performance indicators serve as output targets.

[0021] Based on the trained machine learning model, a multi-dimensional feature space is defined, where each dimension represents a heat treatment parameter or its derived feature; in the multi-dimensional feature space, one or more hyperplanes are determined, and the hyperplanes can best divide seamless steel pipes with different performance levels; the selection of the hyperplanes is based on the prediction ability of the model, that is, it can effectively classify or predict the properties of the steel pipes after heat treatment; then, an independent validation data set is used to test the prediction accuracy of the heat treatment performance analysis hyperplane to ensure that the model has good generalization ability; and according to the validation results, necessary adjustments and optimizations are carried out, including adjusting the network structure, selecting a more suitable algorithm, or optimizing the parameter settings until a satisfactory performance prediction effect is achieved. Through the above analysis, a heat treatment performance analysis hyperplane is constructed to ensure the efficiency and quality control of the heat treatment process.

[0022] Furthermore, through the heat treatment analysis network, the relationship between the heat treatment parameter set and the material properties is analyzed, and a heat treatment performance analysis hyperplane is established. The method of the present application includes: Based on the heat treatment parameter set, correlation extraction is carried out in combination with the material properties to obtain a set of correlation features; through the heat treatment analysis network, the relationship between the heat treatment parameter set and the material properties is analyzed, and a heat treatment performance analysis micro-space is constructed using machine learning; with the set of correlation features as the center, corresponding additional features are added to expand the feature dimension in the heat treatment performance analysis micro-space, and the heat treatment performance analysis hyperplane is established.

[0023] Specifically, parameter data of each stage in the heat treatment process (such as temperature, time, cooling rate, etc.) are comprehensively collected, and at the same time, the corresponding material property data of the seamless steel pipes (hardness, strength, toughness, etc.) are collected; statistical analysis or a preliminary machine learning model (such as random forest, principal component analysis PCA) is used to identify which heat treatment parameters have a significant correlation with the change of material properties, and the parameter combination that has the greatest impact on the material properties is extracted to form a set of correlation features; According to the extracted set of associated features, configure a suitable machine learning model (such as a deep neural network, support vector machine SVM) as the basis of the heat treatment analysis network. The model will take heat treatment parameters as input and the expected material properties as the output target; train the above model to make it learn the non-linear relationship between the set of heat treatment parameters and the material properties. This process essentially defines a local and highly correlated micro-space for heat treatment performance analysis in the feature space. In the micro-space for heat treatment performance analysis, the mapping relationship between heat treatment parameters and properties is quantified.

[0024] Analyze the existing set of associated features to identify other parameters or derived features (such as the square of the processing time, temperature gradient, etc.) that are missing but have supplementary value for performance evaluation. Add the additional features to the original set of associated features; based on the micro-space for heat treatment performance analysis, by adding these additional features, expand the dimension of the original feature space, aiming to more precisely depict the impact of the heat treatment process on the material properties and improve the discrimination and prediction accuracy of the model.

[0025] Based on the expanded feature dimension, use advanced machine learning techniques (such as kernel methods, high-dimensional space projection) to determine one or more hyper-planes for heat treatment performance analysis. These hyper-planes for heat treatment performance analysis can effectively divide seamless steel pipe instances with different performance levels; the position and direction of the hyper-planes reflect the optimal combination of heat treatment parameters, which are used to quickly evaluate and predict the heat treatment quality of seamless steel pipes.

[0026] Conduct cross-validation on the constructed model, and use the data set not involved in training to evaluate its prediction performance, including indicators such as accuracy, recall rate, F1 score, etc.; according to the verification results, make necessary adjustments to the model, such as adjusting model parameters, optimizing feature selection, etc., and then retrain the model until a satisfactory prediction effect is achieved; through the above steps, achieve accurate modeling from heat treatment parameters to material property prediction, providing support for the quality control and process optimization of seamless steel pipes.

[0027] Furthermore, in the micro-space for heat treatment performance analysis, expand the feature dimension and establish the hyper-plane for heat treatment performance analysis. The method of the present application includes: Taking the set of associated features as the center, a first additional feature set is set around the first associated feature element, where the first associated feature element is any one in the set of associated features; traversing the set of associated features, determining the first associated feature element and the first additional feature set, the second associated feature element and the second additional feature set,..., the U-th associated feature element and the U-th additional feature set; based on the first associated feature element and the first additional feature set, the second associated feature element and the second additional feature set,..., the U-th associated feature element and the U-th additional feature set, expanding the feature dimension in the heat treatment performance analysis microspace, and establishing the heat treatment performance analysis hyperplane.

[0028] Randomly select one item from the obtained set of associated features as the first associated feature element; it will serve as the basis for constructing the additional feature set; around the first associated feature element, design and set the first additional feature set; the additional features include derivative variables (such as derivatives, integrals), interaction terms (products with other features), function transformations (logarithmic, exponential conversions), etc. of the central feature, aiming to enrich the feature expression and capture more potential performance influencing factors.

[0029] Traverse the entire set of associated features, repeat the operation for each associated feature element, and respectively construct the corresponding additional feature sets, including the first associated feature element and the first additional feature set, the second associated feature element and the second additional feature set,..., the U-th associated feature element and the U-th additional feature set; integrate all the basic associated features and their respective corresponding additional feature sets together to form an extended feature set, which is essentially adding new dimensions on the basis of the original microspace, and each dimension represents the influence of heat treatment parameters and their derivative effects on different aspects of material properties.

[0030] Then, use machine learning algorithms (such as support vector machines, neural networks, deep learning models) to learn and establish the heat treatment performance analysis hyperplane in this extended feature space; the selection of the hyperplane aims to maximize the separation of seamless steel pipe samples of different quality grades to achieve efficient quality assessment; further, apply cross-validation or reserve a part of the data as the test set to evaluate the prediction performance of the hyperplane, including indicators such as accuracy rate and recall rate; according to the verification results, adjust the feature selection, model parameters or algorithms until the model performance meets the established standards. Through the above steps, expand the feature dimension of the heat treatment performance analysis and establish a hyperplane that can accurately reflect and predict the heat treatment quality of seamless steel pipes, providing a scientific basis for quality control and process optimization.

[0031] Furthermore, the method of the present application includes: In the heat treatment performance analysis microspace, U explosion microspaces are set by comparing the first associated feature element with the first additional feature set, the second associated feature element with the second additional feature set, ……, and the U-th associated feature element with the U-th additional feature set; based on the heat treatment performance analysis microspace, the U explosion microspaces are annexed, and the heat treatment performance analysis microspace is updated; through the updated heat treatment performance analysis microspace, the heat treatment performance analysis hyperplane is established.

[0032] For each associated feature element (from the first associated feature element to the U-th associated feature element), its corresponding additional feature set is identified, which together constitutes the multi-dimensional basis for heat treatment performance analysis; for each associated feature element and its corresponding additional feature set, an explosion microspace is created; this means that each basic feature is regarded as the center, and its additional features form a local and highly concentrated feature space around this center; the U explosion microspaces respectively represent in-depth exploration of each key feature and its influence range.

[0033] Taking into comprehensive consideration the information of the U explosion microspaces, through mathematical merging or fusion operations (such as weighted average, principal component analysis PCA or other dimensionality reduction techniques), the U explosion microspaces are unified into a broader and more complex heat treatment performance analysis microspace, aiming to integrate the results of all local feature explorations and form a performance analysis framework from a global perspective; according to the merged information, the original heat treatment performance analysis microspace is updated, and the update includes adjusting feature weights, redefining the interaction rules between features, introducing new dimensionality divisions, etc., to ensure that the new microspace can more comprehensively and accurately reflect the complexity of the heat treatment process.

[0034] Then, using the updated heat treatment performance analysis microspace, through advanced data analysis or machine learning models (such as support vector machine SVM, deep neural network DNN, etc.), a heat treatment performance analysis hyperplane is constructed. The heat treatment performance analysis hyperplane aims to divide different quality regions, making it feasible and accurate to predict the quality of steel pipes based on heat treatment parameters; further, the heat treatment performance analysis hyperplane is optimized to ensure that it can best classify or regression predict the quality grade of seamless steel pipes; at the same time, the model performance is evaluated through methods such as cross-validation and AUC-ROC curve analysis, and parameter tuning is carried out if necessary. Through the above steps, the understanding of the heat treatment process is deepened, explosion microspaces are created and merged, and the accurate construction of the heat treatment performance analysis hyperplane is achieved, thereby improving the accuracy and efficiency of seamless steel pipe quality inspection.

[0035] Furthermore, the method of the present application further includes: Obtain the quality standards for seamless steel pipe finished products, where the quality standards for seamless steel pipe finished products include surface roughness, internal cavity size, and wall thickness uniformity; after completing the steel pipe heat treatment process and the steel pipe surface treatment process, conduct a quality grade assessment by comparing each quality standard threshold in the quality standards for seamless steel pipe finished products to obtain the quality grade of the seamless steel pipe; based on the quality grade of the seamless steel pipe, mark the quality standards for the seamless steel pipe finished products.

[0036] Clarify the specific quality standards for seamless steel pipe finished products, including surface roughness, internal cavity size, and wall thickness uniformity. Among them, each standard is a key indicator for measuring the performance of the steel pipe and is directly related to its reliability and service life in specific application scenarios; according to the established heat treatment processes (normalizing, annealing, quenching, tempering, etc.), process the steel pipe and record the heat treatment parameters at each stage to ensure that the material properties meet the expectations; successively carry out processes such as pickling, shot peening, phosphating treatment, and coating treatment, and at the same time collect relevant surface treatment parameters to improve the surface quality and corrosion resistance of the steel pipe.

[0037] Integrate all the parameters collected from the heat treatment and surface treatment processes, including the heat treatment parameter set and the surface treatment parameter set; use the heat treatment parameter set and the surface treatment parameter set to configure the heat treatment analysis network and the surface treatment analysis network, and establish a performance analysis hyperplane through machine learning methods to predict and evaluate the heat treatment effect and surface treatment effect of the steel pipe.

[0038] Use a phased array ultrasonic module to monitor the finished steel pipe from multiple angles to form structural point cloud data, which is an important means for non-destructive testing of the internal structure and defects of the steel pipe; compare the key indicators such as surface roughness, internal cavity size, and wall thickness uniformity analyzed from the point cloud data with the pre-set quality standard thresholds one by one; according to the compliance degree of each indicator with the standard threshold, conduct a quality grade assessment of the seamless steel pipe finished products, including excellent, good, qualified, unqualified, etc.

[0039] Based on the evaluated quality grade of the seamless steel pipe, mark the quality standards for each batch or each steel pipe of the finished products, which helps to trace and distinguish products with different quality levels; according to the quality assessment results and market feedback, make necessary adjustments and optimizations to the heat treatment parameters, surface treatment parameters, and the entire detection process to continuously improve product quality and production efficiency, forming a closed loop, from formulating standards to performing inspections, and then adjusting and optimizing according to the inspection results to ensure that the seamless steel pipe finished products always maintain a high standard of quality.

[0040] Furthermore, the method of this application also includes: Call the CMA-ES database and, in combination with each quality standard threshold in the finished seamless steel pipe quality standard, perform improved decision optimization. The improved decision is optimized and configured with the goal of maximizing the quality grade of the seamless steel pipe and maximizing the processing efficiency.

[0041] In the process of calling the CMA-ES database (Covariance Matrix Adaptation Evolution Strategy) for improved decision optimization, aiming at the finished seamless steel pipe quality standard and its thresholds, with the goal of maximizing the quality grade and processing efficiency of the seamless steel pipe, first, collect historical data from the existing system, including production information, heat treatment parameters, surface treatment parameters, known quality evaluation results, and processing efficiency data of each batch of seamless steel pipes; numericalize the specific threshold values of the quality standards (surface roughness, internal cavity size, wall thickness uniformity) as the target boundary conditions for optimization.

[0042] Design a composite objective function that simultaneously considers the improvement of the quality grade of the seamless steel pipe (for example, by assigning higher scores to higher quality grades) and the improvement of processing efficiency (such as reducing processing time or cost). The objective function needs to balance these two aspects to ensure that both can be improved during the optimization process; set initial parameters for the CMA-ES algorithm, such as population size, initial search range, mutation variance, etc., and these parameters are adjusted according to the complexity and dimension of the problem; according to the objective function, define the fitness evaluation criteria, that is, how to evaluate the degree to which each solution (i.e., a set of specific heat treatment and surface treatment parameters) generated by the algorithm meets the quality standards and improves processing efficiency.

[0043] The CMA-ES algorithm will generate a series of new parameter combinations (strategies) according to the currently estimated best parameter distribution, simulate the application of the corresponding strategies to the production process, calculate the corresponding quality grade and processing efficiency; evaluate the fitness of each new strategy, that is, calculate its score according to the objective function; based on the evaluation results, update the parameter distribution and gradually approach a better solution; determine whether a good enough strategy has been found or whether the algorithm has converged by monitoring the change in the fitness value or reaching a predetermined number of iterations.

[0044] In the actual application process, it also includes selecting one or more strategies from the optimal solutions found in the CMA-ES optimization process for actual production testing; applying the optimal strategy in a part of the production line or during a specific time period for small-batch trial production, and collecting actual data to verify the optimization effect; according to the results of the trial production, evaluate whether the actual quality grade and processing efficiency of the seamless steel pipe meet the expected goals; if not, analyze the reasons, adjust the optimization model or parameters, and perform the optimization iteration again.

[0045] After confirming the effectiveness of the optimization strategy, deploy it to the entire production process to achieve large-scale production optimization; continuously monitor production data, regularly review the optimization effect, and continuously adjust the optimization strategy according to market changes, the application of new materials or new technologies to ensure continuous improvement of the quality and production efficiency of seamless steel pipes; utilize the optimization ability of the CMA-ES database to systematically improve the quality control and production efficiency of seamless steel pipe finished products and achieve more intelligent and efficient production management.

[0046] In summary, the beneficial effects of the embodiments of the present application are as follows: 1. Integrate the comprehensive collection of billet material information, size information, heat treatment, and surface treatment parameters, and construct an all-round quality monitoring system from the source to the finished product, greatly improving the detection efficiency.

[0047] 2. Use phased array ultrasonic technology for multi-angle monitoring, combined with heat treatment performance analysis hyperplane and surface treatment performance analysis hyperplane, to achieve accurate evaluation of the internal structure and surface quality of seamless steel pipes and improve the detection accuracy.

[0048] 3. Introduce machine learning and CMA-ES database optimization technology, make improvement decisions on the production process according to the detection data and quality standard thresholds, and continuously optimize the heat treatment and surface treatment parameters and the feedback mechanism to promote the continuous improvement of product quality and the continuous optimization of the production process, and promote the maximization of seamless steel pipe quality and the optimization of production efficiency.

[0049] 4. Establish clear quality standards, including surface roughness, internal cavity size, wall thickness uniformity, etc., to achieve automatic quality grade assessment and improve the consistency and standardization level of quality control.

[0050] 5. Since in the heat treatment performance analysis microspace, U explosion microspaces are set by comparing the first associated feature element with the first additional feature set, the second associated feature element with the second additional feature set,..., the Uth associated feature element with the Uth additional feature set; based on the heat treatment performance analysis microspace, the U explosion microspaces are annexed and the heat treatment performance analysis microspace is updated; through the updated heat treatment performance analysis microspace, a heat treatment performance analysis hyperplane is established. Through the above steps, the understanding of the heat treatment process is deepened, the explosion microspaces are created and merged, and the accurate construction of the heat treatment performance analysis hyperplane is achieved, thereby improving the accuracy and efficiency of seamless steel pipe quality detection.

[0051] Embodiment 2 Based on the same inventive concept as the quality detection method for seamless steel pipe finished products in the foregoing embodiment, as Figure 2 shown, the embodiments of the present application provide a quality detection system for seamless steel pipe finished products, wherein the system includes: A production information acquisition module M100 is used to acquire the basic production information of seamless steel pipe finished products. The basic production information includes billet material information and billet size information. A heat treatment parameter acquisition module M200 is used to collect a set of heat treatment parameters based on the steel pipe heat treatment process. The set of heat treatment parameters includes subsets of heat treatment parameters corresponding to M heat treatment stages. Among them, the heat treatment stages include normalizing stage, annealing stage, quenching stage, and tempering stage, where M≥4 and M is a positive integer. A surface treatment parameter acquisition module M300 is used to collect a set of surface treatment parameters based on the steel pipe surface treatment process. The set of surface treatment parameters includes subsets of surface treatment parameters corresponding to N surface treatment stages. Among them, the surface treatment stages include pickling stage, shot peening stage, phosphating treatment stage, and coating treatment stage, where N≥4 and N is a positive integer. A heat treatment analysis module M400 is used to configure a heat treatment analysis network through the steel pipe heat treatment process and the set of heat treatment parameters, and establish a heat treatment performance analysis hyperplane. A surface treatment analysis module M500 is used to configure a surface treatment analysis network through the steel pipe surface treatment process and the set of surface treatment parameters, and establish a surface treatment performance analysis hyperplane. A quality assessment module M600 is used to perform multi-angle monitoring using a phased array ultrasonic module based on the basic production information, obtain the structure point cloud of the seamless steel pipe finished product, and perform quality assessment of the seamless steel pipe finished product through the heat treatment performance analysis hyperplane and the surface treatment performance analysis hyperplane.

[0052] Furthermore, the heat treatment analysis module M400 is used to execute the following method: Taking the steel pipe heat treatment process as a line and the set of heat treatment parameters as points, configure a heat treatment analysis network. Through the heat treatment analysis network, analyze the relationship between the set of heat treatment parameters and material properties, and establish a heat treatment performance analysis hyperplane.

[0053] Furthermore, the heat treatment analysis module M400 is also used to execute the following method: Based on the set of heat treatment parameters, perform associated extraction in combination with material properties to obtain an associated feature set. Through the heat treatment analysis network, analyze the relationship between the set of heat treatment parameters and material properties, and construct a heat treatment performance analysis microspace using machine learning. Taking the associated feature set as the center, adding corresponding additional features, expand the feature dimension in the heat treatment performance analysis microspace, and establish the heat treatment performance analysis hyperplane.

[0054] Furthermore, the heat treatment analysis module M400 is also used to execute the following method: Taking the associated feature set as the center, a first additional feature set is set around the first associated feature element, where the first associated feature element is any one in the associated feature set; Traverse the associated feature set to determine the first associated feature element and the first additional feature set, the second associated feature element and the second additional feature set,..., the U-th associated feature element and the U-th additional feature set; Based on the first associated feature element and the first additional feature set, the second associated feature element and the second additional feature set,..., the U-th associated feature element and the U-th additional feature set, expand the feature dimension in the heat treatment performance analysis microspace to establish the heat treatment performance analysis hyperplane.

[0055] Furthermore, the heat treatment analysis module M400 is also used to execute the following method: In the heat treatment performance analysis microspace, set U explosion microspaces by comparing the first associated feature element and the first additional feature set, the second associated feature element and the second additional feature set,..., the U-th associated feature element and the U-th additional feature set; Based on the heat treatment performance analysis microspace, annex the U explosion microspaces and update the heat treatment performance analysis microspace; Establish the heat treatment performance analysis hyperplane through the updated heat treatment performance analysis microspace.

[0056] Furthermore, the quality inspection system for seamless steel pipe finished products is also used to execute the following method: Obtain the quality standards for seamless steel pipe finished products, which include surface roughness, internal cavity size, and wall thickness uniformity; After completing the steel pipe heat treatment process and the steel pipe surface treatment process, conduct a quality grade assessment by comparing each quality standard threshold in the quality standards for seamless steel pipe finished products to obtain the quality grade of the seamless steel pipe; Based on the quality grade of the seamless steel pipe, mark the quality standards for the seamless steel pipe finished products.

[0057] Furthermore, the quality inspection system for seamless steel pipe finished products is also used to execute the following method: Call the CMA-ES database, and combine each quality standard threshold in the quality standards for seamless steel pipe finished products to perform improved decision optimization. The improved decision is optimized and configured with the goal of maximizing the quality grade of the seamless steel pipe and maximizing the processing efficiency.

[0058] In summary, any step can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor, without any further limitation here.

[0059] Furthermore, the above technical solutions only represent the preferred technical solutions of the technical solutions of the embodiments of the present application. Some changes that those skilled in the art may make to some parts thereof all reflect the principles of the novel embodiments of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application.

Claims

1. A quality inspection method for seamless steel pipe finished products, characterized in that: The method comprises: Obtaining basic production information of seamless steel pipe finished products, wherein the basic production information includes tube blank material information and tube blank size information; Based on the heat treatment process of the steel pipe, a heat treatment parameter set is collected, wherein the heat treatment parameter set includes a heat treatment parameter subset corresponding to M heat treatment stages, wherein the heat treatment stages include a normalizing stage, an annealing stage, a quenching stage, and a tempering stage, and M≥4 and M is a positive integer; Based on the steel pipe surface treatment process, a surface treatment parameter set is collected, and the surface treatment parameter set includes a surface treatment parameter subset corresponding to N surface treatment stages, wherein the surface treatment stage includes a pickling stage, a shot peening stage, a phosphating stage, and a coating stage, and N ≥ 4 and N is a positive integer; According to the steel pipe heat treatment process and heat treatment parameter set, a heat treatment analysis network is configured, and a heat treatment performance analysis hyperplane is established; By means of the steel pipe surface treatment process and the surface treatment parameter set, a surface treatment analysis network is configured, and a surface treatment performance analysis hyperplane is established; Based on the basic production information, a phased array ultrasonic module is used to perform multi-angle monitoring to obtain the structural point cloud of the seamless steel pipe finished product, and the quality of the seamless steel pipe finished product is evaluated through the heat treatment performance analysis hyperplane and the surface treatment performance analysis hyperplane.

2. The quality inspection method for seamless steel pipe finished product according to claim 1, characterized in that: By means of the steel pipe heat treatment process and the heat treatment parameter set, a heat treatment analysis network is configured, and a heat treatment performance analysis hyperplane is established. The method comprises: The heat treatment analysis network is configured with the steel pipe heat treatment process as a line and the heat treatment parameter set as a point; The relationship between the heat treatment parameter set and material properties is analyzed through the heat treatment analysis network, and a heat treatment performance analysis hyperplane is established.

3. The quality inspection method for seamless steel pipe finished product according to claim 2, characterized in that: By using the heat treatment analysis network, analyzing the relationship between the heat treatment parameter set and material properties, and establishing a heat treatment performance analysis hyperplane, the method includes: Based on the heat treatment parameter set, associated extraction is performed in combination with material properties to obtain an associated feature set; By means of the heat treatment analysis network, the relationship between the heat treatment parameter set and the material properties is analyzed, and a heat treatment performance analysis microspace is constructed by machine learning; The associated feature set is taken as the center, corresponding additional features are added, the feature dimension is expanded in the heat treatment performance analysis microspace, and the heat treatment performance analysis hyperplane is established.

4. The quality inspection method for seamless steel pipe finished product according to claim 3, characterized in that: Expanding the characteristic dimension in the heat treatment performance analysis microspace and establishing the heat treatment performance analysis hyperplane, the method comprising: Taking the associated feature set as the center, and surrounding the first associated feature element, a first additional feature set is set, wherein the first associated feature element is any one item in the associated feature set; Traversing the associated feature set, determining a first associated feature element and a first additional feature set, a second associated feature element and a second additional feature set, ..., a Uth associated feature element and a Uth additional feature set; Based on the first associated feature element and the first additional feature set, the second associated feature element and the second additional feature set, ..., the Uth associated feature element and the Uth additional feature set, the feature dimension is expanded in the heat treatment performance analysis microspace and the heat treatment performance analysis hyperplane is established.

5. The quality inspection method for seamless steel pipe finished product according to claim 4, characterized in that: The method comprises: In the heat treatment performance analysis microspace, by comparing the first associated characteristic element and the first additional characteristic set, the second associated characteristic element and the second additional characteristic set, ..., the Uth associated characteristic element and the Uth additional characteristic set, U explosion microspaces are set; Based on the heat treatment performance analysis microspace, the U explosion microspaces are annexed, and the heat treatment performance analysis microspace is updated; The heat treatment performance analysis hyperplane is established through the updated heat treatment performance analysis microspace.

6. The quality inspection method for seamless steel pipe finished product according to claim 1, characterized in that: Obtaining quality standards for seamless steel pipe finished products, wherein the quality standards for seamless steel pipe finished products include surface roughness, internal cavity size, and wall thickness uniformity; After completing the steel pipe heat treatment process and the steel pipe surface treatment process, the quality grade is evaluated by comparing each quality standard threshold in the seamless steel pipe finished product quality standard to obtain the quality grade of the seamless steel pipe; Based on the quality grade of the seamless steel pipe, the quality standard of the finished seamless steel pipe is marked.

7. The quality inspection method for seamless steel pipe finished product according to claim 6, characterized in that: The method comprises: The CMA-ES database is called, and each quality standard threshold in the seamless steel pipe finished product quality standard is combined to optimize the improved decision. The improved decision is optimized with the goal of maximizing the seamless steel pipe quality grade and maximizing the processing efficiency.

8. A quality inspection system for seamless steel pipe finished products, characterized in that: For implementing the quality inspection method for seamless steel pipe products according to any one of claims 1 to 7, the system comprises: A production information acquisition module is used to acquire basic production information of seamless steel pipe finished products, wherein the basic production information includes tube blank material information and tube blank size information; A heat treatment parameter collection module is used to collect a heat treatment parameter set based on a steel pipe heat treatment process, wherein the heat treatment parameter set includes a heat treatment parameter subset corresponding to M heat treatment stages, wherein the heat treatment stages include a normalizing stage, an annealing stage, a quenching stage, and a tempering stage, and M≥4 and M is a positive integer; A surface treatment parameter collection module is used to collect a surface treatment parameter set based on the steel pipe surface treatment process, wherein the surface treatment parameter set includes a surface treatment parameter subset corresponding to N surface treatment stages, wherein the surface treatment stages include a pickling stage, a shot peening stage, a phosphating stage, and a coating stage, and N ≥ 4 and N is a positive integer; A heat treatment analysis module, used to configure a heat treatment analysis network through the steel pipe heat treatment process and heat treatment parameter set, and to establish a heat treatment performance analysis hyperplane; A surface treatment analysis module, used to configure a surface treatment analysis network through the steel pipe surface treatment process and surface treatment parameter set, and to establish a surface treatment performance analysis hyperplane; The quality assessment module is used to perform multi-angle monitoring based on the basic production information using a phased array ultrasonic module to obtain the structural point cloud of the seamless steel pipe finished product, and to perform quality assessment of the seamless steel pipe finished product through the heat treatment performance analysis hyperplane and the surface treatment performance analysis hyperplane.

Images