A Stacking-Based Multi-Model Fusion Method for Seamless Steel Pipe Production Monitoring and Anomaly Tracing

Abstract

This invention discloses a method for monitoring and tracing anomalies in seamless steel pipe production based on a Stacking multi-model fusion, belonging to the field of mechanical automation control technology. The method includes: acquiring production process data from historical production processes; selecting excellent sample data; preprocessing the excellent sample data and using a Stacking multi-model fusion dimensionality reduction algorithm to obtain the corresponding feature matrix; real-time acquisition of production process data during the current production process; preprocessing the real-time data and using a Stacking multi-model fusion dimensionality reduction algorithm to obtain the corresponding feature matrix; based on this, determining whether the current production process is abnormal, and performing anomaly tracing when anomalies occur. This invention enables online monitoring and anomaly tracing in seamless steel pipe production, improving production efficiency and product qualification rate.

Description

Technical Field

[0001] This invention relates to the field of mechanical automation control technology, and in particular to a method for monitoring and tracing anomalies in seamless steel pipe production based on Stacking fusion of multiple models. Background Technology

[0002] Seamless steel pipes are a commonly used special steel product, and the quality monitoring algorithms for each process of seamless steel pipes are gradually being valued by major seamless steel pipe manufacturers.

[0003] The continuous rolling process of seamless steel pipes has a significant impact on the dimensions and quality of the final product. However, the continuous rolling process is also highly complex, presenting various unresolved management challenges. In actual production, manual sampling and experience-based judgment are often used to monitor the quality and analyze anomalies of seamless steel pipes. A large amount of available production data is not effectively utilized, resulting in low quality control efficiency and most quality anomalies not being effectively addressed. Therefore, there is an urgent need for an automated intelligent monitoring system that utilizes production data to achieve real-time monitoring, alerts, and analysis of specific processes, thereby improving production efficiency and product quality. Summary of the Invention

[0004] This invention provides a method for monitoring and tracing anomalies in the production of seamless steel pipes based on Stacking and multi-model fusion, in order to solve the technical problem of low quality control efficiency in traditional methods that rely on manual sampling and experience-based judgment to monitor and analyze the quality of seamless steel pipes.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: On the one hand, this invention provides a method for monitoring and tracing anomalies in seamless steel pipe production based on Stacking fusion of multiple models, the method comprising: Obtain production process data from the historical continuous rolling production of seamless steel pipes and record it as historical data; Using the production process data corresponding to each seamless steel pipe as a sample data, and using preset quality evaluation indicators, sample data that meets the preset requirements are selected from the historical data and recorded as excellent sample data. The excellent sample data is preprocessed using a preset preprocessing algorithm to obtain the original matrix corresponding to the excellent sample data, and the original matrix corresponding to the excellent sample data is reduced in dimension using a dimensionality reduction algorithm based on Stacking fusion of multiple models to obtain the feature matrix corresponding to the excellent sample data. Real-time data is collected on the production process of seamless steel pipes during the current continuous rolling production process and recorded as real-time data. The real-time data is preprocessed using a preset preprocessing algorithm to obtain the original matrix corresponding to the real-time data, and the original matrix corresponding to the real-time data is reduced in dimensionality using a dimensionality reduction algorithm based on Stacking fusion of multiple models to obtain the feature matrix corresponding to the real-time data. Based on the feature matrix corresponding to the real-time data and the feature matrix corresponding to the excellent sample data, it is determined whether the current continuous rolling production process of seamless steel pipe is abnormal, and abnormality tracing is performed when an abnormality occurs.

[0006] Furthermore, the production process data includes: the rolling force, speed, torque and current of each roll of the six stands of the continuous rolling mill, the mill inlet temperature and the mandrel exit linear velocity.

[0007] Furthermore, the quality evaluation indicators include: the percentage of lengths with out-of-tolerance wall thickness at both ends. Maximum standard tolerance percentage of wall thickness after sawing Maximum eccentricity of wall thickness after sawing Percentage of head and tail outer diameters exceeding tolerance and the percentage of the maximum standard tolerance of the outer diameter after sawing. ;in,

[0008] in, The length of the head wall thickness exceeding the tolerance; The length of the tail section with excessive wall thickness; This is the total length of the seamless steel pipe;

[0009] in, This indicates the wall thickness at all sampling points after sawing; Indicates the target wall thickness value; This represents the percentage factor indicating the maximum allowable wall thickness error according to production requirements.

[0010] in, This represents the maximum wall thickness on the sampling unit circumference; This represents the minimum wall thickness on the circumference of the sampling unit;

[0011] in, Indicates the out-of-tolerance length of the head's outer diameter; Indicates the length of the tail's outer diameter that is out of tolerance;

[0012] in, This represents the outer diameter of all sampling points after sawing; Indicates the target outer diameter value; This represents the percentage factor indicating the maximum allowable error in the outer diameter according to production requirements.

[0013] Furthermore, the step of using preset quality evaluation indicators to select sample data that meets preset requirements from the historical data and recording them as excellent sample data includes: Based on the production process data corresponding to each seamless steel pipe, the quality evaluation index of each seamless steel pipe is calculated, and the production process data corresponding to seamless steel pipes whose quality evaluation index is within the preset range are used as excellent sample data.

[0014] Furthermore, the execution process of the preprocessing algorithm includes: For the production process data to be processed, the dynamic time warping method is used to transform the data into a standardized matrix that meets the requirements of statistical process control, ensuring that the lengths are uniform.

[0015] Furthermore, the execution process of the dimensionality reduction algorithm based on Stacking and fusion of multiple models includes: For the production process data to be dimensionality reduced, firstly, multiple different data processing algorithms are used to reduce the dimensionality of the data, resulting in various dimensionality reduction results. Then, Stacking ensemble learning technology is used to integrate and fuse the dimensionality reduction results of the multiple different data processing algorithms to achieve the dimensionality reduction of the production process data to be reduced.

[0016] Furthermore, the various data processing algorithms include: kernel principal component analysis algorithm, kernel partial least squares algorithm, and kernel entropy component analysis algorithm.

[0017] Furthermore, each data processing algorithm is optimized using the Adam algorithm.

[0018] Furthermore, based on the feature matrix corresponding to the real-time data and the feature matrix corresponding to the excellent sample data, it is determined whether the current continuous rolling production process of seamless steel pipe is abnormal, and anomaly tracing is performed when an anomaly occurs, including: The Procrustes analysis algorithm is used to calculate the dissimilarity between the feature matrix corresponding to the real-time data and the feature matrix corresponding to the excellent sample data. The calculated dissimilarity is compared with a preset dissimilarity threshold. When the calculated dissimilarity exceeds the preset dissimilarity threshold, the current continuous rolling production process is determined to be abnormal. When an anomaly is determined in the current continuous rolling production process, production process data of the seamless steel pipe continuous rolling production process in which the anomaly occurs are collected and recorded as abnormal data. Calculate the abnormal contribution rate of each variable in the abnormal data, and take the single variable or multiple variables with the largest abnormal contribution rate as the cause of the current abnormality in the continuous rolling production process of seamless steel pipe, so as to realize the abnormality tracing.

[0019] Furthermore, the formula for calculating the outlier contribution rate of each variable in the outlier data is as follows:

[0020] in, Indicates the first abnormal data j Abnormal contribution rate of each variable; Indicates the first number corresponding to the abnormal data i Principal components of the characteristic matrix; T Represents the transpose of a matrix; N This indicates the number of principal components in the feature matrix corresponding to the outlier data; Indicates the first abnormal data j Standardized data for each variable; Indicates the first number corresponding to the abnormal data i Information entropy of principal components of the feature matrix.

[0021] In another aspect, the present invention also provides an electronic device comprising a processor and a memory; wherein the memory stores at least one instruction, which is loaded and executed by the processor to implement the above-described method.

[0022] In another aspect, the present invention also provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the above method.

[0023] The beneficial effects of the technical solution provided by this invention include at least the following: Compared to traditional methods of monitoring and analyzing the quality of seamless steel pipes based on manual experience, the production data statistical process control method of this invention establishes dimensional judgment indicators and utilizes Stacking multi-model fusion, Procrustes analysis algorithm, and kernel space anomaly contribution algorithm to achieve real-time monitoring, real-time alarm, real-time analysis, and timely traceability of anomalies in the continuous rolling production process of seamless steel pipes. This solves the problem of low efficiency in traditional continuous rolling process monitoring and anomaly analysis. Experiments on multiple real-world cases have demonstrated the algorithm's practical effectiveness and applicability in the field, improving the production efficiency and product qualification rate of seamless steel pipes. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1This is a schematic diagram of the execution flow of the seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models provided in this embodiment of the invention; Figure 2 This is a system block diagram of the electronic device provided in the embodiments of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0027] First, it should be noted that in the embodiments of the present invention, the words "exemplarily," "for example," etc., are used to indicate that they are examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the term "exemplarily" is intended to present the concept in a specific manner. Furthermore, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either one or the other.

[0028] First Embodiment

[0029] This embodiment addresses various challenges in quality control during seamless steel pipe production and establishes a status monitoring and anomaly tracing method for the continuous rolling process. Based on statistical process control, it provides a method for seamless steel pipe production monitoring and anomaly tracing based on Stacking and multi-model fusion. This method can be implemented using electronic equipment, such as a terminal or server. The execution flow of this method is as follows: Figure 1 As shown, it includes the following steps: S1, acquire production process data from the historical continuous rolling production process of seamless steel pipes, and record it as historical data; In this embodiment, the production process data selected for monitoring includes effective variables such as process variables of each roll in the six stands of the continuous rolling mill, mill inlet temperature, and mandrel state variables, while redundant or inefficient variables have been removed. Specifically, the production process data selected for monitoring in this embodiment are: rolling force, speed, torque, and current of each roll in the six stands of the continuous rolling mill, mill inlet temperature, and mandrel exit linear velocity.

[0030] S2, using the production process data corresponding to each steel pipe as a sample data, and using preset quality evaluation indicators, selecting sample data that meets the preset requirements from the historical data, and recording them as excellent sample data; In this embodiment, quality evaluation indicators are designed based on actual on-site production conditions, including: the percentage of lengths with out-of-tolerance wall thickness at the beginning and end. Maximum standard tolerance percentage of wall thickness after sawing Maximum eccentricity of wall thickness after sawing Percentage of head and tail outer diameters exceeding tolerance and the percentage of the maximum standard tolerance of the outer diameter after sawing. ;in,

[0031] in, The length of the head wall thickness exceeding the tolerance; The length of the tail section with excessive wall thickness; The total length of the seamless steel pipe; the length exceeding the wall thickness tolerance is determined by the wall thickness requirements of the production process.

[0032]

[0033] The sawing length is collected from the production record sheet; This indicates the wall thickness at all sampling points after sawing; Indicates the target wall thickness value; This represents the percentage factor indicating the maximum allowable wall thickness error for production requirements, and is typically set to 0.1.

[0034]

[0035] in, This represents the maximum wall thickness on the sampling unit circumference; The eccentricity after sawing represents the wall thickness at the minimum value on the circumference of the sampling unit; the eccentricity after sawing represents the wall thickness unevenness and eccentricity. A larger value indicates that the wall thickness distribution is uneven in some parts of the steel pipe and that the center of the circular hole is eccentric. The eccentricity after sawing can measure the wall thickness uniformity of the finished seamless steel pipe.

[0036]

[0037] in, Indicates the out-of-tolerance length of the head's outer diameter; Indicates the length of the tail's outer diameter that is out of tolerance; This refers to the total length of the seamless steel pipe; the length exceeding the outer diameter tolerance is determined by the outer diameter requirements of the production process.

[0038]

[0039] in, This represents the outer diameter of all sampling points after sawing; Indicates the target outer diameter value; This represents the percentage factor indicating the maximum allowable error in the outer diameter according to production requirements, and is typically set between 0.005 and 0.01.

[0040] Furthermore, the step of using preset quality evaluation indicators to select sample data that meet preset requirements from the historical data and recording them as excellent sample data specifically involves: calculating the quality evaluation indicators for each seamless steel pipe based on the production process data corresponding to each seamless steel pipe, and using the production process data corresponding to seamless steel pipes with quality evaluation indicators within the preset range as excellent sample data, thereby constructing an excellent sample library.

[0041] S3, a preset preprocessing algorithm is used to preprocess the excellent sample data to obtain the original matrix corresponding to the excellent sample data, and a dimensionality reduction algorithm based on Stacking fusion of multiple models is used to reduce the dimensionality of the original matrix corresponding to the excellent sample data to obtain the feature matrix corresponding to the excellent sample data. Furthermore, the execution process of the preprocessing algorithm includes: preprocessing the original data with varying lengths and time misalignments, and using dynamic time warping to transform the data into a normalized matrix that satisfies statistical process control, ensuring that their lengths are uniform. The specific calculation formula is as follows: (1) For the collected data array X( n × m The data standardization process is as follows:

[0042]

[0043]

[0044] (2) Let the two time series whose similarity is to be calculated be X and Y, with lengths respectively. | X | and | Y | The resolving path takes the form W. =w 1 ,w 2 ,...,w K ,in Max(| X |,| Y |)<= K <=| X |+| Y | . w k The form is ( i , j ),in i It represents X. i coordinate, j This represents Y. j Coordinates. The resetting path W must start from... w Starting with 1=(1,1), and ending withw K =(| X |,| Y |) At the end, in W w ( i,j )of i and j It must be monotonically increasing:

[0045] Find the shortest normalized path of the sequence trajectory, with a distance of D(|X|,|Y|):

[0046] Furthermore, the execution process of the dimensionality reduction algorithm based on Stacking and fusion of multiple models is as follows: for the new regularized data, kernel principal component analysis, kernel partial least squares method, and kernel entropy component analysis are used to process it respectively, and then the data dimensionality reduction is achieved by integrating and fusing the above multi-layer models based on Stacking.

[0047] (1) The process of kernel principal component analysis (KPCA) is as follows: First, the nonlinear mapping of data is defined as... θ:R N →F , R N F represents the sample space, and F represents the high-dimensional mapping space. The two are connected by a nonlinear mapping function. θ Transformation, then F The sample covariance matrix C in space is represented as:

[0048] Introducing the kernel function, and based on its theory, we can obtain the kernel matrix K, whose elements are expressed as follows:

[0049] This scheme uses the Gaussian kernel function as the kernel matrix mapping function. The Gaussian kernel function mapping formula is as follows:

[0050] In the formula, K (x i , x j ) For the kernel matrix i Line 1 j Column elements, x i For the original data matrix, the first... i Array of column variables xj For the original data matrix, the first... j Array of column variables σ These are the coefficients of the Gaussian kernel function, and the output is a kernel space data matrix.

[0051] (2) The kernel partial least squares (KPLS) algorithm process is as follows: Define the nonlinear mapping function It can take samples x i (1,2,…, N Mapping from the original variable space to the feature space F ,Right now:

[0052] In the formula, the characteristic space F dimensionality m A matrix X, which can be arbitrarily large or even infinite in dimension, is mapped to a matrix F in the characteristic space:

[0053] KPLS requires that To perform zero-mean processing, first calculate... ( x i (average) :

[0054] In the formula, I N =[1 1 … 1] T ∈R N Then calculate the zero mean. :

[0055]

[0056] Therefore, nonlinear data (X, Y) is mapped to linear data ( Then, the PLS algorithm is used to construct... KPLS model between Y and Y:

[0057] In the formula, Y∈R N×γ This is called the score matrix, P∈R M×γ ,yes The load matrix, Q∈R 1×γ The load matrix of Y and Y r They are The residual matrix of Y, γ The number of latent variables.

[0058] (3) The nuclear entropy composition analysis (KECA) process is as follows: Using Renyi entropy as the information entropy function, the formula for quadratic Renyi entropy is as follows:

[0059] Substituting the kernel function into the quadratic Renyi entropy, the derivation is as follows:

[0060] In the formula, I yes N A column vector of size ×1 (all elements are 1), the kernel matrix K is decomposed into eigenvalues, and the eigenvalues ​​of each term in the formula are obtained. λ i and eigenvectors σ i Finally, the quadratic Renyi entropy of each dimension of the kernel matrix is ​​obtained. Dimensions with extremely low information entropy contain little information and are mostly composed of useless information and noise, so they are discarded to obtain the final kernel entropy principal component matrix.

[0061] (4) Establish a prediction model based on the Stacking algorithm and a multi-layer model to reduce the dimensionality of the continuous rolling production process data. The training method for each model in the Stacking framework is uniformly optimized using the Adam algorithm, which dynamically adjusts the learning rate of each parameter in the basic model using the first and second moment estimates of the gradient. The process of Adam algorithm optimizing the basic model is as follows:

[0062]

[0063]

[0064]

[0065]

[0066]

[0067] in, For gradient, m For the sample size, L Let be the objective function. θ The initial weights for the basic model, s and r gradients The first-order moment estimate and the second-order moment estimate, and These are the bias corrections for the first-order moment estimate and the second-order moment estimate, respectively. The first-order momentum decay coefficient, It is the second-order momentum decay coefficient. δ For the smoothing term, Δ θ for θ The change in quantity.

[0068] The formula for calculating the loss function L is as follows:

[0069] In the formula, N It is the number of training samples. and These are the predicted value and the actual value.

[0070] This model, trained through backpropagation, adjusts the weights and bias vectors of individual models, ultimately yielding a comprehensive dimensionality reduction matrix. This model can then be used to achieve dimensionality reduction of data.

[0071] S4 collects production process data in real time during the current continuous rolling production of seamless steel pipes and records it as real-time data; S5, the real-time data is preprocessed using a preset preprocessing algorithm to obtain the original matrix corresponding to the real-time data, and the original matrix corresponding to the real-time data is reduced in dimension using a dimensionality reduction algorithm based on Stacking fusion of multiple models to obtain the feature matrix corresponding to the real-time data. S6. Based on the feature matrix corresponding to the real-time data and the feature matrix corresponding to the excellent sample data, determine whether the current continuous rolling production process of the seamless steel pipe is abnormal, and perform abnormal tracing when an abnormality occurs. Specifically, based on the feature matrix corresponding to real-time data and the feature matrix corresponding to excellent sample data, it is determined whether the current continuous rolling production process of seamless steel pipe is abnormal, and anomaly tracing is performed when anomalies occur, including: S61, using the Procrustes analysis algorithm, calculates the dissimilarity between the feature matrix corresponding to the real-time data and the feature matrix corresponding to the excellent sample data, compares the calculated dissimilarity with the preset dissimilarity threshold, and determines that the current continuous rolling production process is abnormal when the calculated dissimilarity exceeds the preset dissimilarity threshold. The specific method for calculating dissimilarity is as follows: Taking two different batches of data matrices as an example, where X and Y are two different data matrices, the dissimilarity formula is derived as follows: First, move the centers of the two matrices to the origin:

[0072]

[0073] In the formula, Let X be the centered matrix. This is the centered matrix of Y.

[0074] Then, standardization is used to make the two matrices the same size, that is, to standardize their dissimilarity to 1:

[0075]

[0076] In the formula, Let X be the standardized matrix. Let Y be the standardized matrix.

[0077] Then find a rotation matrix R such that as much as possible with Alignment. This rotation matrix is ​​calculated using singular value decomposition (SVD):

[0078]

[0079] In the formula, U and V are orthogonal matrices, Σ is a diagonal matrix, and R is a rotation matrix.

[0080] Finally, calculate the dissimilarity between the aligned matrices:

[0081] In the formula, d (·) represents the dissimilarity between matrices, and its range is [0,1]. The larger the dissimilarity index, the greater the difference between the two original data matrices X and Y.

[0082] When there are multiple excellent samples, the average dissimilarity between the feature matrix corresponding to the real-time data and the feature matrix corresponding to each excellent sample data can be used as the final dissimilarity value to achieve production process monitoring.

[0083] S62, when the current continuous rolling production process is determined to be abnormal, the production process data of the seamless steel pipe continuous rolling production process in which the abnormality occurs is collected and recorded as abnormal data; S63, based on the effective production data in each production process, calculates the abnormal contribution rate of each variable in the abnormal data using the abnormal contribution rate method, and takes the single variable or multiple variables with the largest abnormal contribution rate as the cause of the current seamless steel pipe continuous rolling production process abnormality, so as to realize abnormality tracing. The formula for calculating the outlier contribution rate of each variable in the outlier data is as follows:

[0084] in, Indicates the first j The abnormal contribution rate of each monitoring parameter; Indicates the first i Item stacking fusion matrix principal components; T Represents the transpose of a matrix; N Indicates the number of principal components in the Stacking fusion matrix; Indicates the first j Standardized data for each monitoring parameter; Indicates the first i Stacking fusion matrix principal component information entropy.

[0085] In summary, this embodiment provides a method for monitoring and tracing anomalies in seamless steel pipe production based on a Stacking fusion model. Utilizing statistical process control, combined with a corresponding Stacking fusion model and anomaly contribution rate algorithm, it achieves process monitoring and anomaly tracing in the continuous rolling process of steel pipes, ultimately ensuring normal production of seamless steel pipes and achieving the goal of improving yield. This data-driven method for monitoring and tracing anomalies in seamless steel pipe production is efficient and convenient, based on a large amount of actual production data, and can quickly adjust parameters, training sets, and models according to actual working conditions, making it adaptable to production scenarios of various working conditions and steel grades.

[0086] Second Embodiment

[0087] This embodiment illustrates the implementation process of the seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models provided by the present invention, in conjunction with a specific application scenario. It includes the following steps: S1, select effective variables for monitoring during continuous rolling production.

[0088] This case study uses the production data of 600 seamless steel pipes of the same specification and heat number from a certain seamless steel pipe plant on a certain day, including the continuous rolling entry temperature. T 0, six-stand three-roll rolling force F 1-1 ~ F 6-1 , F 1-2 ~ F 6-2 , F 1-3 ~ F 6-3 ,speed v 1-1 ~ v 6-1 , v 1-2 ~ v 6-2 , v 1-3 ~ v 6-3 Torque M 1-1 ~ M 6-1 , M 1-2 ~ M 6-2 , M 1-3 ~ M 6-3 Current I 1-1 ~ I 6-1 , I 1-2 ~ I 6-2 , I 1-3 ~ I 6-3 Export line velocity v s1 - v s6 As shown in Table 1: Table 1. Parameters of each sample from the field

[0089] S2, based on the data obtained from S1, calculates quality evaluation indicators according to the actual production data samples on site, selects samples with excellent quality, and constructs an excellent sample database, as shown in Table 2. Table 2 Quality Evaluation Indicators

[0090] S3: Preprocess some of the raw data for the parameters used; Dynamic time warping is used to transform data into a normalized matrix that meets the requirements of statistical process control. Linear interpolation is employed to interpolate and sample time series of parameters of varying lengths, unifying the sampling length of each parameter without compromising its original characteristics. This achieves the physical allocation of sampling points for each parameter.

[0091] S4: Data dimensionality reduction is achieved through data analysis based on Stacking integrated multi-layer models. The dissimilarity of the dimensionality reduction matrix is ​​calculated using the Procrustes analysis algorithm to enable process monitoring of continuous rolling production.

[0092] In this embodiment, according to the production failure record table, it is known that the 448th steel pipe had a serious rolling problem, and the entire steel pipe cracked during continuous rolling, ultimately being scrapped. Experiments revealed a production anomaly at the 348th steel pipe, as shown in Table 3. Since the first 100 steel pipes were used as the training set, adding them to the training set resulted in the 448th steel pipe, thus proving the effectiveness of the production anomaly localization. Retrieving its production data and comparing it with normal steel pipes revealed obvious anomaly characteristics.

[0093] Table 3. Anomaly Dissimilarity Values

[0094] S5: Utilize effective production data from each production process and analyze monitoring parameters using the anomaly contribution rate method. Calculate and analyze the single or several parameters with the greatest impact on the anomaly to achieve anomaly tracing. Specifically, in this embodiment, based on step S4, process monitoring detects anomalies, then extracts anomaly sample data and performs anomaly contribution rate analysis. The results are shown in Table 4. Table 4 Analysis of Abnormal Contribution Rate

[0095] As shown in Table 4, the abnormal contribution rate of the mill torque is the largest when corresponding to the process parameter number. The torque reflects the combined effect of rolling force and speed. Based on historical data, after adjusting the mill speed setting in subsequent work, production returned to normal.

[0096] In summary, this case demonstrates that the seamless steel pipe continuous rolling production process detection and anomaly tracing method based on the Stacking integrated multi-layer model provided by this invention can solve the problem of low efficiency in traditional continuous rolling process monitoring and anomaly analysis methods. It can realize online monitoring and anomaly tracing of the seamless steel pipe production process, which is of great significance for improving the production efficiency and product qualification rate of seamless steel pipes.

[0097] Third Embodiment

[0098] This embodiment provides an electronic device, such as... Figure 2 As shown, the electronic device includes a processor and a memory; wherein the processor and the memory can be connected via a communication bus; the memory stores at least one instruction, which is loaded and executed by the processor to implement the method of the first embodiment described above. Furthermore, the electronic device may also include a transceiver, the processor and the transceiver can be connected via a communication bus, and the transceiver is used to communicate with other devices.

[0099] Below, in conjunction with Figure 2 A detailed introduction to each component of this electronic device is provided below: The processor is the control center of the electronic device. The electronic device may include multiple processors, each of which can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The term "processor" can refer to a single processor or a collective term for multiple processing elements. For example, a processor can be one or more central processing units (CPUs), other general-purpose processors, application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), one or more field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor can perform various functions of the electronic device by running or executing software programs stored in memory and by calling data stored in memory.

[0100] In a specific implementation, as one example, the processor may include one or more CPUs, for example... Figure 2 CPU0 and CPU1 shown are, of course, merely illustrative examples.

[0101] The memory is used to store the software program that executes the solution of the present invention, and the processor controls its execution. For specific implementation methods, please refer to the above method embodiments, which will not be repeated here.

[0102] Optionally, the memory may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory may be integrated with the processor or exist independently, and may be accessed through the interface circuit of the electronic device ( Figure 2 (Not shown in the image) is coupled to the processor; however, this embodiment of the invention does not impose specific limitations on this.

[0103] The transceiver may include a receiver and a transmitter. Figure 2 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function. The transceiver can be integrated with the processor or exist independently, and can be connected through the interface circuit of the electronic device (…). Figure 2 (Not shown in the image) is coupled to the processor, and this embodiment of the invention does not specifically limit this.

[0104] In addition, it should be noted that, Figure 2 The structure of the electronic device shown is not intended to limit the device. Actual devices may include more or fewer components than shown, or combine certain components, or have different component arrangements. Furthermore, the technical effects achieved by this electronic device when performing the method of the first embodiment described above can be referenced to the technical effects described in the first embodiment; therefore, they will not be repeated here.

[0105] Fourth embodiment

[0106] This embodiment provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the method of the first embodiment described above. The computer-readable storage medium may be a ROM, random access memory, CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc. The instruction stored therein can be loaded and executed by a processor in a terminal.

[0107] Furthermore, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, embodiments of the present invention can take the form of a completely or partially hardware embodiment, a completely or partially software embodiment, or an embodiment combining software and hardware aspects. Moreover, when implemented in software, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any usable medium accessible to a computer or a data storage device such as a server or data center containing one or more sets of usable media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive (SSD).

[0108] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0109] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element. Furthermore, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Additionally, the character " / " in this text generally indicates an "or" relationship between the preceding and following objects, but it can also indicate an "AND / OR" relationship. Please refer to the context for specific interpretations. "At least one" refers to one or more items, while "more than" refers to two or more items. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can be represented as: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0111] Furthermore, it is understood that in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0112] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0113] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of functional modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings or direct couplings or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms. Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs. Additionally, the functional units in the various embodiments of this invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0114] If the method is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0115] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments of the present invention have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make several improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A method for monitoring and tracing anomalies in seamless steel pipe production based on Stacking multi-model fusion, characterized in that, The method for monitoring and tracing anomalies in seamless steel pipe production based on Stacking fusion of multiple models includes: Obtain production process data from the historical continuous rolling production of seamless steel pipes and record it as historical data; Using the production process data corresponding to each seamless steel pipe as a sample data, and using preset quality evaluation indicators, sample data that meets the preset requirements are selected from the historical data and recorded as excellent sample data. The excellent sample data is preprocessed using a preset preprocessing algorithm to obtain the original matrix corresponding to the excellent sample data, and the original matrix corresponding to the excellent sample data is reduced in dimension using a dimensionality reduction algorithm based on Stacking fusion of multiple models to obtain the feature matrix corresponding to the excellent sample data. Real-time data is collected on the production process of seamless steel pipes during the current continuous rolling production process and recorded as real-time data. The real-time data is preprocessed using a preset preprocessing algorithm to obtain the original matrix corresponding to the real-time data, and the original matrix corresponding to the real-time data is reduced in dimensionality using a dimensionality reduction algorithm based on Stacking fusion of multiple models to obtain the feature matrix corresponding to the real-time data. Based on the feature matrix corresponding to the real-time data and the feature matrix corresponding to the excellent sample data, it is determined whether the current continuous rolling production process of seamless steel pipe is abnormal, and abnormality tracing is performed when an abnormality occurs. Based on the feature matrix corresponding to the real-time data and the feature matrix corresponding to the excellent sample data, it is determined whether the current continuous rolling production process of seamless steel pipe is abnormal, and anomaly tracing is performed when an anomaly occurs, including: The Procrustes analysis algorithm is used to calculate the dissimilarity between the feature matrix corresponding to the real-time data and the feature matrix corresponding to the excellent sample data. The calculated dissimilarity is compared with a preset dissimilarity threshold. When the calculated dissimilarity exceeds the preset dissimilarity threshold, the current continuous rolling production process is determined to be abnormal. When an anomaly is determined in the current continuous rolling production process, production process data of the seamless steel pipe continuous rolling production process in which the anomaly occurs are collected and recorded as abnormal data. Calculate the abnormal contribution rate of each variable in the abnormal data, and take the single variable or multiple variables with the largest abnormal contribution rate as the cause of the current abnormality in the continuous rolling production process of seamless steel pipe, so as to realize the abnormality tracing. The formula for calculating the outlier contribution rate of each variable in the outlier data is: ； in, Indicates the first abnormal data j Abnormal contribution rate of each variable; Indicates the first number corresponding to the abnormal data i Principal components of the characteristic matrix; T Represents the transpose of a matrix; N This indicates the number of principal components in the feature matrix corresponding to the outlier data; Indicates the first abnormal data j Standardized data for each variable; Indicates the first number corresponding to the abnormal data i Information entropy of principal components of the feature matrix.

2. The seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models as described in claim 1, characterized in that, The production process data includes: rolling force, speed, torque and current of each roll of the six stands of the continuous rolling mill, mill inlet temperature and mandrel exit linear velocity.

3. The seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models as described in claim 1, characterized in that, The quality evaluation indicators include: the percentage of lengths with out-of-tolerance wall thickness at the head and tail. Maximum standard tolerance percentage of wall thickness after sawing Maximum eccentricity of wall thickness after sawing Percentage of head and tail outer diameters exceeding tolerance and the percentage of the maximum standard tolerance of the outer diameter after sawing. ;in, ； in, The length of the head wall thickness exceeding the tolerance; The length of the tail section with excessive wall thickness; This is the total length of the seamless steel pipe; ； in, This indicates the wall thickness at all sampling points after sawing; Indicates the target wall thickness value; This represents the percentage factor indicating the maximum allowable wall thickness error according to production requirements. ； in, This represents the maximum wall thickness on the sampling unit circumference; This represents the minimum wall thickness on the circumference of the sampling unit; ； in, Indicates the out-of-tolerance length of the head's outer diameter; Indicates the length of the tail's outer diameter that is out of tolerance; ； in, This represents the outer diameter of all sampling points after sawing; Indicates the target outer diameter value; This represents the percentage factor indicating the maximum allowable error in the outer diameter according to production requirements.

4. The seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models as described in claim 3, characterized in that, The step of using preset quality evaluation indicators to select sample data that meets preset requirements from the historical data and recording it as excellent sample data includes: Based on the production process data corresponding to each seamless steel pipe, the quality evaluation index of each seamless steel pipe is calculated, and the production process data corresponding to seamless steel pipes whose quality evaluation index is within the preset range are used as excellent sample data.

5. The seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models as described in claim 1, characterized in that, The execution process of the preprocessing algorithm includes: For the production process data to be processed, the dynamic time warping method is used to transform the data into a standardized matrix that meets the requirements of statistical process control, ensuring that the lengths are uniform.

6. The seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models as described in claim 1, characterized in that, The execution process of the dimensionality reduction algorithm based on Stacking and fusion of multiple models includes: For the production process data to be dimensionality reduced, firstly, multiple different data processing algorithms are used to reduce the dimensionality of the data, resulting in various dimensionality reduction results. Then, Stacking ensemble learning technology is used to integrate and fuse the dimensionality reduction results of the multiple different data processing algorithms to achieve the dimensionality reduction of the production process data to be reduced.

7. The seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models as described in claim 6, characterized in that, The various data processing algorithms include: kernel principal component analysis algorithm, kernel partial least squares algorithm, and kernel entropy component analysis algorithm.

8. The seamless steel pipe production monitoring and anomaly tracing method based on Stacking fusion of multiple models as described in claim 6, characterized in that, Each data processing algorithm is optimized using the Adam algorithm.

Images