CN116765925A - CNC-based die cutting machining control system and method thereof - Google Patents
CNC-based die cutting machining control system and method thereof Download PDFInfo
- Publication number
- CN116765925A CN116765925A CN202310690429.7A CN202310690429A CN116765925A CN 116765925 A CN116765925 A CN 116765925A CN 202310690429 A CN202310690429 A CN 202310690429A CN 116765925 A CN116765925 A CN 116765925A
- Authority
- CN
- China
- Prior art keywords
- vector
- feature
- temperature
- pressure
- time sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000005520 cutting process Methods 0.000 title claims abstract description 111
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000003754 machining Methods 0.000 title claims description 27
- 238000012545 processing Methods 0.000 claims abstract description 26
- 239000013598 vector Substances 0.000 claims description 235
- 239000011159 matrix material Substances 0.000 claims description 46
- 238000000605 extraction Methods 0.000 claims description 36
- 230000004927 fusion Effects 0.000 claims description 19
- 230000008859 change Effects 0.000 claims description 14
- 238000005538 encapsulation Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims 2
- 238000010276 construction Methods 0.000 claims 1
- 230000008569 process Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 8
- 238000009826 distribution Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 238000004519 manufacturing process Methods 0.000 description 7
- 238000012549 training Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 4
- 230000001939 inductive effect Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 238000003058 natural language processing Methods 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000009530 blood pressure measurement Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000002195 synergetic effect Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Landscapes
- Numerical Control (AREA)
Abstract
The application relates to the field of intelligent control, and particularly discloses a CNC (computer numerical control) -based die cutting and processing control system and a CNC-based die cutting and processing control method.
Description
Technical Field
The application relates to the field of intelligent control, and more particularly relates to a CNC (computer numerical control) -based die cutting machining control system and a CNC-based die cutting machining control method.
Background
The cutting process of the die is a very important link in the manufacturing industry, and the quality and the efficiency of the cutting process directly affect the stability and the economic benefit of the whole production line. In recent years, with the continuous improvement of the industrial automation level, a CNC-based die cutting processing control system is increasingly widely used. However, the die machining requires a high-precision cutting operation, and the conventional die cutting operation often depends on a manual experience control operation, so that the efficiency is low, and the cutting precision is difficult to ensure.
Accordingly, an optimized CNC-based die cutting control system is desired to improve the efficiency and accuracy of the die cutting process, optimizing manufacturing processes and product quality.
Disclosure of Invention
The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides a CNC (computer numerical control) -based die cutting and processing control system and a CNC-based die cutting and processing control method, which are used for accurately pre-warning faults by adopting a neural network model based on deep learning to excavate time sequence collaborative implicit association characteristics of the position, the temperature and the pressure of a cutting machine tool in the die cutting and processing process, so that the efficiency and the accuracy of die cutting and processing are improved, and the manufacturing process and the product quality are optimized.
According to one aspect of the present application, there is provided a CNC-based die cutting control system comprising: the data acquisition module is used for acquiring position data, temperature data and pressure data of the monitored cutting machine tool at a plurality of preset time points in a preset time period; the position time sequence change feature extraction module is used for arranging the position data of the monitored cutting machine tool into position time sequence input vectors according to time and then obtaining position time sequence correlation feature vectors through the multi-scale neighborhood feature extraction module; the temperature-pressure time sequence association module is used for respectively arranging the temperature data and the pressure data of the monitored cutting machine tool at a plurality of preset time points into a temperature time sequence input vector and a pressure time sequence input vector according to a time dimension, and then carrying out association coding on the temperature time sequence input vector and the pressure time sequence input vector to obtain a temperature-pressure full-time sequence association matrix; the feature association coding module is used for dividing the temperature-pressure full-time sequence association matrix into sequences of feature submatrices by a feature matrix, and obtaining a temperature-pressure global context association feature vector through a context encoder comprising an embedded layer; the feature fusion module is used for fusing the temperature-pressure global context correlation feature vector and the position time sequence time correlation feature vector to obtain a classification feature vector; and
The fault early warning module is used for enabling the classification feature vector to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether a fault alarm prompt is generated or not.
According to another aspect of the present application, there is provided a CNC-based die cutting control method including: acquiring position data, temperature data and pressure data of a monitored cutting machine tool at a plurality of preset time points in a preset time period; the position data of the monitored cutting machine tool are arranged into position time sequence input vectors according to time, and then the position time sequence input vectors are obtained through a multi-scale neighborhood feature extraction module; after arranging the temperature data and the pressure data of the monitored cutting machine tool at a plurality of preset time points into a temperature time sequence input vector and a pressure time sequence input vector according to a time dimension, performing association coding on the temperature time sequence input vector and the pressure time sequence input vector to obtain a temperature-pressure full-time sequence association matrix; after the temperature-pressure full-time sequence correlation matrix is subjected to feature matrix division into sequences of feature submatrices, a context encoder comprising an embedded layer is used for obtaining a temperature-pressure global context correlation feature vector; fusing the temperature-pressure global context correlation feature vector and the position time sequence correlation feature vector to obtain a classification feature vector; and
And the classification feature vector passes through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fault alarm prompt is generated or not.
Compared with the prior art, the CNC-based die cutting processing control system and the CNC-based die cutting processing control method provided by the application have the advantages that the time sequence of the position, the temperature and the pressure of the cutting machine tool in the die cutting processing process is excavated by adopting the neural network model based on deep learning to cooperatively and implicitly correlate the characteristics, so that the fault is accurately early warned, the die cutting processing efficiency and accuracy are further improved, and the manufacturing process and the product quality are optimized.
Other features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments, which proceeds with reference to the accompanying drawings.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing embodiments of the present application in more detail with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and together with the embodiments of the application, and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
FIG. 1 is a block diagram of a CNC-based die cutting machining control system in accordance with an embodiment of the present application.
Fig. 2 is a system architecture diagram of a CNC-based die cutting machining control system according to an embodiment of the present application.
FIG. 3 is a block diagram of a position timing variation feature extraction module in a CNC-based die cutting machining control system according to an embodiment of the present application.
FIG. 4 is a block diagram of a fault pre-warning module in a CNC-based die cutting control system according to an embodiment of the present application.
Fig. 5 is a flowchart of a CNC-based die cutting control method according to an embodiment of the present application.
Fig. 6 is a schematic diagram of a scenario of a CNC-based die cutting control system according to an embodiment of the present application.
Detailed Description
Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
FIG. 1 is a block diagram of a CNC-based die cutting machining control system in accordance with an embodiment of the present application. Fig. 2 is a system architecture diagram of a CNC-based die cutting machining control system according to an embodiment of the present application. As shown in fig. 1 and 2, a CNC-based die cutting machining control system 300 according to an embodiment of the present application includes: a data acquisition module 310, configured to acquire position data, temperature data, and pressure data of the monitored cutting machine at a plurality of predetermined time points within a predetermined time period; the position time sequence change feature extraction module 320 is configured to arrange position data of the monitored cutting machine tool into position time sequence input vectors according to time, and then obtain position time sequence time-related feature vectors through the multi-scale neighborhood feature extraction module; the temperature-pressure time sequence association module 330 is configured to arrange the temperature data and the pressure data of the monitored cutting machine tool at the plurality of predetermined time points into a temperature time sequence input vector and a pressure time sequence input vector according to a time dimension, and then perform association encoding on the temperature time sequence input vector and the pressure time sequence input vector to obtain a temperature-pressure full time sequence association matrix; the feature association encoding module 340 is configured to divide the temperature-pressure full-time sequence association matrix into sequences of feature submatrices, and then obtain a temperature-pressure global context association feature vector through a context encoder including an embedded layer; a feature fusion module 350, configured to fuse the temperature-pressure global context correlation feature vector and the position time sequence time correlation feature vector to obtain a classification feature vector; and a fault early warning module 360, configured to pass the classification feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether a fault alarm prompt is generated.
Specifically, during operation of the CNC-based die-cutting machining control system 300, the data acquisition module 310 is configured to acquire position data, temperature data, and pressure data of the monitored cutting machine at a plurality of predetermined time points within a predetermined time period. In the actual cutting process of the die, in order to ensure the cutting quality of the die, various parameters in the cutting process need to be monitored in real time. In die cutting, parameters such as position, temperature and pressure of a machine tool affect cutting precision and efficiency, and a synergistic effect exists between the parameters. In a specific example, the time-series change information of the position data, the temperature data and the pressure data of the monitored cutting machine may reflect the operation state and the performance change of the monitored cutting machine in the time period. Specifically, by monitoring the position change of the tool in space, information such as the movement track, the processing path, the speed, the precision and the like of the tool can be known, so that the running condition of the cutting machine can be deduced. For example, if there is an abnormal tool motion path or an excessive deviation angle, tool damage or machine tool errors may be indicated. And, the cutting machine tool which is normally operated should be kept in a proper temperature range so as not to cause problems of cutter deformation, material expansion and the like. Meanwhile, the pressure data of the cutting machine tool can reflect the contact state between the cutter and the workpiece, the processing quality and the like. Too much or too little pressure may affect the finish and accuracy of the workpiece surface, while proper cutting pressure helps to improve machining efficiency and reduce energy consumption. Therefore, through collecting, processing and analyzing the data, the real-time monitoring and the classified identification of the running state of the cutting machine tool can be realized, and then the fault early warning and the timely correction are realized. More specifically, first, position data of the monitored cutting machine at a plurality of predetermined time points in a predetermined period of time is acquired by a position sensor, temperature data of the monitored cutting machine at a plurality of predetermined time points in the predetermined period of time is acquired by a temperature sensor, and pressure data of the monitored cutting machine at a plurality of predetermined time points in the predetermined period of time is acquired by a pressure sensor.
Wherein the position sensor is a sensor for measuring the position or change in position of an object. They typically use electromagnetic, optical or mechanical principles to measure the position of an object. The operating principle of the position sensor can be generalized to measure the position or change in position of an object using different physical principles. Different sensors are suitable for different application scenarios, and suitable sensors need to be selected according to specific requirements.
The temperature sensor is a sensor for measuring temperature. Common temperature sensors include thermocouples, thermistors, semiconductor temperature sensors, and the like. The operating principle of the temperature sensor can be generalized to measure the temperature of an object using different physical principles. Different sensors are suitable for different application scenarios, and suitable sensors need to be selected according to specific requirements.
The pressure sensor is a device for measuring pressure, and the working principle of the pressure sensor is based on acting force of a measured medium on an internal sensing element of the sensor. In general, an inductive element of a pressure sensor is implemented by means of resistance, capacitance, inductance, etc., and when the inductive element is subjected to pressure, the physical quantity of resistance, capacitance or inductance, etc. in the inductive element changes, and the change can be converted into an electrical signal for output, so that pressure measurement is implemented. In short, the working principle of the pressure sensor is to convert the acting force of the measured medium on the sensing element into an electric signal for output, so that the pressure measurement is realized.
Specifically, during the operation of the CNC-based die-cutting control system 300, the position-time-sequence-variation feature extraction module 320 is configured to time-arrange the position data of the monitored cutting machine tool into a position-time-sequence input vector, and then obtain a position-time-sequence-correlation feature vector through the multi-scale neighborhood feature extraction module. Considering that the position data of the monitored cutting machine tool has a dynamic change rule in the time dimension and has volatility and uncertainty in the time dimension, the position data presents different time sequence dynamic characteristic information under different time period spans. Therefore, in the technical scheme of the application, in order to fully express the time sequence change characteristics of the position data of the monitored cutting machine tool, the position data of the monitored cutting machine tool is required to be arranged into position time sequence input vectors according to time, and then the position data of the monitored cutting machine tool is processed in a multi-scale neighborhood characteristic extraction module so as to extract dynamic multi-scale neighborhood associated characteristics of the position data of the monitored cutting machine tool under different time spans, thereby obtaining position time sequence associated characteristic vectors. The multi-scale neighborhood feature extraction module is an algorithm used in the fields of image processing and computer vision, and is mainly used for extracting neighborhood features with different scales in an image. The algorithm generally includes the following steps: firstly, scaling an original image according to different scales to obtain a series of images with different scales; then, for each scale image, extracting features in the image by using a Convolutional Neural Network (CNN) and other algorithms; for each feature map, extracting neighborhood features of different scales by using different convolution kernel sizes and step sizes; then, fusing the neighborhood features with different scales to obtain more comprehensive and more accurate feature representation; the feature representation is input to a classifier or a regressor, and tasks such as classification and regression are performed. The multi-scale neighborhood feature extraction module has the advantages that information of different scales in the image can be fully utilized, and accuracy and robustness of image processing and computer vision tasks are improved. Common applications include the fields of image classification, object detection, face recognition, etc. In a specific example of the present application, the multi-scale neighborhood feature extraction module includes: the device comprises a first convolution layer, a second convolution layer parallel to the first convolution layer and a multi-scale feature fusion layer connected with the first convolution layer and the second convolution layer, wherein the first convolution layer uses a one-dimensional convolution kernel with a first length, and the second convolution layer uses a one-dimensional convolution kernel with a second length.
More specifically, as shown in fig. 3, the position timing change feature extraction module 320 includes: a first neighborhood scale feature extraction unit 321, configured to input the position timing input vector into a first convolution layer of the multi-scale neighborhood feature extraction module to obtain a first neighborhood scale position timing associated feature vector, where the first convolution layer has a first one-dimensional convolution kernel with a first length; a second neighborhood scale feature extraction unit 322 configured to input the position timing input vector into a second convolution layer of the multi-scale neighborhood feature extraction module to obtain a second neighborhood scale position timing associated feature vector, where the second convolution layer has a second one-dimensional convolution kernel of a second length, and the first length is different from the second length; and a multi-scale cascading unit 323, configured to cascade the first neighborhood scale location time-series associated feature vector and the second neighborhood scale location time-series associated feature vector to obtain the location time-series associated feature vector. Wherein, the first neighborhood scale feature extraction unit 321 is configured to: performing one-dimensional convolution coding on the position time sequence input vector by using a first convolution layer of the multi-scale neighborhood feature extraction module according to the following one-dimensional convolution formula to obtain a first neighborhood scale position time sequence associated feature vector; wherein, the formula is: Wherein->For the first convolution kernel at->Width in direction, ++>For the first convolution kernel parameter vector, +.>For a local vector matrix operating with a convolution kernel function, < ->For the size of the first one-dimensional convolution kernel, +.>Representing the position timing input vector, +.>Representing one-dimensional convolutional encoding of the position timing input vector; the second neighborhood scale feature extraction unit 322 is configured to: performing one-dimensional convolution coding on the position time sequence input vector by using a second convolution layer of the multi-scale neighborhood feature extraction module according to the following one-dimensional convolution formula to obtain a second neighborhood scale position time sequence associated feature vector; wherein, the formula is: />Wherein->Is the second convolution kernelWidth in direction, ++>For a second convolutionNuclear parameter vector,/->For a local vector matrix operating with a convolution kernel function, < ->For the size of the second one-dimensional convolution kernel, +.>Representing the position timing input vector, +.>Representing one-dimensional convolutional encoding of the position-timing input vector.
Specifically, during the operation of the CNC-based die-cutting machining control system 300, the temperature-pressure time-sequence association module 330 is configured to arrange the temperature data and the pressure data of the monitored cutting machine tool at the plurality of predetermined time points into a temperature time-sequence input vector and a pressure time-sequence input vector according to a time dimension, and then perform association encoding on the temperature time-sequence input vector and the pressure time-sequence input vector to obtain a temperature-pressure full-time-sequence association matrix. That is, as for the temperature data and the pressure data of the monitored cutting machine, it is considered that there is a considerable correlation in time series between the temperature and the pressure, that is, the time series change characteristics between the temperature data and the pressure data of the monitored cutting machine are mutually affected. Therefore, it is necessary to extract the correlation characteristics of both data. Specifically, firstly, the temperature data and the pressure data of the monitored cutting machine tool at a plurality of preset time points are respectively arranged into a temperature time sequence input vector and a pressure time sequence input vector according to a time dimension, and then the temperature time sequence input vector and the pressure time sequence input vector are subjected to association coding to obtain a temperature-pressure full-time sequence association matrix. In a specific example, the correlation encoding of the temperature timing input vector and the pressure timing input vector to obtain a temperature-pressure full timing correlation matrix includes: the temperature time sequence input vector and the pressure time sequence input are input according to the following formula Carrying out association coding on the vectors to obtain a temperature-pressure full-time-sequence association matrix; wherein, the formula is:wherein->Representing the temperature timing input vector, +.>A transpose vector representing the temperature timing input vector, < >>Representing the pressure timing input vector, +.>Representing the temperature-pressure full-time-series correlation matrix,/->Representing vector multiplication.
The associative coding (Associative Encoding) is an unsupervised learning method that is based on mapping input data onto codewords (codes) in a low-dimensional space such that similar input data are closer in codeword space and dissimilar input data are farther in codeword space. The main idea of the associated coding is that the code words can effectively represent the structural information of the input data through the self-adaptive learning process, so that the compression and dimension reduction of the data are realized. The implementation process of the association code can be divided into two stages: a learning phase and a coding phase. In the learning phase, a linear transformation matrix is learned by minimizing reconstruction errors, and the input data is mapped into codeword space. In the encoding stage, the input data is multiplied by the learned linear transformation matrix to obtain the corresponding codeword representation. The distance between codewords may be measured using a metric such as euclidean distance or cosine similarity. The association code may be applied to various fields such as image processing, natural language processing, recommendation systems, etc. In image processing, associative coding can be used for compression and dimension reduction of images; in natural language processing, associative coding can be used for word vector learning and text classification; in a recommendation system, association codes may be used for user and item representations and recommendations.
Specifically, during the operation of the CNC-based die cutting control system 300, the feature-related encoding module 340 is configured to perform feature matrix division on the temperature-pressure full-time-sequence-related matrix into a sequence of feature sub-matrices, and then obtain a temperature-pressure global context-related feature vector by a context encoder including an embedded layer. In the technical scheme of the application, the characteristic mining of the temperature-pressure full-time-sequence correlation matrix is performed by using a convolutional neural network model with excellent performance in the aspect of local implicit correlation characteristic extraction, but the method of pure CNN is difficult to learn explicit global and remote semantic information interaction due to the inherent limitation of convolution operation. Therefore, in order to improve the expression capability of the time sequence collaborative implicit small-scale associated feature of the temperature data and the pressure data, in the technical scheme of the application, the temperature-pressure full-time sequence associated matrix is divided into sequences of feature submatrices, and then the sequences are encoded in a context encoder comprising an embedded layer, so that global context associated feature information based on global context Wen Shixu between the temperature data and the pressure data in a time dimension is extracted, and a temperature-pressure global context associated feature vector is obtained.
Accordingly, in one possible implementation manner, after the temperature-pressure full-time sequence correlation matrix is subjected to feature matrix division into a sequence of feature submatrices, a context encoder including an embedded layer is used to obtain a temperature-pressure global context correlation feature vector, which includes: mapping each feature sub-matrix in the sequence of feature sub-matrices into an embedded vector by using an embedded layer of the context encoder comprising the embedded layer to obtain a sequence of embedded vectors; performing global context semantic coding on the sequence of embedded vectors based on a converter thought by using a converter of the context encoder comprising an embedded layer to obtain a plurality of global context semantic feature vectors; and cascading the plurality of global context semantic feature vectors to obtain the temperature-pressure global context associated feature vector.
Wherein the contextual semantic encoder (Contextual Semantic Encoder) is a technique for natural language processing that can convert input text into a vector representation for subsequent processing. The principle is based on a model such as a Recurrent Neural Network (RNN) or a variant long and short time memory network (LSTM) in deep learning, and a vector representation is obtained by encoding the input text word by word. In this process, the model takes into account the context information, i.e., the influence of the preceding and following words on the current word, thereby better capturing the semantic information of the text. Specifically, the context semantic encoder converts each word in the input text into a vector representation, and then processes the vector representation through a model such as a recurrent neural network or a long and short time memory network to finally obtain an overall vector representation. This vector representation can be used in text classification, emotion analysis, machine translation, etc. In general, the contextual semantic encoder is a technique that can translate natural language into a vector representation that can help computers better understand and process natural language. The embedded layer is part of a context semantic encoder for converting input text into a vector representation. The following is the principle of the embedded layer.
The embedded layer has the main function of converting the input discretized text into a continuous vector representation for subsequent processing and analysis. In particular, the embedding layer maps each word to a vector representation, which can be regarded as the position of the word in semantic space. In general, the purpose of the embedding layer is to convert the one-hot vector representation of the word into a continuous vector representation for subsequent processing and analysis. Through the embedding layer, we can represent the entered text as a matrix, where each row represents an embedded vector representation of a word. This matrix can be used as input to the context semantic encoder for subsequent processing and analysis.
Specifically, during operation of the CNC-based die cutting machining control system 300, the feature fusion module 350 is configured to fuse the temperature-pressure global context-related feature vector and the position-time-related feature vector to obtain a classification feature vector. That is, the temperature-pressure global context correlation feature vector and the position time sequence time correlation feature vector are fused, so that the time sequence global context cooperative correlation feature information based on the temperature and the pressure and the time sequence multi-scale dynamic change feature information of the position data are fused, and the time sequence cooperative correlation feature and the position multi-scale dynamic feature classification feature vector fused with the temperature-pressure are obtained. In particular, in the technical scheme of the application, the temperature-pressure global context correlation feature vector expresses the context semantic correlation feature of the temperature-pressure full-time correlation value under each process-remote correlation part, and the position time sequence correlation feature vector expresses the multi-scale time sequence neighborhood correlation feature of the position data, so that although the source data are time sequence data, the source data are different in the extraction direction of the time sequence semantic feature, and therefore, the characteristic semantic distribution of the source data is also different. Thus, when the classification feature vector is obtained by fusing the temperature-pressure global context-associated feature vector and the position time-sequence-associated feature vector, it is desirable to be able to enhance the fusion effect of the temperature-pressure global context-associated feature vector and the position time-sequence-associated feature vector on the feature semantic level.
Accordingly, in one possible implementation, the temperature-pressure global context associated feature vectorAnd the position-time-sequence-time-associated feature vector +.>Performing deep space encapsulation semantic matching fusion to obtain the classification feature vector, for example, marked as +.>Wherein the classification feature vector +.>The concrete steps are as follows:wherein->Is the temperature-pressure global context associated feature vector,/for>Is the position time sequence time correlation characteristic vector, < >>Is the classification feature vector,/->And->Representing the first and second norms of the vector, respectively, ">And->The weight and bias super-parameters are respectively given,representing a per-position distance matrix between the temperature-pressure global context-associated feature vector and the position-time-associated feature vector, and +.>Is a unitary matrix->、 />、/>Representing addition by location, subtraction by location, and multiplication by location, respectively. Here, the global context-dependent feature vector +.>And the position-time-sequence-time-associated feature vector +.>The semantic expression is packaged into a deep space, so that fine-grained features in the overall distribution of the feature vector simultaneously comprise low-level semantic distribution and high-level semantic distribution, thereby, through the deep space packaging semantic matching fusion, the matching of semantic levels of a classification mode layer can be performed through balancing the low-level semantic distribution and the high-level semantic distribution, so as to realize the semantic controlled compiling fusion of the features in the feature space, and further, the temperature-pressure global context associated feature vector # -is obtained >And the position-time-sequence-time-associated feature vector +.>Semantic collaboration in the feature fusion space improves the optimized classification feature vector +.>A global context-dependent feature vector for said temperature-pressure>And the position-time-sequence-time-associated feature vector +.>The semantic fusion effect of the classification feature vector is improved, so that the accuracy of the classification result obtained by the classification feature vector through the classifier is improved. Thus, the fault can be accurately pre-warned, and then the fault is liftedHigh efficiency and accuracy of die cutting processing, and optimized manufacturing process and product quality.
Specifically, during operation of the CNC-based die cutting machining control system 300, the fault pre-warning module 360 is configured to pass the classification feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether a fault alarm prompt is generated. That is, the classification feature vector is passed through a classifier to obtain a classification result indicating whether a fault alert is generated. Specifically, in the classification process of the classifier, as shown in fig. 4, the fault early-warning module 360 includes: a full-connection encoding unit 361, configured to perform full-connection encoding on the classification feature vector by using multiple full-connection layers of the classifier to obtain an encoded classification feature vector; and a classification result generating unit 362, configured to pass the encoded classification feature vector through a Softmax classification function of the classifier to obtain the classification result. In a specific example of the present application, the labeling of the classifier includes generating a fault alert (first labeling) and not generating a fault alert (second labeling), wherein the classifier determines to which classification label the classification feature vector belongs by a soft maximum function. It should be noted that the first tag p1 and the second tag p2 do not include a manually set concept, and in fact, during the training process, the computer model does not have a concept of "whether to generate a fault alarm prompt", which is simply that there are two kinds of classification tags and the probability that the output feature is under the two classification tags, that is, the sum of p1 and p2 is one. Therefore, the classification result of whether the fault alarm prompt is generated is actually converted into the classified probability distribution conforming to the natural rule through classifying the labels, and the physical meaning of the natural probability distribution of the labels is essentially used instead of the language text meaning of whether the fault alarm prompt is generated. It should be understood that, in the technical scheme of the application, the classification label of the classifier is a control strategy label for generating a fault alarm prompt, so that after the classification result is obtained, the accurate early warning can be performed on the fault based on the classification result, and the efficiency and the accuracy of die cutting processing are improved.
Wherein the classifier is a machine learning model for classifying data into different categories. The principle of the classifier is to learn a classification function or decision rule based on sample features and class labels in the training dataset, for classifying new data. The training process of the classifier is to learn and optimize the training data set to obtain the optimal classification function or decision rule so as to predict the classification of new data to the greatest extent. In summary, a classifier is a model based on training data learning that can be used to classify new data into known classes. Different classifiers have different principles and implementations, and an appropriate classifier can be selected according to a specific application scenario.
In summary, the CNC-based die-cut control system 300 according to the embodiment of the present application is illustrated, which uses a neural network model based on deep learning to mine out the time sequence of the position, temperature and pressure of the cutting machine tool in the die-cut process, so as to accurately pre-warn faults, thereby improving the efficiency and accuracy of the die-cut process, and optimizing the manufacturing process and product quality.
As described above, the CNC-based die cutting machining control system according to the embodiment of the present application may be implemented in various terminal devices. In one example, the CNC-based die cutting machining control system 300 according to embodiments of the present application may be integrated into the terminal device as one software module and/or hardware module. For example, the CNC-based die cutting machining control system 300 may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the CNC-based die cutting control system 300 may likewise be one of a number of hardware modules of the terminal device.
Alternatively, in another example, the CNC-based die cutting process control system 300 and the terminal device may also be separate devices, and the CNC-based die cutting process control system 300 may be connected to the terminal device via a wired and/or wireless network and communicate the interaction information in accordance with a agreed data format.
Further, a CNC-based die cutting machining control method is also provided.
Fig. 5 is a flowchart of a CNC-based die cutting control method according to an embodiment of the present application. As shown in fig. 5, the CNC-based die cutting control method according to an embodiment of the present application includes: s110, acquiring position data, temperature data and pressure data of a monitored cutting machine tool at a plurality of preset time points in a preset time period; s120, arranging the position data of the monitored cutting machine tool into position time sequence input vectors according to time, and then obtaining position time sequence associated feature vectors through a multi-scale neighborhood feature extraction module; s130, after arranging temperature data and pressure data of the monitored cutting machine tool at a plurality of preset time points into a temperature time sequence input vector and a pressure time sequence input vector according to a time dimension, performing association coding on the temperature time sequence input vector and the pressure time sequence input vector to obtain a temperature-pressure full-time sequence association matrix; s140, dividing the temperature-pressure full-time sequence correlation matrix into sequences of feature submatrices by a feature matrix, and then obtaining a temperature-pressure global context correlation feature vector by a context encoder comprising an embedded layer; s150, fusing the temperature-pressure global context correlation feature vector and the position time sequence correlation feature vector 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 for indicating whether a fault alarm prompt is generated or not.
Fig. 6 is a schematic diagram of a scenario of a CNC-based die cutting control system according to an embodiment of the present application. As shown in fig. 6, in this application scenario, position data of the monitored cutting machine tool at a plurality of predetermined time points within a predetermined period is acquired by a position sensor (e.g., V1 as illustrated in fig. 1), temperature data of the monitored cutting machine tool at a plurality of predetermined time points within a predetermined period is acquired by a temperature sensor (e.g., V2 as illustrated in fig. 1), and pressure data of the monitored cutting machine tool at a plurality of predetermined time points within a predetermined period is acquired by a pressure sensor (e.g., V3 as illustrated in fig. 1). The data is then input to a server (e.g., S in fig. 1) deployed with a CNC-based die-cut control algorithm, where the server is capable of processing the input data with the CNC-based die-cut control algorithm to generate a classification result indicating whether a fault alert is generated.
The foregoing description of the embodiments of the present disclosure has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (10)
1. A CNC-based die cutting control system, comprising: the data acquisition module is used for acquiring position data, temperature data and pressure data of the monitored cutting machine tool at a plurality of preset time points in a preset time period; the position time sequence change feature extraction module is used for arranging the position data of the monitored cutting machine tool into position time sequence input vectors according to time and then obtaining position time sequence correlation feature vectors through the multi-scale neighborhood feature extraction module; the temperature-pressure time sequence association module is used for respectively arranging the temperature data and the pressure data of the monitored cutting machine tool at a plurality of preset time points into a temperature time sequence input vector and a pressure time sequence input vector according to a time dimension, and then carrying out association coding on the temperature time sequence input vector and the pressure time sequence input vector to obtain a temperature-pressure full-time sequence association matrix; the feature association coding module is used for dividing the temperature-pressure full-time sequence association matrix into sequences of feature submatrices by a feature matrix, and obtaining a temperature-pressure global context association feature vector through a context encoder comprising an embedded layer; the feature fusion module is used for fusing the temperature-pressure global context correlation feature vector and the position time sequence time correlation feature vector to obtain a classification feature vector; and the fault early warning module is used for enabling the classification feature vector to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fault alarm prompt is generated or not.
2. The CNC-based die cutting machining control system of claim 1, wherein the multi-scale neighborhood feature extraction module comprises: the device comprises a first convolution layer, a second convolution layer parallel to the first convolution layer and a multi-scale feature fusion layer connected with the first convolution layer and the second convolution layer, wherein the first convolution layer uses a one-dimensional convolution kernel with a first length, and the second convolution layer uses a one-dimensional convolution kernel with a second length.
3. The CNC-based die cutting machining control system of claim 2, wherein the position timing variation feature extraction module comprises: a first neighborhood scale feature extraction unit, configured to input the position timing input vector into a first convolution layer of the multi-scale neighborhood feature extraction module to obtain a first neighborhood scale position timing associated feature vector, where the first convolution layer has a first one-dimensional convolution kernel with a first length; a second neighborhood scale feature extraction unit configured to input the position timing input vector to a second convolution layer of the multi-scale neighborhood feature extraction module to obtain a second neighborhood scale position timing associated feature vector, where the second convolution layer has a second one-dimensional convolution kernel of a second length, and the first length is different from the second length; the multi-scale cascading unit is used for cascading the first neighborhood scale position time sequence association feature vector and the second neighborhood scale position time sequence association feature vector to obtain the position time sequence association feature vector, wherein the first neighborhood scale feature extraction unit is used for: extracting a model using the multi-scale neighborhood feature The first convolution layer of the block carries out one-dimensional convolution coding on the position time sequence input vector by using the following one-dimensional convolution formula to obtain a first neighborhood scale position time sequence associated feature vector; wherein, the formula is:wherein->For the first convolution kernel at->Width in direction, ++>For the first convolution kernel parameter vector, +.>For a local vector matrix operating with a convolution kernel function, < ->For the size of the first one-dimensional convolution kernel, +.>Representing the position timing input vector, +.>Representing one-dimensional convolutional encoding of the position timing input vector; the second neighborhood scale feature extraction unit is configured to: performing one-dimensional convolution coding on the position time sequence input vector by using a second convolution layer of the multi-scale neighborhood feature extraction module according to the following one-dimensional convolution formula to obtain a second neighborhood scale position time sequence associated feature vector; wherein, the formula is:wherein->For the second convolution kernel>Width in direction, ++>For a second convolution kernel parameter vector, +.>For a local vector matrix operating with a convolution kernel function, < ->For the size of the second one-dimensional convolution kernel, +.>Representing the position timing input vector, +.>Representing one-dimensional convolutional encoding of the position-timing input vector.
4. The CNC-based die cutting machining control system of claim 3, wherein the temperature-pressure timing correlation module is configured to: performing association coding on the temperature time sequence input vector and the pressure time sequence input vector by using the following formula to obtain a temperature-pressure full-time sequence association matrix; wherein, the formula is:wherein->Representing the temperature timing input vector, +.>A transpose of the temperature timing input vector,/>representing the pressure timing input vector, +.>Representing the temperature-pressure full-time-series correlation matrix,/->Representing vector multiplication.
5. The CNC-based die cutting machining control system of claim 4, wherein the feature-related encoding module comprises: a matrix embedding unit, configured to map each feature sub-matrix in the sequence of feature sub-matrices into an embedded vector by using an embedding layer of the context encoder including the embedding layer, so as to obtain a sequence of embedded vectors; a context coding unit, configured to perform global context semantic coding on the sequence of embedded vectors using the converter of the context encoder including the embedded layer, where the global context semantic coding is based on a converter thought, so as to obtain a plurality of global context semantic feature vectors; and a cascade unit, configured to cascade the plurality of global context semantic feature vectors to obtain the temperature-pressure global context associated feature vector.
6. The CNC-based die cutting machining control system of claim 5, wherein the context encoding unit comprises: a query vector construction subunit, configured to perform one-dimensional arrangement on the sequence of the embedded vectors to obtain a global feature vector; a self-attention subunit, configured to calculate a product between the global feature vector and a transpose vector of each embedded vector in the sequence of embedded vectors to obtain a plurality of self-attention correlation matrices; the normalization subunit is used for respectively performing normalization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of normalized self-attention correlation matrices; the attention calculating subunit is used for obtaining a plurality of probability values through a Softmax classification function by each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; an attention applying subunit, configured to weight each embedded vector in the sequence of embedded vectors with each probability value in the plurality of probability values as a weight to obtain the plurality of context semantic feature vectors; and a concatenation subunit, configured to concatenate the plurality of context semantic feature vectors to obtain the global context semantic feature vector.
7. The CNC-based die cutting machining control system of claim 6, wherein the feature fusion module is configured to: carrying out deep space encapsulation semantic matching fusion on the temperature-pressure global context associated feature vector and the position time sequence associated feature vector by using the following fusion formula to obtain the classification feature vector; wherein, the fusion formula is:wherein->Is the temperature-pressure global context associated feature vector,/for>Is the position time sequence time correlation characteristic vector, < >>Is the classification feature vector,/->And->Representing the first and second norms of the vector, respectively, ">And->Respectively weight and bias superparameter, +.>Representing a per-position distance matrix between the temperature-pressure global context-associated feature vector and the position-time-associated feature vector, and +.>Is a unitary matrix->、 />、/>Representing addition by location, subtraction by location, and multiplication by location, respectively.
8. The CNC-based die cutting machining control system of claim 7, wherein the fault pre-warning module comprises: the full-connection coding unit is used for carrying out full-connection coding on the classification characteristic vectors by using a plurality of full-connection layers of the classifier so as to obtain coded classification characteristic vectors; and the classification result generating unit is used for passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
9. A CNC-based die cutting control method, comprising: acquiring position data, temperature data and pressure data of a monitored cutting machine tool at a plurality of preset time points in a preset time period; the position data of the monitored cutting machine tool are arranged into position time sequence input vectors according to time, and then the position time sequence input vectors are obtained through a multi-scale neighborhood feature extraction module; after arranging the temperature data and the pressure data of the monitored cutting machine tool at a plurality of preset time points into a temperature time sequence input vector and a pressure time sequence input vector according to a time dimension, performing association coding on the temperature time sequence input vector and the pressure time sequence input vector to obtain a temperature-pressure full-time sequence association matrix; after the temperature-pressure full-time sequence correlation matrix is subjected to feature matrix division into sequences of feature submatrices, a context encoder comprising an embedded layer is used for obtaining a temperature-pressure global context correlation feature vector; fusing the temperature-pressure global context correlation feature vector and the position time sequence correlation feature vector to obtain a classification feature vector; and passing the classification feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fault alarm prompt is generated or not.
10. The CNC-based die cutting process control method according to claim 9, wherein fusing the temperature-pressure global context-related feature vector and the position-time-series time-related feature vector to obtain a classification feature vector, comprises: carrying out deep space encapsulation semantic matching fusion on the temperature-pressure global context associated feature vector and the position time sequence associated feature vector by using the following fusion formula to obtain the classification feature vector; wherein, the fusion formula is:wherein->Is the temperature-pressure global context associated feature vector,/for>Is the position time sequence time correlation characteristic vector, < >>Is the classification feature vector,/->And->Representing the first and second norms of the vector, respectively, ">And->Respectively weight and bias superparameter, +.>Representing a per-position distance matrix between the temperature-pressure global context-associated feature vector and the position-time-associated feature vector, and +.>Is a unitary matrix->、 />、/>Representing addition by location, subtraction by location, and multiplication by location, respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310690429.7A CN116765925A (en) | 2023-06-12 | 2023-06-12 | CNC-based die cutting machining control system and method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310690429.7A CN116765925A (en) | 2023-06-12 | 2023-06-12 | CNC-based die cutting machining control system and method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116765925A true CN116765925A (en) | 2023-09-19 |
Family
ID=88010820
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310690429.7A Pending CN116765925A (en) | 2023-06-12 | 2023-06-12 | CNC-based die cutting machining control system and method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116765925A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117087023A (en) * | 2023-10-17 | 2023-11-21 | 杭州泓芯微半导体有限公司 | Double-station linear cutting machine and control method thereof |
| CN117102988A (en) * | 2023-10-16 | 2023-11-24 | 杭州泓芯微半导体有限公司 | Centerless grinding machine and control method thereof |
| CN117252926A (en) * | 2023-11-20 | 2023-12-19 | 南昌工控机器人有限公司 | Mobile phone shell auxiliary material intelligent assembly control system based on visual positioning |
| CN117315749A (en) * | 2023-09-25 | 2023-12-29 | 惠州市沃生照明有限公司 | Intelligent light regulation and control method and system for desk lamp |
-
2023
- 2023-06-12 CN CN202310690429.7A patent/CN116765925A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117315749A (en) * | 2023-09-25 | 2023-12-29 | 惠州市沃生照明有限公司 | Intelligent light regulation and control method and system for desk lamp |
| CN117102988A (en) * | 2023-10-16 | 2023-11-24 | 杭州泓芯微半导体有限公司 | Centerless grinding machine and control method thereof |
| CN117102988B (en) * | 2023-10-16 | 2024-01-05 | 杭州泓芯微半导体有限公司 | Centerless grinding machine and control method thereof |
| CN117087023A (en) * | 2023-10-17 | 2023-11-21 | 杭州泓芯微半导体有限公司 | Double-station linear cutting machine and control method thereof |
| CN117252926A (en) * | 2023-11-20 | 2023-12-19 | 南昌工控机器人有限公司 | Mobile phone shell auxiliary material intelligent assembly control system based on visual positioning |
| CN117252926B (en) * | 2023-11-20 | 2024-02-02 | 南昌工控机器人有限公司 | Mobile phone shell auxiliary material intelligent assembly control system based on visual positioning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116765925A (en) | CNC-based die cutting machining control system and method thereof | |
| CN116627708B (en) | Storage fault analysis system and method thereof | |
| CN116245513B (en) | Automatic operation and maintenance system and method based on rule base | |
| CN115456789B (en) | Abnormal transaction detection method and system based on transaction pattern recognition | |
| CN117116498A (en) | Mobile ward-round data processing system and method thereof | |
| CN116151845A (en) | Product full life cycle management system and method based on industrial Internet of things technology | |
| CN116703642A (en) | Intelligent management system of product manufacturing production line based on digital twin technology | |
| CN116929815A (en) | Equipment working state monitoring system and method based on Internet of things | |
| CN116629977A (en) | Electronic commerce decision method and system based on big data | |
| CN116482524A (en) | Power transmission and distribution switch state detection method and system | |
| CN116299051B (en) | Overcurrent short circuit detection circuit and detection protection system | |
| CN117231590A (en) | Fault prediction system and method for hydraulic system | |
| Luo et al. | An ELM-embedded deep learning based intelligent recognition system for computer numeric control machine tools | |
| Yassaie et al. | Data-Driven Fault Classification in Large-Scale Industrial Processes Using Reduced Number of Process Variables | |
| CN116757773A (en) | Clothing electronic commerce sales management system and method thereof | |
| CN116954113A (en) | Intelligent robot driving sensing intelligent control system and method thereof | |
| CN116363441A (en) | Pipeline corrosion detection system with marking function | |
| CN115762667A (en) | Chemical reaction type identification method and device and computer equipment | |
| CN116975731B (en) | Cross-domain cutter damage monitoring method and system based on transfer learning | |
| Zhang et al. | Deep Trajectory Similarity Model: A Fast Method for Trajectory Similarity Computation | |
| CN116561814B (en) | Textile chemical fiber supply chain information tamper-proof method and system thereof | |
| CN116821745B (en) | Control method and system of intelligent linear cutting slow wire-moving equipment | |
| CN117151223B (en) | Multi-modal entity identification and relation extraction method based on learning prompt | |
| CN117784710B (en) | Remote state monitoring system and method for numerical control machine tool | |
| Jang et al. | Multi-instance attention network for anomaly detection from multivariate time series |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |