Automobile parts intelligent stamping control system and method thereof

Abstract

This application relates to the field of intelligent control, specifically disclosing an intelligent stamping control system and method for automotive parts. It acquires stamping force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period, and further uses artificial intelligence technology based on a deep neural network model to extract and fuse data at first and second multi-scale levels to determine whether the stamping force should be increased or decreased. This intelligent stamping control scheme for automotive parts can intelligently control the stamping force, improving production efficiency and product quality.

Description

Technical Field

[0001] This application relates to the field of intelligent control, and more specifically, to an intelligent stamping control system and method for automotive parts. Background Technology

[0002] Stamping is a crucial step in automotive parts manufacturing. It involves applying pressure (through the press and the die) to metal sheets of various specifications according to the shape of the die in a stamping press, causing them to plastically deform. However, excessive or insufficient stamping pressure will affect the forming of parts and cannot be precisely controlled, thus affecting production efficiency and forming quality.

[0003] Therefore, an optimized intelligent stamping control scheme for automotive parts is needed. This scheme would collect relevant parameter data such as stamping force and workpiece deformation through devices such as pressure sensors and displacement sensors installed on the machine tool, and obtain classification results to indicate whether the stamping force should be increased or decreased. Summary of the Invention

[0004] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide an intelligent stamping control system and method for automotive parts. This system acquires stamping force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period. Furthermore, it uses artificial intelligence technology based on a deep neural network model to extract and fuse data at first and second multi-scale levels to determine whether the stamping force should be increased or decreased. This intelligent stamping control scheme for automotive parts enables intelligent control of stamping force, improving production efficiency and product quality.

[0005] According to one aspect of this application, an intelligent stamping control system for automotive parts is provided, comprising:

[0006] The data acquisition module is used to acquire the impact force value and part displacement value at multiple predetermined time points within a predetermined time period;

[0007] The timing arrangement module is used to arrange the punching force values ​​and part displacement values ​​at the multiple predetermined time points into punching force input vectors and displacement input vectors according to the time dimension, respectively.

[0008] The punching force feature extraction module is used to extract the punching force input vector through the first multi-scale neighborhood feature extraction module to obtain the punching force feature vector.

[0009] The displacement feature extraction module is used to extract the displacement input vector and then pass it through the second multi-scale neighborhood feature extraction module to obtain the displacement feature vector.

[0010] The projection module is used to perform a spatial simultaneous projection of the impact force feature vector and the displacement feature vector based on the sub-dimensions of the feature set to obtain the classification feature vector; and

[0011] The classification result module is used to pass the classification feature vector through a classifier to obtain the classification result, which indicates whether the force of the stamping needs to be increased or decreased.

[0012] In the aforementioned intelligent stamping control system for automotive parts, the stamping force feature extraction module includes: a first-scale stamping force extraction unit, used to input the stamping force input vector into the first convolutional layer of the first multi-scale neighborhood feature extraction module to obtain a first-scale stamping force feature vector, wherein the first convolutional layer has a first one-dimensional convolutional kernel of a first length; a second-scale stamping force extraction unit, used to input the stamping force input vector into the second convolutional layer of the first multi-scale neighborhood feature extraction module to obtain a second-scale stamping force feature vector, wherein the second convolutional layer has a second one-dimensional convolutional kernel of a second length, the first length being different from the second length; and a stamping force feature extraction unit, used to cascade the first-scale stamping force feature vector and the second-scale stamping force feature vector using a cascaded layer of the first multi-scale neighborhood feature extraction module to obtain the stamping force feature vector.

[0013] In the aforementioned intelligent stamping control system for automotive parts, the first-scale stamping force extraction unit is used to: perform one-dimensional convolutional encoding on the stamping force input vector using the first convolutional layer of the first multi-scale neighborhood feature extraction module with the following first one-dimensional convolution formula to obtain the first-scale stamping force feature vector; wherein, the first one-dimensional convolution formula is:

[0014]

[0015] Where a is the width of the first one-dimensional convolution kernel in the x direction, F(a) is the parameter vector of the first one-dimensional convolution kernel, G(xa) is the local vector matrix operated with the first convolution kernel function, w is the size of the first one-dimensional convolution kernel, X represents the punch input vector, and Cov(X) is the one-dimensional convolution encoding of the punch input vector.

[0016] In the aforementioned intelligent stamping control system for automotive parts, the second-scale stamping force extraction unit is used to: perform one-dimensional convolutional encoding on the stamping force input vector using the second convolutional layer of the first multi-scale neighborhood feature extraction module with the following second one-dimensional convolution formula to obtain the second-scale stamping force feature vector; wherein, the second one-dimensional convolution formula is:

[0017]

[0018] Where b is the width of the second one-dimensional convolution kernel in the x direction, F(b) is the parameter vector of the second one-dimensional convolution kernel, G(xb) is the local vector matrix of the operation with the second convolution kernel function, m is the size of the second one-dimensional convolution kernel, X represents the punch input vector, and Cov(X) is the one-dimensional convolution encoding of the punch input vector.

[0019] In the aforementioned intelligent stamping control system for automotive parts, the displacement feature extraction module includes: a first-scale displacement extraction unit, used to input the displacement input vector into the third convolutional layer of the second multi-scale neighborhood feature extraction module to obtain a first-scale displacement feature vector, wherein the third convolutional layer has a third one-dimensional convolutional kernel of a third length; a second-scale displacement extraction unit, used to input the displacement input vector into the fourth convolutional layer of the second multi-scale neighborhood feature extraction module to obtain a second-scale displacement feature vector, wherein the fourth convolutional layer has a fourth one-dimensional convolutional kernel of a fourth length, and the third length is different from the fourth length; and a displacement feature extraction unit, used to cascade the first-scale displacement feature vector and the second-scale displacement feature vector using a cascaded layer of the second multi-scale neighborhood feature extraction module to obtain the displacement feature vector.

[0020] In the aforementioned intelligent stamping control system for automotive parts, the first-scale displacement extraction unit is used to: perform one-dimensional convolutional encoding on the displacement input vector using the third convolutional layer of the second multi-scale neighborhood feature extraction module with the following third one-dimensional convolution formula to obtain the first-scale displacement feature vector; wherein, the third one-dimensional convolution formula is:

[0021]

[0022] Where c is the width of the first one-dimensional convolution kernel in the y direction, F(c) is the parameter vector of the first one-dimensional convolution kernel, G(yc) is the local vector matrix operated with the third convolution kernel function, z is the size of the first one-dimensional convolution kernel, Y represents the displacement input vector, and Cov(Y) is the one-dimensional convolution encoding of the displacement input vector.

[0023] In the aforementioned intelligent stamping control system for automotive parts, the second-scale displacement extraction unit is used to: perform one-dimensional convolutional encoding on the displacement input vector using the fourth convolutional layer of the second multi-scale neighborhood feature extraction module with the following fourth one-dimensional convolution formula to obtain the second-scale displacement feature vector; wherein, the fourth one-dimensional convolution formula is:

[0024]

[0025] Where d is the width of the second one-dimensional convolution kernel in the y direction, F(d) is the parameter vector of the second one-dimensional convolution kernel, G(yd) is the local vector matrix operated with the fourth convolution kernel function, n is the size of the second one-dimensional convolution kernel, Y represents the displacement input vector, and Cov(Y) is the one-dimensional convolution encoding of the displacement input vector.

[0026] In the aforementioned intelligent stamping control system for automotive parts, the classification result module includes: a fully connected encoding unit, used to perform fully connected encoding on the classification feature vector using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and a classification result unit, used to pass the encoded classification feature vector through the Softmax classification function of the classifier to obtain the classification result.

[0027] According to another aspect of this application, an intelligent stamping control method for automotive parts is provided, comprising:

[0028] Obtain the impact force value and part displacement value at multiple predetermined time points within a predetermined time period;

[0029] The punching force values ​​and part displacement values ​​at the multiple predetermined time points are arranged according to the time dimension to form punching force input vector and displacement input vector, respectively.

[0030] The punching force input vector is processed by the first multi-scale neighborhood feature extraction module to obtain the punching force feature vector.

[0031] The displacement input vector is then processed by the second multi-scale neighborhood feature extraction module to obtain the displacement feature vector.

[0032] The classification feature vector is obtained by performing a spatial simultaneous projection between the sub-dimensions of the feature set on the impact force feature vector and the displacement feature vector; and

[0033] The classification feature vector is passed through a classifier to obtain a classification result, which is used to indicate whether the force of the stamping should be increased or decreased.

[0034] Compared with existing technologies, this application provides an intelligent stamping control system and method for automotive parts. It acquires stamping force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period. Furthermore, it uses artificial intelligence technology based on a deep neural network model to extract and fuse data at first and second multi-scale levels to determine whether the stamping force should be increased or decreased. This intelligent stamping control scheme for automotive parts can intelligently control the stamping force, improving production efficiency and product quality. Attached Figure Description

[0035] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0036] Figure 1 This is an application scenario diagram of the intelligent stamping control system for automotive parts according to an embodiment of this application.

[0037] Figure 2 This is a block diagram of an intelligent stamping control system for automotive parts according to an embodiment of this application.

[0038] Figure 3 This is a schematic diagram of the architecture of an intelligent stamping control system for automotive parts according to an embodiment of this application.

[0039] Figure 4 This is a block diagram of the displacement feature extraction module in the intelligent stamping control system for automotive parts according to an embodiment of this application.

[0040] Figure 5 This is a flowchart of an intelligent stamping control method for automotive parts according to an embodiment of this application.

[0041] Figure 6 This is a block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0042] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0043] Application Overview

[0044] As mentioned above, traditional restaurants rely on manual payment for food, leading to excessively long waiting times. Currently, most restaurants use the shape and color of tableware for payment, requiring the purchase of specially made tableware, which lacks a high level of intelligence. Therefore, there is a need for an optimized intelligent stamping control solution for automotive parts.

[0045] To address the aforementioned technical issues, the applicant of this application collects relevant parameter data such as stamping force and workpiece deformation by using pressure sensors and displacement sensors installed on the machine tool, and obtains classification results to indicate whether the stamping force should be increased or decreased.

[0046] Accordingly, in the technical solution of this application, it is considered that during the stamping process, the force applied by the punch affects the forming and deformation of the metal material by the die, and also affects the wear and tear of the cutting edge. By acquiring stamping force data, the force applied by the punch can be analyzed and adjusted in real time to achieve the optimal state, thereby ensuring the accuracy and stability of the stamping. During the stamping process, the deformation and displacement of the workpiece directly affect the dimensional accuracy and surface quality of the stamping. By acquiring displacement data, the deformation and displacement of the workpiece can be monitored and controlled in real time, thereby adjusting the position and force of the punch in a timely manner to meet the design requirements, improving product quality and reputation.

[0047] In recent years, deep learning and neural networks have been widely applied in fields such as computer vision, natural language processing, and text signal processing. Furthermore, deep learning and neural networks have demonstrated near-human or even superior performance in areas such as image classification, object detection, semantic segmentation, and text translation.

[0048] Specifically, in the technical solution of this application, firstly, the stamping force value and part displacement value at multiple predetermined time points within a predetermined time period are obtained. Next, considering that the force applied by the punch and the deformation of the workpiece change over time during the stamping process, arranging the data into a vector according to time sequence can reflect the dynamic changes in force and deformation during the stamping process, thereby achieving real-time monitoring and control of the stamping process, timely detection of abnormalities and adjustments; it also facilitates the analysis and comparison of the dynamic changes in force and deformation during the stamping process, identifying its patterns and characteristics. Therefore, the stamping force value and part displacement value at the multiple predetermined time points are arranged into a stamping force input vector and a displacement input vector respectively according to the time dimension.

[0049] Furthermore, considering that a multi-scale approach allows for the observation and analysis of punching force and displacement data from different perspectives, fully exploring their time-domain and frequency-domain characteristics, different workpieces and dies may produce different punching force intensities, deformation patterns, and displacement variation laws. A multi-scale approach can better adapt to different workpieces and dies. By employing information from multiple scales, more punching force and displacement features can be obtained, improving the accuracy and stability of the classifier. Therefore, the punching force input vector is processed through a first multi-scale neighborhood feature extraction module to obtain a punching force feature vector; the displacement input vector is then processed through a second multi-scale neighborhood feature extraction module to obtain a displacement feature vector.

[0050] Next, considering that stamping force and displacement are two important factors affecting the quality of stamping processes, reflecting the force applied by the punch and the deformation of the workpiece respectively, fusing the data features of both allows for full exploitation of data complementarity, enabling the acquisition of more global information and enhancing the comprehensiveness and reliability of the data. Fusing multiple feature vectors can improve the accuracy and stability of the classifier, reduce false positives and false negatives, and increase overall accuracy. Differences may exist between different regions, manufacturers, and products; fusing multiple feature vectors can overcome these differences and establish a more universal and comprehensive model. Therefore, the stamping force feature vector and the displacement feature vector are fused to obtain the classification feature vector.

[0051] Then, the classification feature vector is passed through a classifier to obtain the classification result, which indicates whether the stamping force should be increased or decreased. Considering that data is classified by a classifier, it is possible to detect in real time whether the force applied by the punch during stamping is appropriate, promptly identify abnormalities, and make adjustments. The force applied by the punch can be automatically adjusted, reducing manual intervention and improving production efficiency and quality. Real-time monitoring and control of stamping parameters through a classifier can improve the stability and reliability of processing, and reduce product defect rates and scrap rates.

[0052] Specifically, in the technical solution of this application, when fusing the stamping force feature vector and the displacement feature vector to obtain a classification feature vector, the stamping force feature vector represents the temporal correlation feature of the stamping force value, and the displacement feature vector is used to represent the temporal correlation feature of the part displacement value. The classification feature vector fuses the temporal correlation features of the stamping force value and the temporal correlation features of the part displacement value to more comprehensively characterize the stamping process and stamping result. However, in the technical solution of this application, part of the reason for the displacement of the part is that stamping force is applied to the part. Therefore, if the stamping force feature vector and the displacement feature vector are directly fused in a position-weighted sum manner to obtain the classification feature vector, the feature redundancy between the stamping force feature vector and the displacement feature vector will reduce the accuracy of the classification result obtained by the classifier from the classification feature vector.

[0053] Based on this, in the technical solution of this application, the impact force feature vector and the displacement feature vector are subjected to spatial simultaneous projection based on the sub-dimensions of the feature set to obtain the classification feature vector. Specifically, firstly, the covariance matrix between the impact force feature vector and the displacement feature vector is calculated; then, the covariance matrix is ​​decomposed into eigenvalues ​​to obtain multiple eigenvalues ​​and multiple feature vectors corresponding to the multiple eigenvalues; finally, the multiple feature vectors are arranged into the classification feature vector.

[0054] In this way, by performing a spatial simultaneous projection of the impact force feature vector and the displacement feature vector based on the sub-dimensions of the feature set, the direction and magnitude of the data changes in the impact force feature vector and the displacement feature vector can be fully considered. This allows the classification feature vector obtained by the spatial simultaneous projection to retain the main feature information of the impact force feature vector and the displacement feature vector to the greatest extent while reducing noise and redundant information in the data. In this way, patterns and regularities in the classification feature vector can be better observed and analyzed. This improves the accuracy of the classification judgment of the classification feature vector.

[0055] Based on this, this application provides an intelligent stamping control system for automotive parts, comprising: a data acquisition module for acquiring stamping force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period; a time sequence arrangement module for arranging the stamping force values ​​and part displacement values ​​at the multiple predetermined time points into a stamping force input vector and a displacement input vector respectively according to the time dimension; a stamping force feature extraction module for extracting the stamping force input vector through a first multi-scale neighborhood feature extraction module to obtain a stamping force feature vector; a displacement feature extraction module for extracting the displacement input vector through a second multi-scale neighborhood feature extraction module to obtain a displacement feature vector; a projection module for performing a spatial joint projection of the stamping force feature vector and the displacement feature vector based on the sub-dimensions of the feature set to obtain a classification feature vector; and a classification result module for passing the classification feature vector through a classifier to obtain a classification result, wherein the classification result indicates whether the stamping force should be increased or decreased.

[0056] Figure 1 This is an application scenario diagram of the intelligent stamping control system for automotive parts according to an embodiment of this application. For example... Figure 1 As shown, in this application scenario, by installing on a machine tool (e.g., such as...) Figure 1 The pressure sensor on the M shown (e.g., such as) Figure 1 The P shown is shown) and the displacement sensor (e.g., such as Figure 1 As shown in D), the stamping force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period are obtained. Then, this information is input to a server deployed with intelligent stamping control algorithms for automotive parts (e.g., [example server]). Figure 1 In S), the server is able to process the input data using the intelligent stamping control algorithm for automotive parts to generate a classification result indicating whether the stamping force should be increased or decreased.

[0057] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0058] Exemplary System

[0059] Figure 2 This is a block diagram of an intelligent stamping control system for automotive parts according to an embodiment of this application. Figure 2 As shown, the intelligent stamping control system 100 for automotive parts according to an embodiment of this application includes: a data acquisition module 110, used to acquire stamping force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period; a time sequence arrangement module 120, used to arrange the stamping force values ​​and part displacement values ​​at the multiple predetermined time points into a stamping force input vector and a displacement input vector respectively according to the time dimension; a stamping force feature extraction module 130, used to extract the stamping force input vector through a first multi-scale neighborhood feature extraction module to obtain a stamping force feature vector; a displacement feature extraction module 140, used to extract the displacement input vector through a second multi-scale neighborhood feature extraction module to obtain a displacement feature vector; a projection module 150, used to perform a spatial joint projection of the stamping force feature vector and the displacement feature vector based on the spatial joint projection between each sub-dimension of the feature set to obtain the classification feature vector; and a classification result module 160, used to pass the classification feature vector through a classifier to obtain a classification result, wherein the classification result indicates whether the stamping force should be increased or decreased.

[0060] Figure 3 This is a schematic diagram of the architecture of an intelligent stamping control system for automotive parts according to an embodiment of this application. Figure 3 As shown, firstly, the punching force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period are obtained. Next, the punching force values ​​and part displacement values ​​at these multiple predetermined time points are arranged according to the time dimension to form punching force input vectors and displacement input vectors, respectively. Then, the punching force input vector is processed through a first multi-scale neighborhood feature extraction module to obtain a punching force feature vector. Simultaneously, the displacement input vector is processed through a second multi-scale neighborhood feature extraction module to obtain a displacement feature vector. Next, the punching force feature vector and the displacement feature vector are subjected to a spatial simultaneous projection based on the sub-dimensions of the feature set to obtain a classification feature vector. Finally, the classification feature vector is processed by a classifier to obtain a classification result, which indicates whether the punching force should be increased or decreased.

[0061] In this embodiment, the data acquisition module 110 is used to acquire the punching force value and part displacement value at multiple predetermined time points within a predetermined time period. Considering that the force applied by the punch during the stamping process affects the forming and deformation of the metal material by the die, and also affects the wear and tear of the cutting edge, acquiring the punching force data allows for real-time analysis and adjustment of the punching force to achieve optimal performance, thereby ensuring the accuracy and stability of the stamping process. During the stamping process, the deformation and displacement of the workpiece directly affect the dimensional accuracy and surface quality of the stamping. Acquiring displacement data allows for real-time monitoring and control of the workpiece's deformation and displacement, enabling timely adjustment of the punch position and force to meet design requirements, thereby improving product quality and reputation.

[0062] In this embodiment, the time-series arrangement module 120 is used to arrange the punching force values ​​and part displacement values ​​at multiple predetermined time points into punching force input vectors and displacement input vectors according to the time dimension, respectively. Considering that the force applied by the punch and the deformation of the workpiece change over time during the stamping process, arranging the data into vectors in time order can reflect the dynamic changes in force and deformation during the stamping process, thereby enabling real-time monitoring and control of the stamping process, timely detection of abnormalities and adjustments; it also facilitates the analysis and comparison of the dynamic changes in force and deformation during the stamping process, identifying its patterns and characteristics.

[0063] In this embodiment, the stamping force feature extraction module 130 is used to extract the stamping force input vector through a first multi-scale neighborhood feature extraction module to obtain a stamping force feature vector. Considering that a multi-scale approach allows for observation and analysis of stamping force data from different perspectives, fully exploring its time-domain, frequency-domain, and other features, and that different workpieces and dies may produce different stamping forces, a multi-scale approach can better adapt to different workpieces and dies. By using information from multiple scales, more stamping force features can be obtained, improving the accuracy and stability of the classifier.

[0064] Specifically, in this embodiment, the impact pressure feature extraction module includes: a first-scale impact pressure extraction unit, configured to input the impact pressure input vector into a first convolutional layer of the first multi-scale neighborhood feature extraction module to obtain a first-scale impact pressure feature vector, wherein the first convolutional layer has a first one-dimensional convolutional kernel of a first length; a second-scale impact pressure extraction unit, configured to input the impact pressure input vector into a second convolutional layer of the first multi-scale neighborhood feature extraction module to obtain a second-scale impact pressure feature vector, wherein the second convolutional layer has a second one-dimensional convolutional kernel of a second length, the first length being different from the second length; and an impact pressure feature extraction unit, configured to cascade the first-scale impact pressure feature vector and the second-scale impact pressure feature vector using a cascaded layer of the first multi-scale neighborhood feature extraction module to obtain the impact pressure feature vector.

[0065] More specifically, in this embodiment, the first-scale impact force extraction unit is configured to: use the first convolutional layer of the first multi-scale neighborhood feature extraction module to perform one-dimensional convolutional encoding on the impact force input vector using the following first one-dimensional convolution formula to obtain the first-scale impact force feature vector; wherein, the first one-dimensional convolution formula is:

[0066]

[0067] Where a is the width of the first one-dimensional convolution kernel in the x direction, F(a) is the parameter vector of the first one-dimensional convolution kernel, G(xa) is the local vector matrix operated with the first convolution kernel function, w is the size of the first one-dimensional convolution kernel, X represents the punch input vector, and Cov(X) is the one-dimensional convolution encoding of the punch input vector.

[0068] More specifically, in this embodiment, the second-scale impact force extraction unit is configured to: use the second convolutional layer of the first multi-scale neighborhood feature extraction module to perform one-dimensional convolutional encoding on the impact force input vector using the following second one-dimensional convolution formula to obtain the second-scale impact force feature vector; wherein, the second one-dimensional convolution formula is:

[0069]

[0070] Where b is the width of the second one-dimensional convolution kernel in the x direction, F(b) is the parameter vector of the second one-dimensional convolution kernel, G(xb) is the local vector matrix of the operation with the second convolution kernel function, m is the size of the second one-dimensional convolution kernel, X represents the punch input vector, and Cov(X) is the one-dimensional convolution encoding of the punch input vector.

[0071] In this embodiment, the displacement feature extraction module 140 is used to extract the displacement input vector and then pass it through the second multi-scale neighborhood feature extraction module to obtain a displacement feature vector. Considering that a multi-scale approach allows for observation and analysis of displacement data from different perspectives, fully exploring its time-domain, frequency-domain, and other features, and that different workpieces and dies may exhibit different deformation patterns and displacement variation laws, a multi-scale approach can better adapt to different workpieces and dies. By employing information from multiple scales, more displacement features can be obtained, improving the accuracy and stability of the classifier.

[0072] Figure 4 This is a block diagram of a displacement feature extraction module in an intelligent stamping control system for automotive parts according to an embodiment of this application. Specifically, in an embodiment of this application, as shown... Figure 4 As shown, the displacement feature extraction module 140 includes: a first-scale displacement extraction unit 141, used to input the displacement input vector into the third convolutional layer of the second multi-scale neighborhood feature extraction module to obtain a first-scale displacement feature vector, wherein the third convolutional layer has a third one-dimensional convolutional kernel of a third length; a second-scale displacement extraction unit 142, used to input the displacement input vector into the fourth convolutional layer of the second multi-scale neighborhood feature extraction module to obtain a second-scale displacement feature vector, wherein the fourth convolutional layer has a fourth one-dimensional convolutional kernel of a fourth length, and the third length is different from the fourth length; and a displacement feature extraction unit 143, used to cascade the first-scale displacement feature vector and the second-scale displacement feature vector using a cascaded layer of the second multi-scale neighborhood feature extraction module to obtain the displacement feature vector.

[0073] More specifically, in this embodiment, the first-scale displacement extraction unit is configured to: use the third convolutional layer of the second multi-scale neighborhood feature extraction module to perform one-dimensional convolutional encoding on the displacement input vector using the following third one-dimensional convolution formula to obtain the first-scale displacement feature vector; wherein, the third one-dimensional convolution formula is:

[0074]

[0075] Where c is the width of the first one-dimensional convolution kernel in the y direction, F(c) is the parameter vector of the first one-dimensional convolution kernel, G(yc) is the local vector matrix operated with the third convolution kernel function, z is the size of the first one-dimensional convolution kernel, Y represents the displacement input vector, and Cov(Y) is the one-dimensional convolution encoding of the displacement input vector.

[0076] More specifically, in this embodiment, the second-scale displacement extraction unit is configured to: use the fourth convolutional layer of the second multi-scale neighborhood feature extraction module to perform one-dimensional convolutional encoding on the displacement input vector using the following fourth one-dimensional convolution formula to obtain the second-scale displacement feature vector; wherein, the fourth one-dimensional convolution formula is:

[0077]

[0078] Where d is the width of the second one-dimensional convolution kernel in the y direction, F(d) is the parameter vector of the second one-dimensional convolution kernel, G(yd) is the local vector matrix operated with the fourth convolution kernel function, n is the size of the second one-dimensional convolution kernel, Y represents the displacement input vector, and Cov(Y) is the one-dimensional convolution encoding of the displacement input vector.

[0079] In this embodiment, the projection module 150 is used to perform a spatial simultaneous projection of the punching force feature vector and the displacement feature vector based on the sub-dimensions of the feature set to obtain the classification feature vector. Considering that punching force and displacement are two important factors affecting the quality of stamping processing, reflecting the force applied by the punch and the deformation of the workpiece respectively, fusing the data features of both can fully explore the complementarity of the data, obtain more global information, and enhance the comprehensiveness and reliability of the data. Fusing multiple feature vectors can improve the accuracy and stability of the classifier, reduce the false positive and false negative rates, and improve accuracy. Differences may exist between different regions, manufacturers, and products; fusing multiple feature vectors can overcome these differences and establish a more universal and comprehensive model.

[0080] Specifically, in the technical solution of this application, when fusing the stamping force feature vector and the displacement feature vector to obtain a classification feature vector, the stamping force feature vector represents the temporal correlation feature of the stamping force value, and the displacement feature vector is used to represent the temporal correlation feature of the part displacement value. The classification feature vector fuses the temporal correlation features of the stamping force value and the temporal correlation features of the part displacement value to more comprehensively characterize the stamping process and stamping result. However, in the technical solution of this application, part of the reason for the displacement of the part is that stamping force is applied to the part. Therefore, if the stamping force feature vector and the displacement feature vector are directly fused in a position-weighted sum manner to obtain the classification feature vector, the feature redundancy between the stamping force feature vector and the displacement feature vector will reduce the accuracy of the classification result obtained by the classifier from the classification feature vector.

[0081] Specifically, in this embodiment, the projection module includes: a covariance unit for calculating the covariance matrix between the impact force feature vector and the displacement feature vector; a decomposition unit for performing eigenvalue decomposition on the covariance matrix to obtain multiple feature values ​​and multiple feature vectors corresponding to the multiple feature values; and an arrangement unit for arranging the multiple feature vectors into the classification feature vector.

[0082] In this embodiment, the classification result module 160 is used to pass the classification feature vector through a classifier to obtain a classification result. The classification result indicates whether the stamping force should be increased or decreased. Considering that data is classified by a classifier, it is possible to detect in real time whether the force applied by the punch during stamping is appropriate, promptly identify abnormalities, and make adjustments. The force applied by the punch can be automatically adjusted, reducing manual intervention and improving production efficiency and quality. Real-time monitoring and control of stamping parameters through a classifier can improve processing stability and reliability, and reduce product defect rates and scrap rates.

[0083] Specifically, in this embodiment of the application, the classification result module includes: a fully connected encoding unit, used to perform fully connected encoding on the classification feature vector using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and a classification result unit, used to pass the encoded classification feature vector through the Softmax classification function of the classifier to obtain the classification result.

[0084] In summary, the intelligent stamping control system 100 for automotive parts based on the embodiments of this application is explained. It acquires stamping force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period. Furthermore, it uses artificial intelligence technology based on a deep neural network model to extract and fuse data at first and second multi-scale levels to determine whether the stamping force should be increased or decreased. Thus, the intelligent stamping control scheme for automotive parts can intelligently control the stamping force, improving production efficiency and product quality.

[0085] Exemplary methods

[0086] Figure 5 This is a flowchart of an intelligent stamping control method for automotive parts according to an embodiment of this application. Figure 5As shown, the intelligent stamping control method for automotive parts according to an embodiment of this application includes: S110, acquiring stamping force values ​​and part displacement values ​​at multiple predetermined time points within a predetermined time period; S120, arranging the stamping force values ​​and part displacement values ​​at the multiple predetermined time points into a stamping force input vector and a displacement input vector respectively according to the time dimension; S130, obtaining a stamping force feature vector by passing the stamping force input vector through a first multi-scale neighborhood feature extraction module; S140, obtaining a displacement feature vector by passing the displacement input vector through a second multi-scale neighborhood feature extraction module; S150, performing a spatial joint projection between the stamping force feature vector and the displacement feature vector based on the sub-dimensions of the feature set to obtain a classification feature vector; and S160, passing the classification feature vector through a classifier to obtain a classification result, wherein the classification result is used to indicate whether the stamping force should be increased or decreased.

[0087] In one example, in the above-mentioned intelligent stamping control method for automotive parts, the stamping force input vector is passed through a first multi-scale neighborhood feature extraction module to obtain a stamping force feature vector. This includes: inputting the stamping force input vector into a first convolutional layer of the first multi-scale neighborhood feature extraction module to obtain a first-scale stamping force feature vector, wherein the first convolutional layer has a first one-dimensional convolutional kernel of a first length; inputting the stamping force input vector into a second convolutional layer of the first multi-scale neighborhood feature extraction module to obtain a second-scale stamping force feature vector, wherein the second convolutional layer has a second one-dimensional convolutional kernel of a second length, the first length being different from the second length; and cascading the first-scale stamping force feature vector and the second-scale stamping force feature vector using a cascaded layer of the first multi-scale neighborhood feature extraction module to obtain the stamping force feature vector.

[0088] Here, those skilled in the art will understand that the specific operations of each step in the above-described intelligent stamping control method for automotive parts have been referenced above. Figures 1 to 4 The intelligent stamping control system for automotive parts is described in detail therein, and therefore, its repeated description will be omitted.

[0089] Exemplary electronic devices

[0090] Below, for reference Figure 6 This describes an electronic device according to embodiments of the present application. Figure 6 This is a block diagram of an electronic device according to an embodiment of this application. (e.g.) Figure 6 As shown, the electronic device 10 includes one or more processors 11 and memory 12.

[0091] The processor 11 may be a central processing module (CPU) or other form of processing module with data processing and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0092] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may execute the program instructions to implement the functions and / or other desired functions in the intelligent stamping control system and method for automotive parts of the various embodiments of this application described above. Various contents such as stamping force values ​​and part displacement values ​​may also be stored in the computer-readable storage medium.

[0093] In one example, the electronic device 10 may also include an input device 13 and an output device 14, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0094] The input device 13 may include, for example, a keyboard, a mouse, etc.

[0095] The output device 14 can output various information to the outside, including classification results. The output device 14 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0096] Of course, for the sake of simplicity, Figure 6 Only some of the components of the electronic device 10 relevant to this application are shown in this illustration; components such as buses, input / output interfaces, etc., are omitted. In addition, the electronic device 10 may include any other suitable components depending on the specific application.

[0097] Exemplary computer program products and computer-readable storage media

[0098] In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps of the functions in the intelligent stamping control method for automotive parts according to various embodiments of this application described in the "Exemplary Methods" section of this specification.

[0099] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0100] Furthermore, embodiments of this application may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps of the functions in the intelligent stamping control method for automotive parts according to various embodiments of this application described in the "Exemplary Methods" section above.

[0101] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0102] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0103] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0104] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0105] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0106] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

[0107] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0108] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0109] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0110] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0111] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. An intelligent stamping control system for automotive parts, characterized in that, include: The data acquisition module is used to acquire the impact force value and part displacement value at multiple predetermined time points within a predetermined time period; The timing arrangement module is used to arrange the punching force values ​​and part displacement values ​​at the multiple predetermined time points into punching force input vectors and displacement input vectors according to the time dimension, respectively. The punching force feature extraction module is used to extract the punching force input vector through the first multi-scale neighborhood feature extraction module to obtain the punching force feature vector. The displacement feature extraction module is used to extract the displacement input vector and then pass it through the second multi-scale neighborhood feature extraction module to obtain the displacement feature vector. The projection module is used to perform spatial simultaneous projection of the impact force feature vector and the displacement feature vector based on the sub-dimensions of the feature set to obtain the classification feature vector; as well as The classification result module is used to pass the classification feature vector through a classifier to obtain a classification result, which indicates whether the force of the stamping needs to be adjusted to increase or decrease. The punching force feature extraction module includes: The first-scale impact force extraction unit is used to input the impact force input vector into the first convolutional layer of the first multi-scale neighborhood feature extraction module to obtain the first-scale impact force feature vector, wherein the first convolutional layer has a first one-dimensional convolutional kernel of a first length. A second-scale impact force extraction unit is used to input the impact force input vector into the second convolutional layer of the first multi-scale neighborhood feature extraction module to obtain a second-scale impact force feature vector, wherein the second convolutional layer has a second one-dimensional convolutional kernel of a second length, and the first length is different from the second length; and The impact force feature extraction unit is used to cascade the first-scale impact force feature vector and the second-scale impact force feature vector using the cascaded layer of the first multi-scale neighborhood feature extraction module to obtain the impact force feature vector. The first-scale impact force extraction unit is used to: use the first convolutional layer of the first multi-scale neighborhood feature extraction module to perform one-dimensional convolutional encoding on the impact force input vector using the following first one-dimensional convolution formula to obtain the first-scale impact force feature vector. The first one-dimensional convolution formula is as follows: in, For the first one-dimensional convolution kernel in x Width in direction For the first one-dimensional convolution kernel parameter vector, The local vector matrix operated on with the first convolution kernel function, The size of the first one-dimensional convolution kernel. This represents the punch force input vector. One-dimensional convolutional encoding is performed on the hedging pressure input vector.

2. The intelligent stamping control system for automotive parts according to claim 1, characterized in that, The second-scale impact force extraction unit is used to: use the second convolutional layer of the first multi-scale neighborhood feature extraction module to perform one-dimensional convolutional encoding on the impact force input vector using the following second one-dimensional convolution formula to obtain the second-scale impact force feature vector; The second one-dimensional convolution formula is as follows: in, For the second one-dimensional convolution kernel in x Width in direction For the second one-dimensional convolution kernel parameter vector, For the local vector matrix operated on with the second convolution kernel function, The size of the second one-dimensional convolution kernel. This represents the punch force input vector. One-dimensional convolutional encoding is performed on the hedging pressure input vector.

3. The intelligent stamping control system for automotive parts according to claim 2, characterized in that, The displacement feature extraction module includes: The first-scale displacement extraction unit is used to input the displacement input vector into the third convolutional layer of the second multi-scale neighborhood feature extraction module to obtain the first-scale displacement feature vector, wherein the third convolutional layer has a third one-dimensional convolutional kernel of a third length; The second-scale displacement extraction unit is used to input the displacement input vector into the fourth convolutional layer of the second multi-scale neighborhood feature extraction module to obtain a second-scale displacement feature vector, wherein the fourth convolutional layer has a fourth one-dimensional convolutional kernel of a fourth length, and the third length is different from the fourth length; and The displacement feature extraction unit is used to cascade the first-scale displacement feature vector and the second-scale displacement feature vector using the cascaded layer of the second multi-scale neighborhood feature extraction module to obtain the displacement feature vector.

4. The intelligent stamping control system for automotive parts according to claim 3, characterized in that, The first-scale displacement extraction unit is used to: use the third convolutional layer of the second multi-scale neighborhood feature extraction module to perform one-dimensional convolutional encoding on the displacement input vector using the following third one-dimensional convolution formula to obtain the first-scale displacement feature vector. The third one-dimensional convolution formula is as follows: in, The width of the first one-dimensional convolution kernel in the y-direction, For the first one-dimensional convolution kernel parameter vector, For the local vector matrix operated on with the third convolution kernel function, The size of the first one-dimensional convolution kernel. This represents the displacement input vector. This is for one-dimensional convolutional encoding of the displacement input vector.

5. The intelligent stamping control system for automotive parts according to claim 4, characterized in that, The second-scale displacement extraction unit is used to: use the fourth convolutional layer of the second multi-scale neighborhood feature extraction module to perform one-dimensional convolutional encoding on the displacement input vector using the following fourth one-dimensional convolution formula to obtain the second-scale displacement feature vector; The fourth one-dimensional convolution formula is as follows: in, For the second one-dimensional convolution kernel in y Width in direction For the second one-dimensional convolution kernel parameter vector, For the local vector matrix operated on with the fourth convolution kernel function, The size of the second one-dimensional convolution kernel. This represents the displacement input vector. This is for one-dimensional convolutional encoding of the displacement input vector.

6. The intelligent stamping control system for automotive parts according to claim 5, characterized in that, The projection module includes: Covariance unit, used to calculate the covariance matrix between the impact force eigenvector and the displacement eigenvector; A decomposition unit is configured to perform eigenvalue decomposition on the covariance matrix to obtain multiple eigenvalues ​​and multiple eigenvectors corresponding to the multiple eigenvalues; and An arrangement unit is used to arrange the plurality of feature vectors into the classification feature vector.

7. The intelligent stamping control system for automotive parts according to claim 6, characterized in that, The classification result module includes: A fully connected encoding unit is configured to perform fully connected encoding on the classification feature vector using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and The classification result unit is used to pass the encoded classification feature vector through the Softmax classification function of the classifier to obtain the classification result.

8. A method for intelligent stamping control of automotive parts, employing the intelligent stamping control system for automotive parts as described in claim 1, characterized in that, include: Obtain the impact force value and part displacement value at multiple predetermined time points within a predetermined time period; The punching force values ​​and part displacement values ​​at the multiple predetermined time points are arranged according to the time dimension to form punching force input vector and displacement input vector, respectively. The punching force input vector is processed by the first multi-scale neighborhood feature extraction module to obtain the punching force feature vector. The displacement input vector is then processed by the second multi-scale neighborhood feature extraction module to obtain the displacement feature vector. The classification feature vector is obtained by performing a spatial simultaneous projection between the sub-dimensions of the feature set on the punching force feature vector and the displacement feature vector; and the classification feature vector is passed through a classifier to obtain a classification result, which is used to indicate whether the punching force should be increased or decreased.

Images