Steel billet internal defect classification method, system and equipment based on PSO-SVM (Particle Swarm Optimization-Support Vector Machine)

Abstract

The invention discloses a billet internal defect classification method, system and device based on PSO-SVM, and the method comprises the steps: S1, collecting billet internal ultrasonic signal data, and carrying out the preprocessing; s2, inputting the preprocessed data into a pre-trained defect classification model for identification; wherein the defect classification model comprises a particle swarm algorithm and a support vector machine; and S3, outputting a billet internal defect type identification result. The method is suitable for flaw detection of the continuous casting billet, is a more accurate internal flaw detection identification method, provides acceptable improvement measures for production of forgings, and is beneficial to quality control of the forgings.

Description

Technical Field

[0001] This application relates to the technical field of non-destructive testing of metal materials, and particularly relates to a method, system, and device for classifying internal defects of steel billets based on PSO-SVM. Background Art

[0002] In recent years, with the rapid development of the scale of wind power generation, the market demand for wind turbine towers has become increasingly large, and the quality of tower flange, the key connecting part of wind turbine towers, has received more and more attention. The tower flange of wind turbine towers is the key connecting part, supporting part, and stress-bearing part of wind turbine towers, and is an important part of wind power generation equipment. As an intermediate product of wind turbine tower flange, the internal quality of steel billets directly affects their service performance after forging. Therefore, it is extremely important to timely detect different internal defects of steel billets, such as porosity, shrinkage cavity, crack, segregation, etc., and propose different forging process improvement measures for different defects to improve the subsequent quality of products.

[0003] Currently, the main method for flaw detection of large-section steel billets is to use hand-held flaw detection equipment. There are some problems with hand-held detection. For example, it usually requires technicians to have proficient technical skills and make accurate descriptions of the results. Otherwise, there may be differences in the views of different personnel on the results; the flaw detection results cannot well improve the material utilization rate. Therefore, it is necessary to study a more accurate internal defect flaw detection and identification method applicable to continuous casting billets, provide acceptable improvement measures for the production of forgings, and be conducive to the quality control of forgings. Summary of the Invention

[0004] This application provides a method, system, and device for classifying internal defects of steel billets based on PSO-SVM to solve the above problems.

[0005] On the one hand, this application provides a method for classifying internal defects of steel billets based on PSO-SVM, and the method includes the following steps:

[0006] Step S1: Collect ultrasonic signal data inside the steel billet and perform preprocessing;

[0007] Step S2: Input the preprocessed data into a pre-trained defect classification model for identification; wherein, the defect classification model includes a particle swarm optimization algorithm and a support vector machine;

[0008] Step S3: Output the identification result of the internal defect type of the steel billet.

[0009] In an implementation manner of this application, the method further includes:

[0010] Select a linear kernel SVM kernel function according to the characteristics of the data;

[0011] Initialize the parameters of the SVM model, including the penalty parameter and the kernel function parameter.

[0012] In one implementation of the present application, the method further includes:

[0013] Optimize the model based on the particle swarm algorithm;

[0014] Select the optimal parameters;

[0015] Conduct model evaluation.

[0016] In one implementation of the present application, optimizing the model based on the particle swarm algorithm specifically includes:

[0017] Set the particle swarm size N, and each particle represents a set of SVM parameter combinations;

[0018] Train each particle using the training set to obtain an SVM model, and use the test set to evaluate the classification accuracy of the model. Take the classification accuracy as the fitness value of the particle;

[0019] According to the fitness value of the particle, update the velocity and position of each particle to make the particle move towards a better solution;

[0020] To improve the search efficiency, a crossover strategy can be introduced, that is, in each iteration, randomly select two particles and exchange their position information with a certain probability to fuse the information between different particles;

[0021] Check whether the stop condition is met, such as reaching the maximum number of iterations or the fitness value no longer changes. If the condition is met, stop the iteration; otherwise, repeat updating the velocity and position of each particle and continue the iteration.

[0022] In one implementation of the present application, selecting the optimal parameters specifically includes:

[0023] After the iteration ends, select the particle with the highest fitness value, and the corresponding parameter combination is the optimal parameter of the SVM model;

[0024] Construct the final SVM model using the optimal parameters and use it for the classification of internal defects in steel billets.

[0025] In one implementation of the present application, the identification results of the internal defects in the steel billet include: internal cracks, pores, inclusions, and no defects; the classification results are output in the form of probabilities.

[0026] In one implementation of the present application, the internal ultrasonic signal data of the steel billet is obtained through an ultrasonic flaw detector or an X-ray imaging device, and the preprocessing includes: noise reduction, normalization, and standardization operations.

[0027] The present application also provides a billet internal defect classification system based on PSO-SVM, and the system includes:

[0028] A data collection module, configured to collect ultrasonic signal data inside the billet and perform preprocessing;

[0029] A prediction module, configured to input the preprocessed data into a pre-trained defect classification model for recognition; wherein, the defect classification model includes a particle swarm algorithm and a support vector machine;

[0030] A result output module, configured to output the recognition result of the internal defect type of the billet.

[0031] The present application also provides a billet internal defect classification device based on PSO-SVM, and the device includes:

[0032] At least one processor; and,

[0033] A memory communicatively connected to the at least one processor; wherein,

[0034] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can complete the foregoing billet internal defect classification method based on PSO-SVM.

[0035] The billet internal defect classification method, system, and device provided by the present application have the following beneficial effects:

[0036] (1). Through iterative search, PSO can find the optimal parameter combination in the SVM model, thereby constructing a hyperplane with better classification performance, improving the classification accuracy. The PSO algorithm has a powerful global search ability, can explore different regions of the solution space, avoid falling into local optimal solutions, and thus find better SVM model parameters;

[0037] (2). Compared with traditional methods such as grid search and random search, the PSO algorithm can converge to the optimal solution faster through an intelligent search strategy, thereby reducing the parameter search space and time. The PSO algorithm is essentially parallel and suitable for implementation on a multi-processor system; through parallel processing, the execution efficiency of the algorithm can be further improved and the training speed can be accelerated;

[0038] (3). By optimizing the SVM model parameters with PSO, overfitting of the model can be avoided while ensuring the classification accuracy, and the generalization ability of the model can be improved. The PSO-SVM model can adapt to data sets of different scales and complexities and maintain stable classification performance;

[0039] (4) The concept of the PSO algorithm is simple and easy to implement programmatically. It does not involve complex mathematical formulas or profound mathematical theories. Compared with other evolutionary algorithms, the PSO algorithm has fewer parameters to be adjusted, reducing the complexity and difficulty of parameter tuning. Description of the Drawings

[0040] The drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0041] Figure 1 It is a flowchart of a method for classifying internal defects of steel billets based on PSO-SVM provided by an embodiment of the present application;

[0042] Figure 2 It is the best fitness curve graph provided by an embodiment of the present application;

[0043] Figure 3 It is a comparison graph of prediction results of the training set provided by an embodiment of the present application;

[0044] Figure 4 It is a comparison graph of prediction results of the test set provided by an embodiment of the present application;

[0045] Figure 5 It is a confusion matrix graph of the training set provided by an embodiment of the present application;

[0046] Figure 6 It is a confusion matrix graph of the test set provided by an embodiment of the present application;

[0047] Figure 7 It is a composition graph of a system for classifying internal defects of steel billets based on PSO-SVM provided by an embodiment of the present application;

[0048] Figure 8 It is a schematic diagram of a device for classifying internal defects of steel billets based on PSO-SVM provided by an embodiment of the present application. Detailed Embodiments

[0049] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0050] The embodiments of the present application provide a method, system, and device for classifying internal defects of steel billets based on PSO-SVM. The technical solutions proposed in the embodiments of the present application will be described in detail below with reference to the drawings.

[0051] Figure 1 This is a flowchart of a method for classifying internal defects of steel billets based on PSO-SVM provided by an embodiment of the present application. As Figure 1 shown, the method mainly includes the following steps:

[0052] Step S1: Collect ultrasonic signal data inside the steel billet and perform preprocessing;

[0053] Step S2: Input the preprocessed data into a pre-trained defect classification model for recognition; wherein, the defect classification model includes a particle swarm algorithm and a support vector machine;

[0054] Step S3: Output the recognition result of the internal defect type of the steel billet.

[0055] In the embodiment of the present application, low-magnification samples of porosity, shrinkage cavity, crack, and segregation of round billets with the grade of S355NL and the specification of Ф800 are selected. A handheld flaw detection device is used to take points for circumferential flaw detection to obtain ultrasonic CSV file data, which is used as a sample training set and a test set for training the SVM model and evaluating the model performance.

[0056] The data set is a matrix sequence of 356 * 13. The data set includes 12 features and 4 labels. Columns 1 - 12 are sample features, and the 13th column is the sample label. Each row is a sample, with a total of 356 samples. The training set includes 240 samples, and the test set includes 116 samples. The sample composition is shown in Table 1.

[0057] Table 1 Sample composition of the training set and the test set

[0058]

[0059] Therefore, rows 1 to 240 and columns 1 to 12 are used as the input of the training set; rows 1 to 240 and the 13th column are used as the output of the training set. Rows 241 to 356 and columns 1 to 12 are used as the input of the test set; rows 241 to 356 and the 13th column are used as the output of the test set. M is the number of samples in the training set, and N is the number of samples in the test set. The specific algorithm is shown in Table 2.

[0060] Table 2 Partitioning algorithm for the training set and the test set

[0061]

[0062] The divided data is normalized to eliminate the influence of the dimension on the model. The input of the training set is normalized to the range of [0, 1] according to the Min-Max normalization criterion to obtain the normalized input of the training set. The test set is processed according to the normalization criterion of the training set to obtain the normalized input of the test set. The specific algorithm is shown in Table 3.

[0063] Data Normalization Algorithm in Table 3

[0064]

[0065] The normalization requires each column to be a sample, which does not meet the requirement that each row of the model is a sample. Therefore, the normalized data is transposed.

[0066] The settings of the ion swarm hyperparameter algorithm are shown in Table 4. Among them, c1: initially 1.5, representing the local search ability of the pso parameter; c2: initially 1.7, representing the global search ability of the pso parameter; maxgen represents the maximum number of iterations set to 100, that is, an optimization algorithm iterates 100 times; 5 ions are optimized each time; k represents the relationship coefficient between the rate and the ion (V = kX), and the initial value is set to 0.6; wV represents the rate update weight coefficient, and the initial value is set to 1; wP is initially 1, representing the population update elasticity coefficient; v: initially 3, representing the number of SVM cross-validation times; c and g are the SVM population search ranges, both belonging to the range of [0 - 100].

[0067] Table 4 Settings of the Ion Swarm Hyperparameter Algorithm

[0068]

[0069]

[0070] After setting the parameters, run the optimization algorithm, put the training set input, training set output and other parameters into the function, and extract the best parameters c and g. And generate the best fitness curve. As Figure 2 shown. From Figure 2 it can be seen that as the particle swarm optimization algorithm iterates, the error rate of cross-validation continuously decreases and finally converges, with a total of 100 iterations.

[0071] Table 5 Definition Method of the pso-svm-class Function

[0072]

[0073] The optimization algorithm of the particle swarm optimization support vector machine includes function definition (shown in Table 5), parameter initialization, setting the maximum speed, determining the error threshold, population initialization, initializing the extreme value and extreme point, determining the average fitness, iterative optimization, determining the fitness curve, and optimal value assignment process. First, parameter initialization is required. Secondly, the initial velocity of the ion swarm needs to be set. Using the maximum range of parameters c and g, multiplying by the coefficient k, the maximum and minimum values (negative values) of its velocity are obtained. Define the error threshold, and then perform population initialization, randomly generate the population velocity, calculate the initial fitness, and thus calculate the initial nodes to obtain the global optimal value and the optimal population. Initialize the extreme value and extreme point, including global extreme value, individual extreme value initialization, global extreme point, and individual extreme point initialization. During the initialization and recording of the average fitness, in the implementation of the particle swarm optimization algorithm, in the iterative optimization process, as shown in Table 6, first, the iteration number loop and the population class loop need to be carried out. In the population class loop, first, the velocity is updated according to the ion swarm optimization formula, and then the velocity is restricted, that is, whether it exceeds the maximum speed and whether it is less than the minimum speed. If so, set it to the corresponding maximum and minimum speeds. After the velocity update, the population is updated according to the velocity and the coefficient, the population range is restricted, and then adaptive particle mutation is carried out. After obtaining the fitness value, individual optimal update and group optimal update are carried out, the average fitness and the best fitness are recorded, the best fitness curve is obtained, and finally the optimal value is assigned.

[0074] Table 6 pso-svm - Iterative Optimization Algorithm

[0075]

[0076]

[0077] Model establishment and evaluation. After obtaining the best parameters c and g, the input and output of the training set are copied to the PSO-SVM network to obtain the optimized best model. Then, simulation tests are carried out, that is, the training set input and the test set are input into the model to obtain the predicted values of the training set and the test set. Then, the class order is sorted, and the predicted values and the true values are compared to obtain the accuracies of the training set and the test set, which are 98.75% and 94.02% respectively. The visualization results are as Figure 3 、 Figure 4 shown. The red dots represent the true values, and the green dots represent the predicted values. When the two points coincide, it means the prediction is correct, and vice versa. Finally, the confusion matrices of the training set and the test set are obtained, as Figure 5 、 Figure 6 shown. Among them, 1, 2, 3, and 4 represent the four sample labels of porosity, shrinkage cavity, crack, and segregation respectively.

[0078] The above is a method for classifying internal defects of steel billets based on PSO-SVM provided by an embodiment of the present application. Based on the same inventive concept, an embodiment of the present application also provides a system for classifying internal defects of steel billets based on PSO-SVM. Figure 7 It is a composition diagram of a system for classifying internal defects of steel billets based on PSO-SVM provided by an embodiment of the present application. As Figure 7 shown, the system mainly includes: a data collection module 701, configured to collect ultrasonic signal data inside the steel billet and perform preprocessing; a prediction module 702, configured to input the preprocessed data into a pre-trained defect classification model for identification; wherein, the defect classification model includes a particle swarm algorithm and a support vector machine; a result output module 703, configured to output the identification result of the internal defect type of the steel billet.

[0079] The above is a system for classifying internal defects of steel billets based on PSO-SVM provided by an embodiment of the present application. Based on the same inventive concept, an embodiment of the present application also provides a device for classifying internal defects of steel billets based on PSO-SVM. Figure 8 It is a schematic diagram of a device for classifying internal defects of steel billets based on PSO-SVM provided by an embodiment of the present application. As Figure 8 shown, the device mainly includes: at least one processor 801; and a memory 802 communicatively connected to the at least one processor; wherein, the memory 802 stores instructions executable by the at least one processor 801, and the instructions are executed by the at least one processor 801 so that the at least one processor 801 can complete the foregoing method for classifying internal defects of steel billets based on PSO-SVM.

[0080] In addition, an embodiment of the present application also provides a non-volatile computer storage medium for classifying internal defects of steel billets based on PSO-SVM, storing computer-executable instructions, and the computer-executable instructions are executed by a processor to implement the foregoing method for classifying internal defects of steel billets based on PSO-SVM.

[0081] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in one Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0082] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction means that implements the functions specified in one or more of the processes and / or blocks Figure 1 of one or more of the processes and / or blocks Figure 1 specified in the block or blocks.

[0083] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one or more of the processes and / or blocks Figure 1 of one or more of the processes and / or blocks Figure 1 specified in the block or blocks.

[0084] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0085] Each embodiment in this application is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the apparatus embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and reference can be made to the partial description of the method embodiments for the relevant parts.

[0086] It should also be noted that the term "comprising", "including" or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process, method, commodity or device including a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the statement "including a..." does not exclude the presence of additional identical elements in the process, method, commodity or device including the element.

[0087] The above are only the embodiments of the present application and are not used to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A steel billet internal defect classification method based on PSO-SVM, characterized in that: The method comprises the following steps: Step S1: collecting ultrasonic signal data inside the billet and performing preprocessing; Step S2: inputting the preprocessed data into a pre-trained defect classification model for identification; wherein the defect classification model includes a particle swarm algorithm and a support vector machine; Step S3: Output the identification result of the internal defect type of the steel billet.

2. The method for classifying internal defects of steel billets based on PSO-SVM according to claim 1, characterized in that: The method further comprises: Select the linear kernel SVM kernel function according to the characteristics of the data; Initialize the parameters of the SVM model, including penalty parameters and kernel function parameters.

3. The method for classifying internal defects of steel billets based on PSO-SVM according to claim 1, characterized in that: The method further comprises: Optimize the model based on particle swarm algorithm; Select optimal parameters; Perform model evaluation.

4. The method for classifying internal defects of steel billets based on PSO-SVM according to claim 3 is characterized in that: The model is optimized based on the particle swarm algorithm, specifically: Set the particle swarm size N, where each particle represents a set of SVM parameter combinations; Use the training set to train each particle to obtain the SVM model, and use the test set to evaluate the classification accuracy of the model, and use the classification accuracy as the fitness value of the particle; According to the fitness value of the particle, update the speed and position of each particle to make the particle move towards a better solution; In order to improve the search efficiency, a crossover strategy can be introduced, that is, in each iteration, two particles are randomly selected and their position information is exchanged according to a preset probability to fuse the information between different particles; Check whether the stopping condition is met, such as reaching the maximum number of iterations or the fitness value no longer changes. If the condition is met, stop the iteration; otherwise, repeatedly update the speed and position of each particle and continue the iteration.

5. The method for classifying internal defects of steel billets based on PSO-SVM according to claim 3 is characterized in that: Select the optimal parameters, specifically: After the iteration, the particle with the highest fitness value is selected, and its corresponding parameter combination is the optimal parameter of the SVM model; The final SVM model was constructed using the optimal parameters and used to classify internal defects of steel billets.

6. The method for classifying internal defects of steel billets based on PSO-SVM according to claim 1, characterized in that: The identification results of the internal defect types of the steel billet include: internal cracks, pores, inclusions and no defects; the classification results are output in the form of probability.

7. The method for classifying internal defects of steel billets based on PSO-SVM according to claim 1, characterized in that: The ultrasonic signal data inside the billet is obtained through an ultrasonic flaw detector or X-ray imaging equipment, and the preprocessing includes: noise reduction, normalization and standardization operations.

8. A steel billet internal defect classification system based on PSO-SVM, characterized in that: The system comprises: A data collection module is used to collect ultrasonic signal data inside the steel billet and perform preprocessing; A prediction module, used to input the preprocessed data into a pre-trained defect classification model for identification; wherein the defect classification model includes a particle swarm algorithm and a support vector machine; The result output module is used to output the identification results of the internal defect types of the steel billet.

9. A steel billet internal defect classification device based on PSO-SVM, characterized in that: The device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can complete the steel billet internal defect classification method based on PSO-SVM as described in any one of claims 1-7.

Images