A pass category monitoring method for radial forging machine

Abstract

The present invention discloses a method for monitoring the pass categories of a radial forging machine, and relates to the field of radial forging technology. An objective function for clustering is obtained; the objective function is constructed based on the kernel space distance function between a data point and the cluster center of the pass category, and the membership function of the data point to the pass category; feature extraction is performed on the forging process sample data and the equipment status sample data at each time point, and the features of multiple data points are determined; the cluster center of each pass category is initialized; with the objective of minimizing the objective function value, the cluster center of each pass category is iteratively updated until the iteration meets the convergence condition, and the cluster center that meets the convergence condition is determined as the target cluster center of each pass category; the pass categories of the radial forging machine are monitored in real time through the target cluster center. This method can accurately monitor the pass categories of the radial forging machine.

Description

Technical Field

[0001] The present invention relates to the technical field of radial forging, and in particular to a pass category monitoring method for a radial forging machine. Background Art

[0002] Currently, radial forging often requires multiple passes to process the workpiece from its initial state to the final desired shape and size, describing the development of the workpiece geometry from its initial state to its final state.

[0003] For example, when forging a rough shaft into a precision shaft part, it may take four to five passes, each with specific deformation and process parameters, gradually changing the workpiece's diameter, length, and other dimensions while also improving its internal structure and properties. Therefore, monitoring the pass type is extremely necessary.

[0004] However, the prior art lacks a method for accurately monitoring the pass category of a radial forging machine. Summary of the Invention

[0005] Based on this, it is necessary to provide a pass category monitoring method for a radial forging machine in response to the above technical problems. This method can accurately monitor the pass category of the radial forging machine.

[0006] The present invention adopts the following technical solutions:

[0007] The present invention provides a pass category monitoring method for a radial forging machine, comprising:

[0008] Obtaining an objective function for clustering; the objective function is constructed based on a kernel space distance function between a data point and a cluster center of a pass category, and a membership function of the data point to the pass category; the kernel space distance function includes a kernel function of the data point itself, a kernel function of the cluster center itself, and a kernel function between the data point and the cluster center; the kernel function includes a univariate Gaussian kernel term for representing the similarity between two univariate data points, and a bivariate interaction kernel term for representing the interaction effect between the two variables;

[0009] Acquire forging process sample data and equipment status sample data of the radial forging machine at multiple time points, and perform feature extraction on the forging process sample data and equipment status sample data at each time point to determine the features of the multiple data points;

[0010] Initialize the cluster center of each pass category;

[0011] With the goal of minimizing the objective function value, the cluster center of each pass category is iteratively updated according to the characteristics of multiple data points until the iteration meets the convergence condition. The cluster center that meets the convergence condition is determined as the target cluster center of each pass category.

[0012] The target features of the forging process data and equipment status data of the radial forging machine at the current moment are extracted, and the membership of the target features to the target cluster center of each pass category is calculated. The pass category corresponding to the maximum membership is determined as the pass category of the radial forging machine at the current moment.

[0013] Optionally, the kernel function is:

[0014] ;

[0015] in, For data points and data points The kernel function between are the univariate Gaussian kernel term and the bivariate interaction term, For data points and data points The characteristic dimension of is the feature weight, , is the nuclear sensitivity factor, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, represents the L2 norm, , Indicates the data point The characteristics of the dimension and the The nonlinear interaction effect weights between the features of the dimension, Indicates the data point The characteristics of the dimension and the The sensitivity factor of the interaction term between the features of the dimension, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics.

[0016] Optionally, data points Cluster centers of the pass categories The kernel space distance function between is:

[0017] ;

[0018] in, is the cluster center of the lane category, Represents data points Its own kernel function, is the cluster center Its own kernel function, For data points and cluster centers The kernel function between

[0019] No. data points For the first Membership function of the pass category for:

[0020] ;

[0021] in, is the number of pass categories, For the data points With the The cluster center of each pass category The kernel function between For the data points With the The cluster center of each pass category The kernel function between for The reciprocal of

[0022] Objective function for:

[0023] ;

[0024] in, is the total number of data points, For the data points For the first The membership function of the pass category, For the data points Its own kernel function, For the The cluster center of each pass category Its own kernel function, For the data points With the The cluster center of each pass category The kernel function between is the first regularization coefficient, is the second regularization coefficient, is the feature weight, , for The L1 norm of for The L2 norm of is the fuzzy index.

[0025] Optionally, feature weights During the clustering process, the soft threshold algorithm is used for iterative updates; the update formula is:

[0026] ;

[0027] in, For the The feature weights of the iteration , For the The first iteration data points For the first The cluster center of each pass category The kernel space distance function between For the The feature weight of the iteration, For the The first iteration data points For the first The membership function of the pass category, represents the soft threshold operator, is the input of the soft threshold operator, is the threshold, for The L2 norm of .

[0028] Optionally, a momentum mechanism is introduced into the cluster center update process; the cluster center update formula is:

[0029] ;

[0030] in, For the The first iteration The cluster centers of the lane categories, For the The first iteration The cluster centers of the lane categories, is the momentum coefficient.

[0031] Optionally, the convergence condition is:

[0032] The objective function value is less than a first preset threshold, and the displacement of the cluster center of each pass category between the current iteration and the previous iteration is less than a second preset threshold.

[0033] Optionally, feature extraction is performed on the forging process sample data and the equipment status sample data at each time point to determine the features of multiple data points, including:

[0034] The 3-sigma principle is used to filter out abnormal data points that are beyond the range of forging process sample data and equipment status sample data, and the mean filling method is used to fill in the missing values after filtering out abnormal data points to obtain complete data points.

[0035] Perform denoising on the complete data points to obtain denoised data points;

[0036] Perform time domain feature extraction and frequency domain feature extraction on the forging process sample data and equipment status sample data of the denoised data points, respectively, to obtain multiple time domain features and multiple frequency domain features for each data point;

[0037] Calculate the importance scores of all features and select the K features with the largest importance scores as the features of each data point.

[0038] Optionally, target features of the forging process data and equipment status data of the radial forging machine at the current moment are extracted, including:

[0039] The key features of the forging process data and equipment status data of the radial forging machine at the current moment are extracted, and the key features are determined as target features.

[0040] The present invention provides a pass category monitoring device for a radial forging machine, comprising:

[0041] The first acquisition module is used to obtain the objective function for clustering; the objective function is constructed based on the kernel space distance function between the data point and the cluster center of the pass category, and the membership function of the data point to the pass category; the kernel space distance function includes the kernel function of the data point itself, the kernel function of the cluster center itself, and the kernel function between the data point and the cluster center; the kernel function includes a univariate Gaussian kernel term for representing the similarity between two univariate data points and a bivariate interaction kernel term for representing the interaction effect between the two variables;

[0042] a second acquisition module, configured to acquire forging process sample data and equipment status sample data of the radial forging machine at multiple time points, and perform feature extraction on the forging process sample data and equipment status sample data at each time point to determine features of the multiple data points;

[0043] Initialization module, used to initialize the cluster center of each pass category;

[0044] An updating module is used to iteratively update the cluster center of each pass category based on the characteristics of multiple data points with the goal of minimizing the objective function value until the iteration meets the convergence condition, and determine the cluster center that meets the convergence condition as the target cluster center of each pass category;

[0045] The monitoring module is used to extract the target features of the forging process data and equipment status data of the radial forging machine at the current moment, and calculate the membership of the target features to the target cluster center of each pass category respectively, and determine the pass category corresponding to the maximum membership as the pass category of the radial forging machine at the current moment.

[0046] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the pass category monitoring method of the radial forging machine is implemented.

[0047] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the pass category monitoring method for the radial forging machine is implemented.

[0048] At least one of the above technical solutions adopted by the present invention can achieve the following beneficial effects:

[0049] In the present invention, the contribution of each feature to the sample similarity can be captured through the univariate Gaussian kernel term, reflecting the characteristics of a single feature; the nonlinear interaction information between different features is mined based on the bivariate interaction kernel term. In this way, the two are combined to construct a kernel function, which comprehensively considers the features themselves and the interaction between features, so that the kernel function more accurately characterizes the similarity between data points, laying a good foundation for subsequent analysis. Furthermore, the kernel function of the data point itself, the kernel function of the cluster center itself, and the kernel function of the data point and the cluster center are integrated to construct a kernel space distance function, which comprehensively measures the relationship between the data point and the cluster center, and combines the membership to construct an objective function, taking into account the degree of belonging of the data points to different pass categories, so that the objective function can effectively reflect the degree of fit between the data and the pass category clustering, guiding the clustering process to be more accurate; and, when performing data feature extraction, the equipment status and process data of the radial forging machine are simultaneously collected, enriching the feature dimension, thereby learning the pass features from multiple angles, improving the ability to distinguish different pass categories, and thus improving the pass category monitoring accuracy of the radial forging machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0051] Figure 1 A schematic flow chart of a pass category monitoring method for a radial forging machine provided by the present invention;

[0052] Figure 2 A schematic diagram of the results of outlier cleaning of an analog signal provided by the present invention;

[0053] Figure 3 A schematic diagram of the result of filling missing items of a simulation signal provided by the present invention;

[0054] Figure 4 A schematic diagram of the noise removal result of an analog signal provided by the present invention;

[0055] Figure 5 A schematic diagram of the result of data scale unification of an analog signal provided by the present invention;

[0056] Figure 6 A schematic flow chart of another method for monitoring pass categories of a radial forging machine provided by the present invention;

[0057] Figure 7 A schematic diagram of a computer device for implementing a pass category monitoring method for a radial forging machine provided by the present invention. DETAILED DESCRIPTION

[0058] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0059] Currently, the dimensional accuracy and surface quality of rotating parts (such as hydraulic cylinders and shafts) during radial forging are highly dependent on precise control of forging pass parameters. Pass monitoring is crucial in radial forging and a key component of fully automated intelligent production lines. Accurate pass monitoring is a prerequisite for intelligent monitoring of production line status. Its purpose is to: prevent scrap caused by manual feeding errors in manual feeding scenarios; and in fully automated scenarios, dynamically switching between dedicated detection models through pass monitoring improves anomaly identification accuracy.

[0060] Traditional pass monitoring methods rely primarily on offline measurements or fixed rule-based judgments, making them difficult to adapt to the complex and changing dynamic characteristics of processes in high-variety, small-batch production models. This is especially true in precision forging scenarios, where material deformation is nonlinear and process parameters are highly coupled. Existing technologies have the following prominent issues:

[0061] (1) Offline prediction and real-time monitoring are disconnected: Traditional methods rely on static modeling of historical data and cannot capture dynamic changes in the forging process (such as sudden changes in material rheology and accumulation of mold thermal deformation), resulting in delayed parameter adjustment.

[0062] (2) Insufficient coordination of multi-source data: process parameters, equipment status (vibration, oil pressure) and part characteristics are scattered in independent systems. The existing threshold method or single model is difficult to characterize the multi-factor coupling effect, and the false alarm rate is as high as 15%-20%.

[0063] (3) Poor adaptability to small samples: In the multi-variety small batch mode, deep learning methods are prone to overfitting due to data sparsity, and traditional clustering algorithms such as K-means are sensitive to noise and have high feature redundancy.

[0064] (4) Computational efficiency bottleneck: Traditional machine learning methods have a processing delay of over 500ms, which makes it difficult to meet real-time monitoring needs.

[0065] Based on this, the present invention provides a pass category monitoring method for a radial forging machine, which aims to achieve real-time perception and adaptive adjustment of the forging process parameters of rotary parts through multi-source data fusion and improved sparse kernel fuzzy clustering, breaking through the existing technology's dependence on static rules and large amounts of data, and improving the monitoring accuracy and process stability under complex working conditions.

[0066] The execution subject of the method of the present invention can be a server set up on a business platform, or a device such as a desktop computer or a laptop computer that can execute the solution of the present invention. The technical solutions provided by various embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0067] Figure 1 The present invention is a flow chart of a method for monitoring the pass category of a radial forging machine, which specifically includes the following steps:

[0068] S101, obtaining the objective function for clustering; the objective function is constructed based on the kernel space distance function between the data point and the cluster center of the pass category, and the membership function of the data point to the pass category; the kernel space distance function includes the kernel function of the data point itself, the kernel function of the cluster center itself, and the kernel function between the data point and the cluster center; the kernel function includes a univariate Gaussian kernel term for representing the similarity between two univariate data points and a bivariate interaction kernel term for representing the interaction effect between the two variables.

[0069] The kernel function is a structured ANOVA kernel function, and the kernel function is:

[0070] (1);

[0071] in, For data points and data points The kernel function between are the univariate Gaussian kernel term and the bivariate interaction term, For data points and data points The characteristic dimension of is the feature weight, , filtering dominant classification features through sparse constraints; is the nuclear sensitivity factor, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, represents the L2 norm, , Indicates the data point The characteristics of the dimension and the The nonlinear interaction effect weights between the features of the dimension, Indicates the data point The characteristics of the dimension and the The sensitivity factor of the interaction term between the features of the dimension, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics.

[0072] in, , For data points and data points No. The standard deviation of the feature of dimension, , For all data points Dimension and The standard deviation of the interaction feature between dimensions.

[0073] in, , .

[0074] Optionally, data points Cluster centers of the pass categories The kernel space distance function between is:

[0075] (2);

[0076] in, is the cluster center of the lane category, Represents data points Its own kernel function, is the cluster center Its own kernel function, For data points and cluster centers The kernel function between .

[0077] According to the kernel space distance function, the kernel space weighted distance term is determined. The calculation form of the kernel space weighted distance term is:

[0078] (3);

[0079] in, Indicates the data points For the first The membership function of the pass category, is the fuzzy index, is the number of pass categories, is the total number of data points, Indicates the data points Its own kernel function, Indicates the The cluster center of each pass category Its own kernel function, Expressed as data points With the The cluster center of each pass category The kernel function between .

[0080] Optionally, data points For the first Membership function of the pass category for:

[0081] (4);

[0082] in, is the number of pass categories, For the data points With the The cluster center of each pass category The kernel function between For the data points With the The cluster center of each pass category The kernel function between for The reciprocal of .

[0083] The objective function is defined as the combination of the kernel space weighted distance term and the regularization term, that is, the objective function for:

[0084] (5);

[0085] in, is the total number of data points, For the data points For the first The membership function of the pass category satisfies ,and ; For the data points Its own kernel function, For the The cluster center of each pass category Its own kernel function, For the data points With the The cluster center of each pass category The kernel function between is the first regularization coefficient, is the second regularization coefficient, is the feature weight, , for The L1 norm of for The L2 norm of is the fuzzy index, the first regularization coefficient , enforce sparsity to remove redundant features; the second regularization coefficient , to prevent weight overfitting.

[0086] is a fuzzy index that controls the sharpness of the membership distribution (the larger the value, the fuzzier the classification boundaries).

[0087] S102 , obtaining forging process sample data and equipment status sample data of the radial forging machine at multiple time points, and performing feature extraction on the forging process sample data and equipment status sample data at each time point to determine features of multiple data points.

[0088] Optionally, feature extraction is performed on the forging process sample data and the equipment status sample data at each time point to determine the features of multiple data points, including: filtering out abnormal data points that exceed the range of the forging process sample data and the equipment status sample data using the 3 standard deviation principle, and using the mean filling method to fill in the missing values after filtering out the abnormal data points to obtain complete data points; denoising the complete data points to obtain denoised data points; performing time domain feature extraction and frequency domain feature extraction on the forging process sample data and the equipment status sample data of the denoised data points to obtain multiple time domain features and multiple frequency domain features of each data point; calculating the importance scores of all features, and selecting the one with the largest importance score K features as the features of each data point; optionally, multiple feature sets are obtained based on all the features, and each feature set is traversed to score the clustering index (mainly referring to the classification coefficient), and the set with the best clustering effect is selected. The feature corresponding to the best set is the one with the largest importance score. K Features.

[0089] Specifically, assuming the total number of features is W , you need to select K Features, generate A set of features, for example, if , , then the number of feature sets is .

[0090] Taking the silhouette coefficient as an example, for each subset, the characteristic data of the subset is used to train a clustering model (such as K-means). Based on the clustering results, the silhouette coefficient of each sample data in the subset is calculated, and the mean of the silhouette coefficients in the subset is calculated. The subset with the largest mean silhouette coefficient is determined as the one with the largest importance score. K Features.

[0091] Specifically, the forging process sample data includes forging frequency and pressing amount; the equipment status sample data includes vibration acceleration (0-20kHz), oil pressure fluctuation (0-100MPa), and mold temperature (20-800℃).

[0092] Outlier cleaning: Using the 3 standard deviation criterion, by calculating the mean ( ) and variance ( ), set the threshold , filter out abnormal data points that are out of range, such as Figure 2 As shown, Figure 2 This is a schematic diagram of the results of outlier cleaning of a simulation signal. Figure 2 Figure (a) is a schematic diagram of the signal before outlier cleaning. Figure 2 Figure (b) is a schematic diagram of the signal after outlier cleaning.

[0093] Missing item filling: Use the mean filling method to fill in missing values to ensure data integrity, such as Figure 3 As shown, Figure 3 This is a schematic diagram of the result of filling in the missing items of a simulation signal. Figure 3 Figure (a) is a schematic diagram of the signal before the missing items are filled. Figure 3 Figure (b) is a schematic diagram of the signal after the missing items are filled.

[0094] Noise removal: Apply Savitzky-Golay filter to smooth the signal with a window size of 7 and a polynomial order of 2, which can suppress high-frequency noise while retaining the signal characteristics. Figure 4 As shown, Figure 4 This is a schematic diagram of the noise removal result of an analog signal provided by the present invention. Figure 4 Figure (a) is a schematic diagram of the signal before noise removal. Figure 4 Figure (b) is a schematic diagram of the signal after noise removal.

[0095] Data scale unification: using normalization algorithm , eliminating dimensional differences, such as Figure 5 As shown, Figure 5 This is a schematic diagram of the result of data scale unification for an analog signal. Figure 5 Figure (a) is a schematic diagram of the signal before data scale unification. Figure 5 Figure (b) is a schematic diagram of the signal after data scale is unified.

[0096] The forging process sample data and equipment status sample data of the unified data points are subjected to time domain feature extraction and frequency domain feature extraction, respectively. The forging process sample data includes forging frequency and pressing amount; the equipment status sample data includes vibration acceleration, oil pressure fluctuation, and mold temperature.

[0097] Among them, time domain feature extraction includes: extracting parameters such as mean, standard deviation, maximum, minimum, root mean square, waveform factor (dimensional), as well as kurtosis and skewness (dimensionless).

[0098] The specific calculation method is: ; ; ; ; ; ; ; .in, is the number of data points of the time series data corresponding to a single sample data in the forging process sample data and the equipment status sample data, Indicates that the corresponding sample data is in The value of a time series.

[0099] Frequency domain feature extraction includes: extracting spectrum mean, spectrum maximum, spectrum energy, and spectrum entropy (dimensioned). The specific method is:

[0100] ; ; ; ,in, ;in, The corresponding sample data is The spectrum of a time series.

[0101] According to the above calculation method, the time domain features and frequency domain features of each time point are calculated respectively, and then the feature importance is evaluated based on the classification coefficient to screen the most important ones. K Key features (such as standard deviation, minimum value, root mean square, waveform factor and spectrum mean with a classification coefficient > 0.85 can be considered as key features).

[0102] Will get K The key features are used as the features of each time point. A time point can be regarded as a data point, that is, multiple features of each data point are obtained.

[0103] S103, initialize the cluster center of each pass category, and with the goal of minimizing the objective function value, iteratively update the cluster center of each pass category based on the characteristics of multiple data points until the iteration meets the convergence condition, and determine the cluster center that meets the convergence condition as the target cluster center of each pass category.

[0104] Among them, the pass categories of radial forging machines can include four categories, as shown in Table 1.

[0105] Table 1

[0106]

[0107] Optionally, feature weights During the clustering process, the soft threshold algorithm is used for iterative updates; the update formula is:

[0108] (6);

[0109] in, For the The feature weights of the iteration , For the The first iteration data points For the first The cluster center of each pass category The kernel space distance function between For the The feature weight of the iteration, For the The first iteration data points For the first The membership function of the pass category, represents the soft threshold operator, is the input of the soft threshold operator, is the threshold, for The L2 norm of . , Indicates in Iteration , Indicates in Iteration , Indicates in Iteration .

[0110] Determine the weight shrinkage ( ); : sparse parameters, is the feature dimension of the data point, controlling the number of retained features (such as while retaining about 4 dimensions).

[0111] Optionally, a momentum mechanism is introduced into the cluster center update process; the cluster center update formula is:

[0112] (7);

[0113] in, For the The first iteration The cluster centers of the lane categories, For the The first iteration The cluster centers of the lane categories, is the momentum coefficient, .

[0114] Optionally, the convergence condition is: the objective function value is less than a first preset threshold, and the displacement of the cluster center of each pass category between the current iteration and the previous iteration is less than a second preset threshold.

[0115] The first preset threshold and the second preset threshold can be set according to actual needs.

[0116] S104, extracting the target features of the forging process data and equipment status data of the radial forging machine at the current moment, and calculating the membership of the target features to the target cluster center of each pass category respectively, and determining the pass category corresponding to the maximum membership as the pass category of the radial forging machine at the current moment.

[0117] Optionally, extracting target features of the forging process data and equipment status data of the radial forging machine at the current moment includes: extracting key features of the forging process data and equipment status data of the radial forging machine at the current moment, and determining the key features as target features.

[0118] Based on the target features, the membership of the data points at the current moment to each pass category is calculated according to formula (4), and the pass category corresponding to the maximum membership is determined as the pass category of the radial forging machine at the current moment.

[0119] In one embodiment, the present invention also provides a method for monitoring the pass category of a radial forging machine, such as Figure 6 As shown, this embodiment specifically includes:

[0120] S601, collecting signal groups.

[0121] The signal group includes forging process sample data and equipment status sample data.

[0122] S602, data preprocessing.

[0123] S603, training data.

[0124] The preprocessed data is used as training data.

[0125] S604, feature extraction.

[0126] Perform feature extraction on the training data.

[0127] S605: Clustering the extracted features.

[0128] The extracted features are clustered using an improved sparse kernel fuzzy clustering algorithm. The specific clustering method is the same as that in the above embodiment and is not limited in this embodiment.

[0129] S606: Determine whether the clustering effect is optimal.

[0130] Determine whether the clustering effect is optimal: determine the clustering accuracy through the target cluster center after clustering. If the accuracy is less than the accuracy threshold, the clustering effect is not optimal; if not, re-execute S604 (adjust the selected features), otherwise, execute step S608.

[0131] S607, test data.

[0132] S608, pass classification.

[0133] S609: Output the classification result.

[0134] This paper proposes a method for monitoring the pass classification of radial forging machines. Through multi-source data fusion, ANOVA kernel decomposition, and dual sparse constraints, the algorithm significantly improves classification performance under complex working conditions through structured kernel functions and adaptive optimization mechanisms. The core of the algorithm is to map equipment status and process parameters to a high-dimensional feature space, using this method to enhance feature separability and achieve dynamic screening of key features through sparse constraints. This method breaks through the limitations of traditional offline prediction and static models, achieving real-time monitoring of forging pass classification and process adaptive optimization. Technical Effect:

[0135] Real-time performance: processing delay ≤ 200ms (processor i9-14900k), a 60% improvement over traditional methods; precision: lane category recognition accuracy ≥ 98.1%, false alarm rate ≤ 4.7%; fast training: training cycle shortened to 2-3 batches.

[0136] When applying the pass category monitoring method of the radial forging machine provided by the present invention, it is not necessary to Figure 1 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0137] The above is a method for monitoring the pass type of a radial forging machine provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding device for monitoring the pass type of a radial forging machine, which includes:

[0138] The first acquisition module is used to obtain the objective function for clustering; the objective function is constructed based on the kernel space distance function between the data point and the cluster center of the pass category, and the membership function of the data point to the pass category; the kernel space distance function includes the kernel function of the data point itself, the kernel function of the cluster center itself, and the kernel function between the data point and the cluster center; the kernel function includes a univariate Gaussian kernel term for representing the similarity between two univariate data points and a bivariate interaction kernel term for representing the interaction effect between the two variables;

[0139] a second acquisition module, configured to acquire forging process sample data and equipment status sample data of the radial forging machine at multiple time points, and perform feature extraction on the forging process sample data and equipment status sample data at each time point to determine features of the multiple data points;

[0140] Initialization module, used to initialize the cluster center of each pass category;

[0141] The update module is used to iteratively update the cluster center of each pass category based on the characteristics of multiple data points, with the goal of minimizing the objective function value, until the iteration meets the convergence condition. The cluster center that meets the convergence condition is determined as the target cluster center of each pass category; the momentum mechanism is introduced in the cluster center update process;

[0142] The monitoring module is used to extract the target features of the forging process data and equipment status data of the radial forging machine at the current moment, and calculate the membership of the target features to the target cluster center of each pass category respectively, and determine the pass category corresponding to the maximum membership as the pass category of the radial forging machine at the current moment.

[0143] The specific definitions of the radial forging machine's pass classification monitoring device can be found in the definitions of the radial forging machine's pass classification monitoring method described above and will not be repeated here. Each module in the aforementioned radial forging machine's pass classification monitoring device can be implemented in whole or in part via software, hardware, or a combination thereof. Each of these modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device's memory in software form, allowing the processor to call and execute operations corresponding to each of these modules.

[0144] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 A pass category monitoring method for a radial forging machine is provided.

[0145] The present invention also provides Figure 7 The structural diagram of the computer equipment shown in FIG. Figure 7 As shown in the figure, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 A pass category monitoring method for a radial forging machine is provided.

[0146] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware using a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes in the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0147] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A method for monitoring pass categories of a radial forging machine, characterized in that: include: Get the objective function for clustering; The objective function is constructed based on the kernel space distance function between the data point and the cluster center of the pass category, as well as the membership function of the data point to the pass category. The kernel space distance function includes the kernel function of the data point itself, the kernel function of the cluster center itself, and the kernel function between the data point and the cluster center. The kernel function includes a univariate Gaussian kernel term used to represent the similarity between two univariate data points and a bivariate interaction kernel term representing the interaction effect between the two variables. The kernel function is: ； in, For data points and data points The kernel function between and are the univariate Gaussian kernel term and the bivariate interaction term, For data points and data points The characteristic dimension of is the feature weight, , is the nuclear sensitivity factor, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, represents the L2 norm, , Indicates the data point The characteristics of the dimension and the The nonlinear interaction effect weights between the features of the dimension, Indicates the data point The characteristics of the dimension and the The sensitivity factor of the interaction term between the features of the dimension, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics, For data points No. Dimensional characteristics; Acquire forging process sample data and equipment status sample data of the radial forging machine at multiple time points, and perform feature extraction on the forging process sample data and equipment status sample data at each time point to determine the features of the multiple data points; Initialize the cluster center of each pass category; with the goal of minimizing the objective function value, iteratively update the cluster center of each pass category based on the characteristics of multiple data points until the iteration meets the convergence condition, and determine the cluster center that meets the convergence condition as the target cluster center of each pass category; The target features of the forging process data and equipment status data of the radial forging machine at the current moment are extracted, and the membership of the target features to the target cluster center of each pass category is calculated. The pass category corresponding to the maximum membership is determined as the pass category of the radial forging machine at the current moment.

2. The method according to claim 1, characterized in that Data Points Cluster centers of the pass categories The kernel space distance function between is: ； in, is the cluster center of the lane category, Represents data points Its own kernel function, is the cluster center Its own kernel function, For data points and cluster centers The kernel function between No. data points For the first Membership function of the pass category for: ； in, is the number of pass categories, For the data points With the The cluster center of each pass category The kernel function between For the data points With the The cluster center of each pass category The kernel function between for The reciprocal of Objective function for: ； in, is the total number of data points, For the data points For the first The membership function of the pass category, For the data points Its own kernel function, For the The cluster center of each pass category Its own kernel function, For the data points With the The cluster center of each pass category The kernel function between is the first regularization coefficient, is the second regularization coefficient, is the feature weight, , for The L1 norm of for The L2 norm of is the fuzzy index.

3. The method according to claim 2, characterized in that Feature weights During the clustering process, the soft threshold algorithm is used for iterative updates; the update formula is: ； in, For the The feature weights of the iteration , For the The first iteration data points For the first The cluster center of each pass category The kernel space distance function between For the The feature weight of the iteration, For the The first iteration data points For the first The membership function of the pass category, represents the soft threshold operator, is the input of the soft threshold operator, is the threshold, for The L2 norm of .

4. The method according to claim 3, characterized in that The momentum mechanism is introduced into the cluster center update process; the cluster center update formula is: ； in, For the The first iteration The cluster centers of the lane categories, For the The first iteration The cluster centers of the lane categories, is the momentum coefficient.

5. The method according to claim 1, characterized in that The convergence conditions are: The objective function value is less than a first preset threshold, and the displacement of the cluster center of each pass category between the current iteration and the previous iteration is less than a second preset threshold.

6. The method according to claim 1, wherein The feature extraction is performed on the forging process sample data and the equipment status sample data at each time point to determine the features of multiple data points, including: The 3-sigma principle is used to filter out abnormal data points that are beyond the range of forging process sample data and equipment status sample data, and the mean filling method is used to fill in the missing values after filtering out abnormal data points to obtain complete data points. Performing denoising on the complete data points to obtain denoised data points; Perform time domain feature extraction and frequency domain feature extraction on the forging process sample data and equipment status sample data of the denoised data points, respectively, to obtain multiple time domain features and multiple frequency domain features for each data point; Calculate the importance scores of all features and select the K features with the largest importance scores as the features of each data point.

7. The method according to claim 6, characterized in that The target features of extracting the forging process data and equipment status data of the radial forging machine at the current moment include: The key features of the forging process data and equipment status data of the radial forging machine at the current moment are extracted, and the key features are determined as target features.

Images