Stranding tension monitoring system for a stranding machine and method thereof

Abstract

The application discloses a stranding tension monitoring system and method of a stranding machine. The system collects tension values of multiple cables through a tension sensor, introduces data processing and analysis algorithms at a rear end to perform time sequence analysis on the tension values of the multiple cables, and determines whether the tension state of each cable is normal. In this way, the system can realize automatic monitoring of the cable tension, reduce the need for manual intervention, improve the monitoring and control capability of the cable tension in the cable manufacturing process, and thus ensure the uniformity of the cable tension in the stranding process and improve the stranding quality and reliability of the cable.

Description

Technical Field

[0001] The present application relates to the field of intelligent monitoring, and more specifically, to a wire stranding tension monitoring system and method for a wire stranding machine. Background Art

[0002] A stranding machine, a device that twists multiple cables into a single strand, is widely used in fields such as power, communications, and aviation. The quality of the stranded wire directly impacts the performance and lifespan of the stranded wire. A key factor in stranding quality is the uniformity of the tension between the individual cables. Excessive or insufficient tension in a single cable can lead to uneven stranding and even breakage or loosening. Therefore, real-time monitoring and control of the stranding tension in the stranding machine is crucial for improving stranding quality.

[0003] Existing wire stranding machine tension monitoring systems typically use a traditional threshold determination method, comparing the tension value of each cable at a certain point in time with preset upper and lower limits. If the tension exceeds the range, the cable is considered to have abnormal tension. However, while this method is simple and easy to implement, it has the following shortcomings: First, it fails to fully utilize the temporal variation characteristics of tension data and ignores the temporal characteristics of tension changes; second, it fails to consider the mutual influence between individual cables and ignores the correlation between tension data; third, it cannot adapt to different types and specifications of cables, requiring manual setting of different threshold parameters, which is inefficient and prone to errors.

[0004] Therefore, an optimized stranding tension monitoring system for a stranding machine is desired. Summary of the Invention

[0005] In order to solve the above-mentioned technical problems, the present application is proposed. The embodiment of the present application provides a wire stranding tension monitoring system and method for a wire stranding machine. After collecting the tension values ​​of multiple cables through a tension sensor, a data processing and analysis algorithm is introduced at the back end to perform a time series analysis of the tension values ​​of multiple cables, so as to determine whether the tension state of each cable is normal. In this way, automatic monitoring of cable tension can be achieved, the need for manual intervention is reduced, and the ability to monitor and control cable tension during the cable manufacturing process is improved, thereby ensuring the uniformity of cable tension during the stranding process and improving the stranding quality and cable reliability.

[0006] According to one aspect of the present application, a wire stranding tension monitoring system for a wire stranding machine is provided, comprising:

[0007] The stranding machine body is used to strand multiple cables into one stranded wire;

[0008] The tension sensor is used to detect the tension value of each cable at a plurality of predetermined time points within a predetermined time period;

[0009] The data collector is used to receive and process the tension values ​​of the plurality of cables collected by the tension sensor at a plurality of predetermined time points within a predetermined time period to obtain an output result, wherein the output result is used to determine whether the tension state of each cable is normal;

[0010] The controller is used to control the operating parameters of the stranding machine body according to the output results of the data collector.

[0011] According to another aspect of the present application, a method for monitoring the tension of a stranded wire of a stranding machine is provided, comprising:

[0012] Arranging the tension values ​​of the plurality of cables at a plurality of predetermined time points within a predetermined time period according to a time dimension to obtain a plurality of cable tension time series input vectors;

[0013] Performing feature extraction on each of the plurality of cable tension time series input vectors using a time series feature extractor based on a deep neural network model to obtain a sequence of cable tension time series feature vectors;

[0014] Extracting a cable tension time series feature vector corresponding to the cable to be analyzed from the sequence of cable tension time series feature vectors as a query feature vector;

[0015] performing embedded feature analysis on the sequence of the cable tension time series feature vectors and the query feature vector to obtain an embedded query cable tension time series feature;

[0016] Based on the embedded query cable tension time sequence characteristics, it is determined whether the tension state of the cable to be analyzed is normal.

[0017] Compared to existing technologies, the present application provides a wire stranding machine wire tension monitoring system and method. After collecting the tension values ​​of multiple cables through tension sensors, the system introduces a data processing and analysis algorithm on the back end to perform a time-series analysis of the multiple cable tension values ​​to determine whether the tension state of each cable is normal. This enables automated monitoring of cable tension, reduces the need for manual intervention, and improves the ability to monitor and control cable tension during the cable manufacturing process, thereby ensuring the uniformity of cable tension during the stranding process and improving the quality and reliability of the stranded cables. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0019] Figure 1 is a block diagram of a wire stranding tension monitoring system for a wire stranding machine according to an embodiment of the present application;

[0020] Figure 2 1 is a system architecture diagram of a wire stranding tension monitoring system for a wire stranding machine according to an embodiment of the present application;

[0021] Figure 3 is a block diagram of a training phase of a wire stranding tension monitoring system for a wire stranding machine according to an embodiment of the present application;

[0022] Figure 4 A block diagram of a data collector in a wire stranding tension monitoring system of a wire stranding machine according to an embodiment of the present application;

[0023] Figure 5 4 is a flow chart of a method for monitoring the tension of a stranded wire of a stranding machine according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0025] As used in this application and the claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not intended to refer to the singular but may include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0026] Although the present application makes various references to certain modules in the system according to embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are illustrative only, and different aspects of the system and method can use different modules.

[0027] Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the various steps may be processed in reverse order or simultaneously, as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0028] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0029] Existing wire tension monitoring systems for stranding machines typically use a traditional threshold determination method. This method compares the tension value of each cable at a specific point in time with preset upper and lower limits. If the tension exceeds the range, the cable is considered to have abnormal tension. However, while simple and easy to implement, this method has the following shortcomings: First, it fails to fully utilize the temporal variation characteristics of tension data, ignoring the temporal characteristics of tension changes; second, it fails to consider the mutual influence between individual cables, ignoring the correlation between tension data; and third, it cannot adapt to cables of different types and specifications, requiring manual setting of different threshold parameters, resulting in low efficiency and prone to errors. Therefore, an optimized wire tension monitoring system for stranding machines is desired.

[0030] In the technical solution of the present application, a wire stranding tension monitoring system for a wire stranding machine is proposed. Figure 1 FIG. 1 is a block diagram of a wire stranding tension monitoring system for a wire stranding machine according to an embodiment of the present application. Figure 1 As shown, the stranding tension monitoring system 300 of the stranding machine according to the embodiment of the present application includes: a stranding machine body 310, wherein the stranding machine body is used to strand multiple cables into one strand; a tension sensor 320, wherein the tension sensor is used to detect the tension value of each of the cables at multiple predetermined time points within a predetermined time period; a data collector 330, wherein the data collector is used to receive and process the tension values ​​of the multiple cables at multiple predetermined time points within a predetermined time period collected by the tension sensor to obtain an output result, and the output result is used to determine whether the tension state of each of the cables is normal; a controller 340, wherein the controller is used to control the operating parameters of the stranding machine body according to the output result of the data collector.

[0031] In particular, the stranding machine body 310 is used to twist multiple cables into a single strand. It is worth noting that the primary function of a stranding machine is to twist two or more cables or wires together using a specific twisting pattern. This twisting pattern improves the flexibility, interference resistance, and wear resistance of the cables or wires. It is commonly used in industries such as wire and cable manufacturing, telecommunication cable manufacturing, and optical fiber manufacturing.

[0032] In particular, the tension sensor 320 is used to detect the tension value of each cable at multiple predetermined time points within a predetermined time period. Considering that in the actual process of monitoring the tension of a stranding machine, it is particularly important to monitor the tension data of multiple cables to comprehensively analyze the cable tension status, which is key to ensuring uniform tension among the cables. Therefore, in the technical solution of this application, first, the tension sensor is used to obtain the tension value of each cable at multiple predetermined time points within a predetermined time period.

[0033] It's worth noting that a tension sensor is a device used to measure the tension or pulling force in an object or system. It converts the physical quantity of tension into an electrical signal or other form of output. The operating principle of a tension sensor is based on the effect of tension on the sensor. When an object or system is subjected to tension, the tension sensor experiences a corresponding force or strain. Sensitive components within the sensor (such as strain gauges, resistors, and inductors) respond to this force or strain and convert it into an electrical signal. Tension sensors are widely used in various fields, such as textiles, printing, packaging, and automated production lines.

[0034] In particular, the data collector 330 is used to receive and process the tension values ​​of the plurality of cables collected by the tension sensor at a plurality of predetermined time points within a predetermined time period to obtain an output result, and the output result is used to determine whether the tension state of each cable is normal. In particular, in a specific example of the present application, if Figure 4 As shown, the data collector 330 includes: a tension time series arrangement module 331, which is used to arrange the tension values ​​of the multiple cables at multiple predetermined time points within a predetermined time period according to the time dimension to obtain multiple cable tension time series input vectors; a cable tension time series feature analysis module 332, which is used to perform feature extraction on the multiple cable tension time series input vectors respectively through a time series feature extractor based on a deep neural network model to obtain a sequence of cable tension time series feature vectors; a cable tension time series feature query module 333 to be analyzed, which is used to extract the cable tension time series feature vector corresponding to the cable to be analyzed from the sequence of cable tension time series feature vectors as a query feature vector; a feature embedded fusion module 334, which is used to perform embedded feature analysis on the sequence of cable tension time series feature vectors and the query feature vector to obtain an embedded query cable tension time series feature; a tension state detection module 335, which is used to determine whether the tension state of the cable to be analyzed is normal based on the embedded query cable tension time series feature.

[0035] Specifically, the tension time series arrangement module 331 is used to arrange the tension values ​​of the multiple cables at multiple predetermined time points within a predetermined time period according to the time dimension to obtain multiple cable tension time series input vectors. Considering that the tension values ​​of each cable have a time series dynamic change law in the time dimension, that is, the tension values ​​of each cable at multiple predetermined time points have a time series correlation feature. Therefore, in the technical solution of the present application, it is necessary to arrange the tension values ​​of the multiple cables at multiple predetermined time points within a predetermined time period according to the time dimension to obtain multiple cable tension time series input vectors, so as to respectively integrate the distribution information of the tension values ​​of each cable in the time series, so as to facilitate the subsequent tension time series feature analysis of each cable.

[0036] Specifically, the cable tension time series feature analysis module 332 is configured to perform feature extraction on each of the multiple cable tension time series input vectors using a time series feature extractor based on a deep neural network model to obtain a sequence of cable tension time series feature vectors. In other words, in the technical solution of the present application, each of the multiple cable tension time series input vectors is subjected to feature mining using a time series feature extractor based on a one-dimensional convolutional layer to extract time series feature information of the tension values ​​of each cable in the time dimension, thereby obtaining a sequence of cable tension time series feature vectors. More specifically, each layer of the one-dimensional convolutional layer-based temporal feature extractor performs the following operations on the input data in the forward pass of the layer: convolution processing on the input data to obtain a convolution feature map; pooling based on a feature matrix on the convolution feature map to obtain a pooled feature map; and nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the output of the last layer of the one-dimensional convolutional layer-based temporal feature extractor is a sequence of the cable tension temporal feature vectors, and the input of the first layer of the one-dimensional convolutional layer-based temporal feature extractor is the multiple cable tension temporal input vectors.

[0037] It's worth noting that the 1D convolutional layer is a common layer type in convolutional neural networks. The input to a 1D convolutional layer is a one-dimensional feature sequence, typically represented as a vector. The 1D convolutional layer extracts local features from the input sequence by applying a 1D convolution operation. The convolution operation uses a learnable filter (also called a convolution kernel or filter) that slides over the input sequence, computing the convolution result at each position. The main parameters of a 1D convolutional layer include the number of filters, the filter width, and the stride. The number of filters determines the depth of the output feature map, with each filter capturing different local features. The filter size defines the filter width, which determines the length of the input sequence covered by a convolution operation. The stride defines the step size of the filter's sliding movement over the input sequence, which determines the size of the output feature map. 1D convolutional layers typically also include activation functions and pooling operations. Activation functions introduce nonlinear transformations to increase the network's expressive power. Common activation functions include ReLU, Sigmoid, and Tanh. Pooling is used to reduce the size of feature maps and extract more significant features. Common pooling operations include max pooling and average pooling. One-dimensional convolutional layers are widely used in many tasks, including natural language processing, speech recognition, and time series analysis.

[0038] Specifically, the cable tension time sequence feature to be analyzed query module 333 is configured to extract the cable tension time sequence feature vector corresponding to the cable to be analyzed from the sequence of cable tension time sequence feature vectors as a query feature vector. In the stranding tension monitoring system of the stranding machine, in order to analyze the tension time sequence feature of each cable respectively to determine whether the tension state is normal, the cable tension time sequence feature vector corresponding to the cable to be analyzed needs to be further extracted from the sequence of cable tension time sequence feature vectors as a query feature vector. In this way, the cable tension time sequence feature vector of the cable to be analyzed can be compared with other cable tension time sequence feature vectors to determine the similarity or difference between them, thereby determining whether the tension state of the cable to be analyzed is normal.

[0039] Correspondingly, in a possible implementation, the cable tension time sequence feature vector corresponding to the cable to be analyzed can be extracted from the sequence of cable tension time sequence feature vectors as a query feature vector by the following steps, for example: preparing the sequence of cable tension time sequence feature vectors and the identifier or index of the cable to be analyzed; locating the feature vector corresponding to the cable to be analyzed in the sequence of cable tension time sequence feature vectors according to the identifier or index of the cable to be analyzed; extracting the located cable tension time sequence feature vector as a query feature vector; and outputting the query feature vector, which is the feature vector corresponding to the cable to be analyzed extracted from the sequence of cable tension time sequence feature vectors.

[0040] Specifically, the feature embedding fusion module 334 is configured to perform embedded feature analysis on the sequence of cable tension time sequence feature vectors and the query feature vector to obtain an embedded query cable tension time sequence feature. Considering that the tension time sequence features of the cables are correlated with each other, in order to make full use of the correlation to more accurately analyze the tension state of the cable to be analyzed, in the technical solution of the present application, the query feature vector and the sequence of cable tension time sequence feature vectors are further input into the feature embedding query module to embed the query feature vector and the sequence of cable tension time sequence feature vectors, so as to obtain an embedded query cable tension time sequence feature vector. It should be understood that the embedded query cable tension time sequence feature vector is obtained by comparing and matching the query feature vector with the cable tension time sequence feature vectors. In this way, by processing through the feature embedding query module, the tension time sequence feature of the cable to be analyzed can be comprehensively embedded based on the tension time sequence features of all the cables as a reference, thereby improving the tension time sequence semantic expression of the query feature vector under the constraint of the tension time sequence features of the cables.

[0041] Correspondingly, in a possible implementation, the query feature vector and the sequence of cable tension time series feature vectors can be input into a feature embedding query module to obtain an embedded query cable tension time series feature vector as the embedded query cable tension time series feature by the following steps, for example: inputting the query feature vector and the sequence of cable tension time series feature vectors; loading the feature embedding query module, which is used to embed the input feature vector sequence; performing feature embedding on the query feature vector: inputting the query feature vector into the feature embedding query module; the feature embedding query module converts the query feature vector into an embedded query feature vector; performing feature embedding on the sequence of cable tension time series feature vectors: inputting the sequence of cable tension time series feature vectors into the feature embedding query module one by one; the feature embedding query module converts each cable tension time series feature vector into an embedded query cable tension time series feature vector; repeating the above steps until all cable tension time series feature vectors are embedded into embedded query cable tension time series feature vectors; outputting the embedded query cable tension time series feature vector, which is obtained by converting the query feature vector and the sequence of cable tension time series feature vectors through the feature embedding query module.

[0042] Specifically, the tension state detection module 335 is configured to determine whether the tension state of the cable to be analyzed is normal based on the embedded query cable tension time series feature. That is, in the technical solution of the present application, the embedded query cable tension time series feature vector is input into a classifier to obtain a classification result, which is used to indicate whether the tension state of the cable to be analyzed is normal. That is, the classification is performed on the tension state time series feature related to the cable to be analyzed after the overall cable tension time series feature embedding expression, so as to determine whether the tension state of each cable is normal, and the operation parameters of the stranding machine are controlled according to the classification result to ensure the quality and uniformity of the stranding. In this way, the automatic monitoring of the cable tension can be realized, the need for manual intervention is reduced, and the monitoring and control capability of the cable tension in the cable manufacturing process is improved, thereby ensuring the uniformity of the cable tension in the stranding process and improving the stranding quality and the reliability of the cable. More specifically, the embedded query cable tension time series feature vector is input into a classifier to obtain a classification result, which is used to indicate whether the tension state of the cable to be analyzed is normal, including: using a plurality of fully connected layers of the classifier to perform fully connected coding on the embedded query cable tension time series feature vector to obtain a coded classification feature vector; and inputting the coded classification feature vector into a Softmax classification function of the classifier to obtain the classification result.

[0043] The classifier refers to a machine learning model or algorithm used to classify input data into different categories or labels. The classifier is part of supervised learning, which learns the mapping relationship from input data to output categories to perform the classification task.

[0044] A fully connected layer is a common layer type in neural networks. In a fully connected layer, each neuron is connected to all neurons in the previous layer, and each connection has a weight. This means that each neuron in a fully connected layer receives input from all neurons in the previous layer, performs a weighted sum of these inputs using the weights, and then passes the result to the next layer.

[0045] The Softmax classification function is a commonly used activation function for multi-classification problems. It converts each element of the input vector into a probability value between 0 and 1, where the sum of these probabilities equals 1. The Softmax function is often used in the output layer of neural networks and is particularly well-suited for multi-classification problems because it maps the network output into a probability distribution for each category. During training, the output of the Softmax function is used to calculate the loss function and update the network parameters through the backpropagation algorithm. It is worth noting that the output of the Softmax function does not change the relative size of the elements; it simply normalizes them. Therefore, the Softmax function does not change the characteristics of the input vector; it simply converts it into a probability distribution.

[0046] Specifically, the controller 340 is configured to control the operating parameters of the stranding machine based on the output of the data collector. In one example, the controller receives the output of the data collector and analyzes and processes the data. It can adjust the stranding machine's operating parameters based on pre-set algorithms and control strategies to achieve the desired stranding effect and quality control. By monitoring and controlling the stranding machine's operating parameters in real time, the controller can improve the stability, consistency, and efficiency of the stranding process, thereby enhancing production quality and reducing production costs.

[0047] Accordingly, in one possible implementation, the operating parameters of the stranding machine body can be controlled based on the output of the data collector through the following steps: for example, setting the data collector to monitor desired parameters. These parameters may include cable tension, line speed, temperature, etc.; starting the data collector and beginning to record data of the desired parameters; connecting the data collector to the stranding machine body to receive the collected data in real time; providing a control system in the stranding machine body to receive and process the data output by the data collector; defining appropriate algorithms or rules to control the operating parameters of the stranding machine body based on the data output by the data collector. These algorithms or rules can be based on preset thresholds, logical judgments, or other relevant conditions; receiving the data output by the data collector in real time and transmitting it to the control system for processing; the control system analyzing the data output by the data collector based on the preset algorithms or rules and determining the operating parameters of the stranding machine body that need to be adjusted; adjusting the operating parameters of the stranding machine body based on the instructions of the control system; and continuously monitoring the output of the data collector and adjusting the parameters as needed to ensure that the stranding machine body operates within the expected range.

[0048] It should be understood that before using the above-mentioned neural network model for inference, the temporal feature extractor based on the one-dimensional convolution layer, the feature embedding query module, and the classifier need to be trained. In other words, the wire stranding machine tension monitoring system 300 according to the present application also includes a training stage 400 for training the temporal feature extractor based on the one-dimensional convolution layer, the feature embedding query module, and the classifier.

[0049] Figure 3 FIG. 1 is a block diagram of the training phase of the stranding tension monitoring system of the stranding machine according to an embodiment of the present application. Figure 3As shown, the stranding tension monitoring system 300 of the stranding machine according to the embodiment of the present application includes: a training stage 400, including: a training data acquisition unit 410, for obtaining training data, wherein the training data includes training tension values ​​of multiple cables at multiple predetermined time points within a predetermined time period, and a real value of whether the tension state of the cable to be analyzed is normal; a training tension data time series arrangement unit 420, for arranging the training tension values ​​of the multiple cables at multiple predetermined time points within a predetermined time period according to the time dimension to obtain multiple training cable tension time series input vectors; a training tension time series feature extraction unit 430, for respectively passing the multiple training cable tension time series input vectors through the time series feature extractor based on the one-dimensional convolution layer to obtain a sequence of training cable tension time series feature vectors; a training cable tension time series feature extraction unit 440 for extracting the training cable tension time series feature vectors from the training cable tension time series feature vectors The cable tension time series feature vector corresponding to the cable to be analyzed is extracted from the sequence and trained as the training query feature vector; the training embedded query cable tension time series feature fusion unit 450 is used to pass the sequence of the training query feature vector and the training cable tension time series feature vector through the feature embedding query module to obtain the training embedded query cable tension time series feature vector; the feature optimization unit 460 is used to optimize the training embedded query cable tension time series feature vector position by position to obtain the optimized training embedded query cable tension time series feature vector; the classification loss unit 470 is used to pass the optimized training embedded query cable tension time series feature vector through the classifier to obtain the classification loss function value; the model training unit 480 is used to train the time series feature extractor based on the one-dimensional convolutional layer, the feature embedding query module and the classifier based on the classification loss function value and through the direction propagation of gradient descent.

[0050] Wherein, the classification loss unit is used to: use the classifier to process the optimized training embedded query cable tension time series feature vector to obtain a training classification result; and calculate the cross entropy loss function value between the training classification result and the true value of whether the tension state of the cable to be analyzed is normal as the classification loss function value.

[0051] In particular, in the technical solution of the present application, each cable tension time sequence feature vector in the sequence of cable tension time sequence feature vectors represents the time sequence correlation feature of the tension value of the corresponding cable, so that the cable tension time sequence feature vector corresponding to the cable to be analyzed is extracted from the sequence of cable tension time sequence feature vectors as a query feature vector, and the query feature vector and the sequence of cable tension time sequence feature vectors are input into the feature embedding query module, so that the time sequence correlation feature of the tension value of the cable to be analyzed can be dynamically constrained under the cable tension time sequence correlation feature of the whole sample domain. However, this will also make the embedded query cable tension time sequence feature vector have a cross-sample domain time sequence feature correlation distribution. Therefore, due to the non-smooth time sequence correlation feature distribution property of the embedded query cable tension time sequence feature vector across the whole sample domain, when the embedded query cable tension time sequence feature vector is classified and regressed by the classifier, the classification and regression efficiency of the cross-domain non-smooth feature distribution needs to be improved. Therefore, preferably, the applicant of the present application optimizes the embedded query cable tension time sequence feature vector position by position when it is classified and regressed by the classifier, which is specifically represented as:

[0052]

[0053] wherein v i is the feature value of the i-th position of the embedded query cable tension time sequence feature vector, is the global mean of all feature values of the embedded query cable tension time sequence feature vector, and v max is the maximum feature value of the embedded query cable tension time sequence feature vector, exp(·) represents the exponential operation, and v i is the optimized training embedded query cable tension time sequence feature vector. That is, by using the concept of regularization function of global distribution parameters, the above optimization represents the parameter vector of the global distribution of the embedded query cable tension time sequence feature vector, and uses the regularization expression of regression probability to simulate the cost function, so as to model the point-by-point regression characteristics of the weight matrix of the classifier based on the classification probability of the feature manifold representation of the embedded query cable tension time sequence feature vector in the high-dimensional feature space, to capture the parameter smooth optimization trajectory of the embedded query cable tension time sequence feature vector to be classified in the parameter space of the classifier model under the scene geometry of the high-dimensional feature manifold, and to improve the training efficiency of the embedded query cable tension time sequence feature vector under the classification probability regression of the classifier. In this way, the automatic monitoring of the cable tension state can be realized, the need for manual intervention is reduced, and the monitoring and control ability of the cable tension in the cable manufacturing process is improved, so as to ensure the uniformity of the cable tension in the stranding process and improve the stranding quality and the reliability of the cable.

[0054] As mentioned above, the stranding tension monitoring system 300 of the stranding machine according to the embodiments of the present application can be implemented in various wireless terminals, such as a server with the stranding tension monitoring algorithm of the stranding machine, etc. In one possible implementation, the stranding tension monitoring system 300 of the stranding machine according to the embodiments of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the stranding tension monitoring system 300 of the stranding machine can be a software module in the operating system of the wireless terminal, or can be an application program developed for the wireless terminal; of course, the stranding tension monitoring system 300 of the stranding machine can also be one of the many hardware modules of the wireless terminal.

[0055] Alternatively, in another example, the stranding tension monitoring system 300 of the stranding machine and the wireless terminal can also be separate devices, and the stranding tension monitoring system 300 of the stranding machine can be connected to the wireless terminal through a wired and / or wireless network, and transmit interactive information in a conventional data format.

[0056] Further, a stranding tension monitoring method of a stranding machine is also provided.

[0057] Figure 5 A flowchart of the stranding tension monitoring method of the stranding machine according to the embodiments of the present application is shown. As shown in Figure 5 the stranding tension monitoring method of the stranding machine according to the embodiments of the present application includes the steps of: S1, arranging the tension values of the plurality of cables at a plurality of predetermined time points within a predetermined time period respectively according to the time dimension to obtain a plurality of cable tension time sequence input vectors; S2, performing feature extraction on the plurality of cable tension time sequence input vectors respectively through a time sequence feature extractor based on a deep neural network model to obtain a sequence of cable tension time sequence feature vectors; S3, extracting the cable tension time sequence feature vector corresponding to the cable to be analyzed from the sequence of cable tension time sequence feature vectors as a query feature vector; S4, performing embedded feature analysis on the sequence of cable tension time sequence feature vectors and the query feature vector to obtain an embedded query cable tension time sequence feature; and S5, determining whether the tension state of the cable to be analyzed is normal based on the embedded query cable tension time sequence feature.

[0058] In summary, the stranding tension monitoring method of the stranding machine according to the embodiments of the present application is illustrated, which collects the tension values of the plurality of cables through the tension sensor, and then introduces data processing and analysis algorithms in the backend to perform time sequence analysis on the tension values of the plurality of cables, so as to determine whether the tension state of each cable is normal. In this way, the automatic monitoring of the cable tension can be realized, the need for manual intervention is reduced, and the monitoring and control capability of the cable tension in the cable manufacturing process is improved, thereby ensuring the uniformity of the cable tension in the stranding process, improving the stranding quality and the reliability of the cable.

[0059] Having described above several embodiments of the disclosure, any modifications and variations that fall within the scope of the described embodiments are also contemplated. It is also contemplated that the application covered by the claims extends to any alternative combination of claim elements not specifically disclosed. The use of the terms "preferably," "preferably," "preferred," and "has been preferred" in the description above indicates that the described feature is but one replacement for the term "the technology."

Claims

1. A wire stranding tension monitoring system for a wire stranding machine, characterized in that: include: Wire stranding machine body, tension sensor, data collector and controller; The stranding machine body is used to strand multiple cables into one stranded wire; The tension sensor is used to detect the tension value of each cable at a plurality of predetermined time points within a predetermined time period; The data collector is used to receive and process the tension values ​​of the plurality of cables collected by the tension sensor at a plurality of predetermined time points within a predetermined time period to obtain an output result, wherein the output result is used to determine whether the tension state of each cable is normal; The controller is used to control the operating parameters of the stranding machine body according to the output result of the data collector; Wherein, the data collector includes: a tension time sequence arrangement module, configured to arrange the tension values ​​of the plurality of cables at a plurality of predetermined time points within a predetermined time period according to a time dimension to obtain a plurality of cable tension time sequence input vectors; a cable tension time series feature analysis module, configured to extract features from each of the plurality of cable tension time series input vectors using a time series feature extractor based on a deep neural network model to obtain a sequence of cable tension time series feature vectors; a cable tension time series feature query module to be analyzed, configured to extract a cable tension time series feature vector corresponding to the cable to be analyzed from the sequence of cable tension time series feature vectors as a query feature vector; a feature embedded fusion module, configured to perform embedded feature analysis on the sequence of the cable tension time series feature vectors and the query feature vector to obtain an embedded query cable tension time series feature; The tension state detection module is used to determine whether the tension state of the cable to be analyzed is normal based on the embedded query cable tension time sequence characteristics.

2. The wire stranding machine wire stranding tension monitoring system according to claim 1, characterized in that: The temporal feature extractor based on the deep neural network model is a temporal feature extractor based on a one-dimensional convolutional layer.

3. The wire stranding machine wire stranding tension monitoring system according to claim 2, characterized in that: The feature embedded fusion module includes: passing the query feature vector and the sequence of the cable tension time series feature vector through a feature embedding query module to obtain an embedded query cable tension time series feature vector as the embedded query cable tension time series feature.

4. The wire stranding machine wire stranding tension monitoring system according to claim 3, characterized in that: The tension state detection module includes: passing the embedded query cable tension time series feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether the tension state of the cable to be analyzed is normal.

5. The wire stranding machine wire stranding tension monitoring system according to claim 4, characterized in that: It also includes a training module for training the one-dimensional convolutional layer-based temporal feature extractor, the feature embedding query module and the classifier.

6. The wire stranding machine wire stranding tension monitoring system according to claim 5, characterized in that: The training module includes: A training data acquisition unit, configured to acquire training data, wherein the training data includes training tension values ​​of a plurality of cables at a plurality of predetermined time points within a predetermined time period, and a true value of whether the tension state of the cable to be analyzed is normal; a training tension data time series arrangement unit, configured to arrange the training tension values ​​of the plurality of cables at a plurality of predetermined time points within a predetermined time period according to a time dimension to obtain a plurality of training cable tension time series input vectors; a training tension time series feature extraction unit, configured to pass the plurality of training cable tension time series input vectors respectively through the time series feature extractor based on the one-dimensional convolutional layer to obtain a sequence of training cable tension time series feature vectors; a training cable tension time series feature extraction unit for extracting a cable tension time series feature vector corresponding to the training cable to be analyzed from the sequence of training cable tension time series feature vectors as a training query feature vector; a training embedded query cable tension time series feature fusion unit, configured to pass the sequence of the training query feature vector and the training cable tension time series feature vector through the feature embedding query module to obtain a training embedded query cable tension time series feature vector; a feature optimization unit, configured to optimize the training embedded query cable tension time series feature vector position by position to obtain an optimized training embedded query cable tension time series feature vector; A classification loss unit, configured to embed the optimization training into the query cable tension time series feature vector through the classifier to obtain a classification loss function value; A model training unit is used to train the temporal feature extractor based on the one-dimensional convolution layer, the feature embedding query module and the classifier based on the classification loss function value and through gradient descent direction propagation.

7. The wire stranding machine wire stranding tension monitoring system according to claim 6, characterized in that: The classification loss unit is used to: Using the classifier to process the optimized training embedded query cable tension time series feature vector to obtain a training classification result: and A cross entropy loss function value between the training classification result and a true value of whether the tension state of the cable to be analyzed is normal is calculated as the classification loss function value.

8. A method for monitoring the tension of a stranded wire of a stranding machine, characterized in that: include: Arranging the tension values ​​of the plurality of cables at a plurality of predetermined time points within a predetermined time period according to the time dimension to obtain a plurality of cable tension time series input vectors; Performing feature extraction on each of the plurality of cable tension time series input vectors using a time series feature extractor based on a deep neural network model to obtain a sequence of cable tension time series feature vectors; Extracting a cable tension time series feature vector corresponding to the cable to be analyzed from the sequence of cable tension time series feature vectors as a query feature vector; performing embedded feature analysis on the sequence of the cable tension time series feature vectors and the query feature vector to obtain an embedded query cable tension time series feature; Based on the embedded query cable tension time sequence characteristics, it is determined whether the tension state of the cable to be analyzed is normal.

Images