Electromagnetic coil winding tension control method, device and equipment and medium

Abstract

The invention relates to the technical field of tension control in the material processing process. The electromagnetic coil winding tension control method comprises the steps that real-time state information is input into a wire rod type recognition model to obtain wire rod type information, and a corresponding tension adjusting model is called from a tension control model library based on the wire rod type information to adjust the winding tension of the electromagnetic coil. And inputting the real-time state information into the tension adjustment model to obtain tension adjustment parameter information, comparing the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information, executing tension adjustment operation according to the limited adjustment instruction information to obtain feedback tension information, and sending the feedback tension information to the server. The feedback tension information and the limited adjustment instruction information are compared and analyzed, a real-time tension adjustment result is obtained, and if the real-time tension adjustment result meets the target tension range, the current tension adjustment state is kept. The wire winding device has the effect of achieving accurate adjustment of the wire in the winding stage.

Description

Technical Field

[0001] The present invention relates to the technical field of tension control in a material processing process, and in particular to a method, device, equipment and medium for controlling the tension of an electromagnetic coil winding. Background Art

[0002] Existing electromagnetic coil winding tension control methods often use fixed tension settings or simple feedback mechanisms, such as using a tension sensor to measure and adjust the motor speed in real time. However, these methods often ignore the impact of different materials, wire diameters, and winding paths on tension adjustment, and lack the ability to deeply identify wire conditions and adaptively adjust multiple factors. Especially in high-precision or complex winding scenarios, single feedback adjustment cannot cope with conditions where the wire condition changes dramatically or winding accuracy is required. This can easily lead to delayed tension adjustment response and lagging control strategies, thereby reducing winding quality. Summary of the Invention

[0003] In order to achieve precise adjustment of the wire during the winding stage, the present application provides an electromagnetic coil winding tension control method, device, equipment and medium.

[0004] The above-mentioned invention objective of this application is achieved through the following technical solutions: A method for controlling the tension of an electromagnetic coil winding, comprising: Obtain real-time status information of the target wire during the winding process; Inputting the real-time status information into a wire type recognition model to obtain wire type information, and calling a corresponding tension adjustment model from a tension control model library based on the wire type information; Inputting the real-time state information into the tension adjustment model to obtain tension adjustment parameter information, and comparing the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information; performing a tension adjustment operation according to the limited adjustment instruction information to obtain feedback tension information, and comparing and analyzing the feedback tension information with the limited adjustment instruction information to obtain a real-time tension adjustment result; If the real-time tension adjustment result meets the target tension range, the current tension adjustment state is maintained.

[0005] By adopting the above technical solution, it is possible to obtain real-time status information of the target wire during the winding process, and accurately identify the material category, wire diameter specification and flexibility parameters in combination with the wire type recognition model, and call the corresponding tension adjustment model in the tension control model library based on the recognition result, so as to achieve dynamic adaptation and personalized matching of the tension adjustment strategy, further input the real-time status information into the tension adjustment model, output the tension adjustment parameter information for the current state, and introduce the preset tension safety boundary information for judgment and restriction, forming a restricted adjustment instruction to ensure the stability and safety of the adjustment process, and then execute the tension adjustment operation according to the restricted adjustment instruction and obtain feedback tension information, and form a real-time tension adjustment result by comparing and analyzing the feedback tension information with the adjustment instruction. When the adjustment result meets the target tension range, the current state is maintained without excessive intervention.

[0006] In a preferred example, the present application may be further configured as follows: inputting the real-time status information into a wire type recognition model to obtain wire type information includes: Performing sliding window processing on the tension change value in the real-time status information to obtain tension feature information; Performing section fitting processing on the bending state information in the real-time state information to obtain bending feature information; Combining the tension characteristic information with the bending characteristic information to form a wire behavior characteristic data set, and inputting the wire behavior characteristic data set into the wire type recognition model; The wire type recognition model performs multi-dimensional feature clustering and type mapping analysis on the wire behavior feature data group, and outputs the wire type information corresponding to the real-time status information.

[0007] By adopting the above technical solution, it is possible to extract short-term dynamic fluctuation characteristics based on sliding window processing of the tension change value in the real-time status information, and at the same time perform segment fitting on the bending status information to reflect the continuity and bending degree of the wire path. The above tension feature information and bending feature information are then combined to form a wire behavior feature data group with comprehensive characteristics of temporal changes and geometric morphology, and input it into the wire type recognition model. The model performs multi-dimensional feature clustering and type mapping on the input data group, thereby realizing accurate identification of wire types with different material properties, physical dimensions and dynamic behavior characteristics.

[0008] In a preferred example, the present application may be further configured as follows: calling a corresponding tension adjustment model from a tension control model library based on the wire type information includes: Using the material type parameter, wire diameter parameter, and flexibility parameter in the wire type information as matching conditions, the matching is performed in sequence with the index information of the candidate tension adjustment models in the tension control model library; performing correlation scoring processing on the candidate tension adjustment models that meet the matching condition to obtain a scoring result, and determining a target tension adjustment model according to the scoring result; According to the identification information of the target tension adjustment model, the model structure and parameter configuration corresponding to the target tension adjustment model are retrieved and extracted from the tension control model library to obtain the tension adjustment model.

[0009] By adopting the above technical solution, after obtaining the wire type information, the material type parameters, wire diameter parameters and flexibility parameters that are closely related to tension control can be extracted, and the index information of each candidate tension adjustment model in the tension control model library can be compared item by item as matching conditions, and a set of candidate models that meet the structural and performance requirements can be screened out. The candidate model set is then scored for correlation, and the target tension adjustment model that best meets the current wire characteristics and control requirements is selected based on the scoring results. Then, according to the identification information of the selected model, the corresponding structure definition and parameter configuration are retrieved from the model library, so as to accurately obtain the tension adjustment model adapted to the current winding task, ensuring that the adjustment strategy is highly consistent with the actual properties of the wire.

[0010] In a preferred example, the present application may be further configured as follows: inputting the real-time status information into the tension adjustment model to obtain tension adjustment parameter information includes: Performing data segmentation processing and value normalization processing on the real-time status information, and extracting the value change and change rate of the real-time status information in multiple time series intervals; The numerical change amount and the change rate are used as input data and input into the tension adjustment model to obtain the tension adjustment parameter information.

[0011] By adopting the above technical solution, the continuously acquired real-time status information can be divided into multiple time series intervals, and the key data change trends can be extracted in each interval, such as the increase or decrease amplitude of the tension value and its change speed per unit time. At the same time, the extracted original numerical data is normalized to eliminate the influence of different measurement units and data amplitude ranges on the consistency of subsequent model inputs, ensuring that each input feature participates in the model calculation at the same scale. The normalized numerical change and change rate are then input into the tension adjustment model. The model maps and analyzes the input features based on the historical training structure, thereby calculating the appropriate tension adjustment parameter information under the current winding state, realizing rapid adaptation of the adjustment strategy to the real-time state, and avoiding adjustment delays or misjudgments caused by drastic state changes.

[0012] In a preferred example, the present application may be further configured as follows: comparing the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information, including: Performing a range judgment on the tension target value in the tension adjustment parameter information, and comparing whether the tension target value is between the tension upper limit value and the tension lower limit value defined by the tension safety boundary information; When the tension target value exceeds the tension upper limit or is lower than the tension lower limit, based on the difference between the tension target value and the tension safety margin information, the tension adjustment parameter information is modified to generate the limited adjustment instruction; When the tension target value is between the tension upper limit value and the tension lower limit value defined by the tension safety margin information, the tension adjustment parameter information is used as the limited adjustment instruction information.

[0013] By adopting the above technical solution, the tension target value in the tension adjustment parameter information can be dynamically checked for compliance. After the system outputs the preliminary adjustment target value, it is further determined whether the target value is within the pre-set tension safety boundary range, where the safety boundary is composed of the tension upper limit value and the tension lower limit value. If the judgment result shows that the tension target value exceeds the boundary limit, the difference between the tension target value and the boundary value is calculated, and the initial target value is corrected in combination with the restriction strategy to generate a restricted adjustment instruction to ensure that the actual adjustment will not cause the risk of over-tension or relaxation. If the judgment result shows that the tension target value is within the boundary range, no correction is required, and the tension adjustment parameter is directly used as the final adjustment instruction to ensure that the instruction output is stably executed within the control tolerance.

[0014] In a preferred example, the present application may be further configured as follows: if the real-time tension adjustment result satisfies the target tension range, the current tension adjustment state is maintained, further comprising: If the real-time tension adjustment result does not meet the target tension range, obtaining historical tension adjustment records; updating the tension adjustment model using the historical tension adjustment records to obtain an updated tension adjustment model; Acquiring reference parameter information from the updated tension adjustment model, and fusing the real-time tension adjustment result with the reference parameter information to form state fusion input information; Based on the state fusion input information, image sequence information is acquired, and the state fusion input information and the image sequence information are input into a path trend prediction model to obtain wire movement trend information; According to the matching of the wire movement trend information with the preset offset identification rule, path pre-adjustment instruction information is generated, and the limited adjustment instruction information is fused with the path pre-adjustment instruction information to form joint control output information.

[0015] By adopting the above technical solution, when the real-time tension adjustment result does not meet the target tension range, the historical tension adjustment record is used to update the tension adjustment model, and the reference parameter information is integrated to generate state fusion input information, thereby improving the adaptability of the adjustment strategy. Further image sequence information is obtained, and the path trend prediction model is jointly driven with the state fusion input information to identify the wire movement trend. The path pre-adjustment instruction information is generated in combination with the preset offset recognition rules, and the joint control output information is formed with the tension adjustment to achieve synchronous adjustment of the tension and the motion path, thereby enhancing the robustness and accuracy of the winding control.

[0016] In a preferred example, the present application may be further configured as follows: the state fusion input information and the image sequence information are inputted into a path trend prediction model to obtain wire movement trend information, including: Constructing a wire force state vector based on the tension change characteristics and reference parameter characteristics in the state fusion input information; Performing target tracking and edge contour extraction between consecutive frames on the image sequence information, extracting the spatial position information of the target wire at each moment, and forming an actual motion trajectory sequence; Performing feature alignment processing on the wire force state vector and the actual motion trajectory sequence to obtain path-related feature information; The path correlation feature information is input into the path trend prediction model, and the correlation trend between the current wire state and the historical offset samples is analyzed by fitting to output the wire movement trend information.

[0017] By adopting the above technical solution, the wire force state vector constructed by fusing the tension change characteristics and the reference parameter characteristics can be aligned with the spatial motion trajectory extracted from the image sequence. On this basis, the path trend prediction model is used to fit and analyze the trend between the current state of the wire and the historical offset samples, and the wire motion trend information is output. This realizes the comprehensive perception of the force state and spatial trajectory in the complex winding process, improves the accuracy of the path offset prediction, and is conducive to further ensuring and optimizing the winding accuracy.

[0018] The second object of the present invention is achieved through the following technical solutions: An electromagnetic coil winding tension control device, the electromagnetic coil winding tension control device comprising: A status acquisition module is used to obtain real-time status information of the target wire during the winding process; a type identification module, configured to input the real-time status information into a wire type identification model to obtain wire type information, and call a corresponding tension adjustment model from a tension control model library based on the wire type information; a parameter generation module, configured to input the real-time state information into the tension adjustment model to obtain tension adjustment parameter information, and compare the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information; a tension execution module, configured to execute a tension adjustment operation according to the limited adjustment instruction information, obtain feedback tension information, and compare and analyze the feedback tension information with the limited adjustment instruction information to obtain a real-time tension adjustment result; The result determination module is configured to maintain the current tension adjustment state if the real-time tension adjustment result satisfies the target tension range.

[0019] By adopting the above technical solution, it is possible to obtain real-time status information of the target wire during the winding process, and accurately identify the material category, wire diameter specification and flexibility parameters in combination with the wire type recognition model, and call the corresponding tension adjustment model in the tension control model library based on the recognition result, so as to achieve dynamic adaptation and personalized matching of the tension adjustment strategy, further input the real-time status information into the tension adjustment model, output the tension adjustment parameter information for the current state, and introduce the preset tension safety boundary information for judgment and restriction, forming a restricted adjustment instruction to ensure the stability and safety of the adjustment process, and then execute the tension adjustment operation according to the restricted adjustment instruction and obtain feedback tension information, and form a real-time tension adjustment result by comparing and analyzing the feedback tension information with the adjustment instruction. When the adjustment result meets the target tension range, the current state is maintained without excessive intervention.

[0020] The third objective of this application is achieved through the following technical solutions: A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method when executing the computer program.

[0021] The fourth objective of this application is achieved through the following technical solutions: A computer-readable storage medium stores a computer program, wherein the computer program implements the steps of the above-mentioned method when executed by a processor.

[0022] In summary, this application includes at least one of the following beneficial technical effects: 1. Based on the real-time status information of the target wire during the winding process, it can accurately identify the material category, wire diameter specification and flexibility parameters in combination with the wire type recognition model, and call the corresponding tension adjustment model in the tension control model library based on the recognition result, so as to achieve dynamic adaptation and personalized matching of the tension adjustment strategy. The real-time status information is further input into the tension adjustment model, and the tension adjustment parameter information for the current state is output. The preset tension safety boundary information is introduced for judgment and restriction to form a restricted adjustment instruction to ensure the stability and safety of the adjustment process. Subsequently, the tension adjustment operation is executed according to the restricted adjustment instruction and feedback tension information is obtained. By comparing and analyzing the feedback tension information with the adjustment instruction, a real-time tension adjustment result is formed. When the adjustment result meets the target tension range, the current state is maintained without excessive intervention. 2. When the real-time tension adjustment result does not meet the target tension range, the tension adjustment model is updated using historical tension adjustment records, and reference parameter information is integrated to generate state fusion input information to improve the adaptability of the adjustment strategy. Image sequence information is further obtained and used together with the state fusion input information to drive the path trend prediction model to identify the wire movement trend. Path pre-adjustment instruction information is generated in combination with preset offset recognition rules, and combined with tension adjustment to form joint control output information, achieving synchronous adjustment of tension and movement path, and enhancing the robustness and accuracy of winding control. 3. It can align the wire force state vector constructed by fusing the tension change characteristics with the reference parameter characteristics with the spatial motion trajectory extracted from the image sequence. On this basis, the path trend prediction model is used to fit and analyze the trend between the current state of the wire and the historical offset samples, and the wire motion trend information is output to achieve a comprehensive perception of the force state and spatial trajectory in the complex winding process, improve the accuracy of path offset prediction, and facilitate further guarantee and optimization of winding accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a schematic structural diagram of a method for controlling the tension of an electromagnetic coil winding according to an embodiment of the application; Figure 2 This is a flowchart for implementing step S20 in a method for controlling the winding tension of an electromagnetic coil in one embodiment of the present application; Figure 3 This is another implementation flow chart of step S20 in a method for controlling the tension of an electromagnetic coil winding in one embodiment of the present application; Figure 4 This is a flowchart for implementing step S30 in a method for controlling the winding tension of an electromagnetic coil in one embodiment of the present application; Figure 5 This is another implementation flow chart of step S30 in a method for controlling the tension of an electromagnetic coil winding in one embodiment of the present application; Figure 6 This is a flowchart for implementing step S50 in a method for controlling the winding tension of an electromagnetic coil in one embodiment of the present application; Figure 7 This is a flowchart for implementing step S504 in a method for controlling the winding tension of an electromagnetic coil in one embodiment of the present application; Figure 8 This is a principle block diagram of an electromagnetic coil winding tension control device in one embodiment of the present application; Figure 9 This is a schematic diagram of a device in one embodiment of the present application. DETAILED DESCRIPTION The present application is further described in detail below with reference to the accompanying drawings.

[0024] In one embodiment, if Figure 1 As shown, the present application discloses a method for controlling the tension of an electromagnetic coil winding, which specifically includes the following steps: S10: Acquire real-time status information of the target wire during the winding process.

[0025] In this embodiment, real-time status information refers to a set of dynamic information collected during the electromagnetic coil winding process to characterize the current motion state of the target wire, including parameters such as tension change value, bending state, speed change and position information. The dynamic information is obtained through means such as tension sensors, displacement sensors and visual acquisition devices and recorded and updated at predetermined time intervals.

[0026] Specifically, by deploying multiple groups of non-contact high-precision tension sensors and laser displacement sensors on the critical path of the winding station, the tension change and wire bending shape parameters are obtained respectively. The tension sensor adopts a fiber optic strain gauge structure, and calculates the tension change value of the wire per unit time based on the interference fringe change of the reflected light signal. The laser displacement sensor continuously scans the lateral offset and bending radius change of the wire in the winding area, and combines the displacement difference within the set sampling period to infer the bending angle change generated during the winding process. By setting timing control rules, the above two types of sensor signals are synchronously triggered and time-aligned, and finally the continuous tension data and bending data are integrated into a state data stream with consistent timestamps to construct real-time status information representing the current force and spatial deformation conditions of the wire.

[0027] S20: Inputting the real-time status information into a wire type recognition model to obtain wire type information, and calling a corresponding tension adjustment model from a tension control model library based on the wire type information.

[0028] In this embodiment, the wire type recognition model refers to a classification model constructed based on multidimensional behavioral feature data. The classification model learns the mapping relationship between tension feature information and bending feature information through a training process and is used to identify the material category, wire diameter specification, and flexibility level of the target wire. The model is constructed using a support vector machine. The tension control model library refers to a collection of tension adjustment models stored in the form of a data structure. The model collection is indexed and classified according to different wire types. Each model corresponds to a specific material type, wire diameter parameter, and flexibility parameter, and contains the corresponding model structure definition and trained parameter configuration, which is used for tension adjustment calculations under specific working conditions during the winding process. The tension adjustment model refers to a model entity generated by a training algorithm and capable of executing tension control calculations. The model is based on the feature quantity input in the real-time status information and outputs tension adjustment parameter information, including control quantities such as target tension value, adjustment rate, and adjustment direction. The model is constructed using a time series regression model.

[0029] Specifically, the tension change values ​​and bending feature data contained in the acquired real-time status information are grouped and processed according to the time axis, and the continuously changing numerical segments in each group of data are input into the wire type recognition model as sliding window samples. The feature extraction mechanism based on the convolution kernel structure is used to extract the time series feature vector formed by the fluctuation frequency of the tension direction and the bending change trend. The input feature vector is matched with the clustering center parameters configured in the model to perform similarity matching to determine the closest wire type label value. The matched type label is used as the wire type information, and the material parameters, wire diameter parameters and flexibility parameters are extracted. The three extracted parameters are used as index conditions and compared one by one with the index configurations of multiple tension adjustment models registered in the tension control model library. During the comparison process, the weight distribution rule is used to perform weighted scoring on the matching degree of each parameter. The item with the highest matching score is selected as the corresponding item of the target tension adjustment model currently called, and the parameter configuration and structure description bound to the target tension adjustment model are retrieved from the tension control model library to complete the model calling process.

[0030] S30: Inputting the real-time state information into the tension adjustment model to obtain tension adjustment parameter information, and comparing the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information.

[0031] In this embodiment, the preset tension safety boundary information refers to the tension upper limit value and tension lower limit value range defined for the tension control requirements of the target wire type during the winding process. The upper and lower limits are determined based on empirical data, material mechanical properties and historical winding failure analysis results, and stored as callable judgment rules for constraining the output range of the tension adjustment parameters.

[0032] Specifically, the tension value and bending value in the real-time status information are segmented according to the set time window, and the tension change amplitude and bending rate in each time period are calculated respectively. The calculated change amount and rate value are standardized to form a multi-dimensional input data vector, and the input data vector is input into the tension adjustment model. According to the control function structure and parameter mapping rules preset in the tension adjustment model, a numerical calculation process is executed to obtain the tension target value of the current stage, and the tension target value is further judged against the tension upper limit value and the tension lower limit value defined in the tension safety boundary information. If the tension target value exceeds the tension upper limit value or is lower than the tension lower limit value, it is corrected in a proportional scaling manner based on the numerical deviation between the tension target value and the boundary value to generate corresponding restricted adjustment instruction information. If the tension target value is between the tension upper limit value and the tension lower limit value, the tension target value is directly used as the restricted adjustment instruction information.

[0033] S40: performing a tension adjustment operation according to the limited adjustment instruction information to obtain feedback tension information, and comparing and analyzing the feedback tension information with the limited adjustment instruction information to obtain a real-time tension adjustment result.

[0034] Specifically, according to the tension target value contained in the limited adjustment instruction information, a corresponding drive signal is output to the control end to adjust the tension control execution structure in the winding device, and the wire tension state is changed by driving the tension release mechanism or the traction mechanism. At the same time, during the tension adjustment execution process, the real-time feedback data collected by the tension sensor is continuously obtained, and a point-by-point comparison is performed with the tension target value. After time alignment of each set of difference values ​​obtained from the comparison, the average deviation value and the change trend value are calculated. Combined with the current winding segment length and the tension change response time, real-time difference aggregation analysis and dynamic leveling processing are performed, and finally a real-time tension adjustment result is formed to reflect whether the current tension adjustment response reaches the target tension range.

[0035] S50: If the real-time tension adjustment result meets the target tension range, the current tension adjustment state is maintained.

[0036] In this embodiment, the target tension range refers to the tension adjustment target interval defined in combination with the winding equipment performance, coil design parameters and wire material characteristics. The target interval is used to evaluate the effectiveness of the real-time tension adjustment results. The upper and lower limits can be set according to specific process requirements and compared with the actual feedback tension value.

[0037] Specifically, when the average deviation value in the real-time tension adjustment result is within the allowable deviation range defined by the target tension range, and the change trend value does not show signs of reversal exceeding the fluctuation threshold, it is judged based on the result that the current tension control state has met the stable control requirements. By interrupting further adjustment instructions, the output state of the current tension control execution structure is maintained unchanged, and the current tension target value and feedback response state are retained as the tension control state.

[0038] In one embodiment, if Figure 2 As shown, in step S20, that is, inputting the real-time status information into the wire type recognition model to obtain the wire type information, the process includes: S201: Perform sliding window processing on the tension change value in the real-time status information to obtain tension feature information.

[0039] Specifically, after obtaining real-time status information, the tension change value is selected as the processing object, and the equally spaced sliding window mechanism is used to divide the tension change value into multiple continuous fixed-length data segments according to the time series. Each data segment slides in sequence according to the set step size to cover all tension change samples. In each sliding window, the mean, range, standard deviation and maximum gradient of the tension change value are calculated respectively. The tension fluctuation characteristics are expressed by the above statistics, and then the feature quantities corresponding to all windows are integrated into tension feature information.

[0040] S202: Performing segment fitting processing on the bending state information in the real-time state information to obtain bending feature information.

[0041] Specifically, when processing the bending state information in the real-time state information, the continuous bending state data is first segmented according to the timestamp or sampling point index. Each segment contains a fixed number of continuous data points. In each data segment, the polynomial fitting method is applied to construct a second-order or third-order fitting curve. The fitting coefficient is calculated by the least squares method and the approximate curvature change trajectory of the wire in the segment is fitted. The curvature peak, inflection point position and fitting residual value of each fitting curve are further extracted, and these numerical features that characterize the geometric change law of the wire are summarized as bending feature information.

[0042] S203: Combining the tension characteristic information and the bending characteristic information to form a wire behavior characteristic data set, and inputting the wire behavior characteristic data set into the wire type recognition model.

[0043] Specifically, the statistics of each window in the tension feature information and the fitting features of each segment in the bending feature information are synchronously arranged in chronological order, the feature dimensions of the two information sources are aligned based on the feature correspondence, and the tension feature values ​​and bending feature values ​​under the same time slice are connected by vector splicing to form a multi-dimensional feature vector sequence. After completing the full-time period combination, a wire behavior feature data group is constructed, and the data group is normalized and format-standardized, and the formatted wire behavior feature data group is input into the trained wire type recognition model.

[0044] S204: The wire type recognition model performs multi-dimensional feature clustering and type mapping analysis on the wire behavior feature data group, and outputs the wire type information corresponding to the real-time status information.

[0045] Specifically, the cluster analysis structure in the wire type recognition model is used to classify the features of each dimension in the wire behavior feature data group. By setting the initial number of cluster centers and gradually iteratively updating the cluster center positions based on the Euclidean distance calculation method, all feature samples converge to stable category labels. After obtaining the cluster labels, each type of clustering result is matched with the known wire type according to the predefined wire type label mapping relationship in the model, thereby completing the process of multi-dimensional feature clustering and type mapping, and outputting the wire type information associated with the current real-time status information.

[0046] In one embodiment, if Figure 3 As shown, in step S20, the corresponding tension adjustment model is called from the tension control model library based on the wire type information, including: S205: Using the material type parameter, wire diameter parameter, and flexibility parameter in the wire type information as matching conditions, sequentially matching them with index information of candidate tension adjustment models in the tension control model library.

[0047] Specifically, the material type parameters, wire diameter parameters and flexibility parameters indicated in the wire type information are extracted as three-dimensional matching vectors, and are compared and matched with the index information attached to each candidate tension adjustment model in the tension control model library in accordance with the set priority order. Among them, the material type parameters are compared by string full matching, the wire diameter parameters are compared by numerical interval matching controlled by error threshold, and the flexibility parameters are matched by level range mapping. During the matching process, a set of matching result vectors are generated for each candidate model, and the corresponding matching status labels are recorded.

[0048] S206: Performing correlation scoring processing on the candidate tension adjustment models that meet the matching condition to obtain a scoring result, and determining a target tension adjustment model according to the scoring result.

[0049] Specifically, the index parameter configurations of all candidate tension adjustment models that meet the matching conditions are extracted, the basic weights are set based on the consistency of the material type parameters, and a multi-factor weighted scoring function is constructed by combining the numerical difference degree of the wire diameter parameters and the grade similarity of the flexibility parameters. By substituting the matching condition parameters and the index parameters of the candidate models into the scoring function respectively, the correlation score value of each candidate model is calculated, and the model with the highest correlation score value among all the scoring results is selected as the target tension adjustment model.

[0050] S207: According to the identification information of the target tension adjustment model, the model structure and parameter configuration corresponding to the target tension adjustment model are retrieved and extracted from the tension control model library to obtain the tension adjustment model.

[0051] Specifically, the identification information of the target tension adjustment model is used as the unique index key to perform a retrieval operation in the tension control model library. First, the corresponding model storage entry is located, and then the model structure information and parameter configuration file recorded in the entry are parsed. The model structure information includes the model type, input dimension, output variable and internal calculation logic structure. The parameter configuration includes the weight parameters, activation function type and adjustment factor range obtained from historical training. The above contents are read and loaded in sequence to complete the complete extraction and construction of the target tension adjustment model.

[0052] In one embodiment, if Figure 4 As shown, in step S30, that is, inputting the real-time state information into the tension adjustment model to obtain tension adjustment parameter information, the process includes: S301: performing data segmentation processing and value normalization processing on the real-time status information, and extracting the value change amount and change rate of the real-time status information in multiple time series intervals.

[0053] Specifically, the real-time status information is divided into equal intervals according to a preset time step to form multiple continuous time series data segments. Each time series data segment contains a continuous record of tension information and bending state information. For each time series data segment, the difference between the tension value and the bending state value between the starting sampling point and the ending sampling point is calculated to obtain the numerical change of the time series data segment. The change rate is calculated based on the tension change between two adjacent sampling points divided by the time interval. The maximum and minimum value normalization processing is performed on the numerical change and the change rate, so that each dimension of data is mapped to the standard interval, forming a unified data input format for subsequent model calculations.

[0054] S302: Inputting the numerical change amount and the change rate as input data into the tension adjustment model to obtain the tension adjustment parameter information.

[0055] Specifically, the extracted numerical changes and change rates are arranged according to the two dimensions of tension and bending state, and an input matrix containing multiple groups of time series features is constructed. The matrix is ​​mapped to the input layer of the tension regulation model according to a preset data dimension format. The feature transformation structure and parameter mapping structure of the tension regulation model are sequentially passed through, and a combination of linear weighted transformation and nonlinear activation function is performed in the feature transformation structure to convert low-dimensional features into high-order feature representations. In the parameter mapping structure, a regression calculation of the tension regulation relationship is performed based on the multi-layer perception unit, and the target tension value, tension response coefficient and adjustment slope coefficient corresponding to the current input state are output. The target tension value, tension response coefficient and adjustment slope coefficient are combined to form tension regulation parameter information.

[0056] In one embodiment, if Figure 5 As shown, in step S30, the tension adjustment parameter information is compared with the preset tension safety boundary information to generate limited adjustment instruction information, including: S303: Performing a range judgment on the tension target value in the tension adjustment parameter information, and comparing whether the tension target value is between the tension upper limit value and the tension lower limit value defined by the tension safety boundary information.

[0057] Specifically, the tension target value in the tension adjustment parameter information and the tension upper limit value and tension lower limit value in the tension safety boundary information are numerically compared respectively. First, the numerical expression of the tension target value is extracted, and the upper and lower limits of the tension allowable range under the current corresponding winding task are read from the preset tension safety boundary information. By executing the interval inclusion judgment logic, it is calculated whether the tension target value is greater than the tension lower limit value and less than the tension upper limit value. If the above two conditions are met at the same time, it is judged that the tension target value is within the tension safety boundary range.

[0058] S304: When the tension target value exceeds the tension upper limit or is lower than the tension lower limit, based on the difference between the tension target value and the tension safety margin information, the tension adjustment parameter information is modified to generate the limited adjustment instruction.

[0059] Specifically, when it is determined through numerical comparison that the tension target value exceeds the tension upper limit or is lower than the tension lower limit, the difference between the tension target value and the tension safety boundary is calculated. If the tension target value is greater than the tension upper limit, the difference between the tension target value and the tension upper limit is used as the upper limit deviation. If the tension target value is less than the tension lower limit, the difference between the tension lower limit and the tension target value is used as the lower limit deviation. According to the deviation, the tension target value in the original tension adjustment parameter information is corrected so that the corrected tension target value does not exceed the tension safety boundary range. During the correction process, the tension target value is adjusted by linear interpolation, while keeping other parameter items in the tension adjustment parameter information unchanged. Finally, the parameter set containing the corrected tension target value is constructed as a restricted adjustment instruction.

[0060] S305: When the tension target value is between the tension upper limit value and the tension lower limit value defined by the tension safety boundary information, use the tension adjustment parameter information as the limited adjustment instruction information.

[0061] Specifically, when it is determined through comparison that the tension target value is between the tension upper limit value and the tension lower limit value, no numerical correction is made to the tension target value in the tension adjustment parameter information, the entire content of the originally generated tension adjustment parameter information is retained, and the unmodified tension adjustment parameter information is directly used as the restricted adjustment instruction information, which includes various control parameters such as the tension target value, adjustment rate, and response delay coefficient.

[0062] In one embodiment, if Figure 6 As shown, in step S50, that is, if the real-time tension adjustment result meets the target tension range, the current tension adjustment state is maintained, and the following further comprises: S501: If the real-time tension adjustment result does not meet the target tension range, obtain historical tension adjustment records.

[0063] Specifically, if the real-time tension adjustment result does not fall within the preset target tension range, the corresponding historical tension adjustment record in the storage unit is retrieved based on the identification information of the current winding task, and the adjustment process data that matches the current wire type and is representative is screened out. The historical tension adjustment record includes the controlled tension value recorded in the previous operation process, the execution time point, the error change between the adjustment response instruction and the actual feedback result, etc., and a time series mapping relationship between the current tension state and the historical record is established through index association for subsequent model updating and control strategy correction.

[0064] S502: updating the tension adjustment model according to the historical tension adjustment records to obtain an updated tension adjustment model.

[0065] Specifically, the tension target value, feedback tension value and their corresponding time series contained in the acquired historical tension adjustment records are used to construct an error backtracking vector to extract the response deviation of the current tension adjustment model at each stage. The weights of the control factors with greater influence in the model are adjusted according to the error distribution. The model parameters are iteratively updated using the gradient descent optimization method until the error change meets the set convergence conditions. The model parameters after the iteration are integrated into the original model structure to form an updated tension adjustment model.

[0066] S503: Acquire reference parameter information from the updated tension adjustment model, and fuse the real-time tension adjustment result with the reference parameter information to form state fusion input information.

[0067] Specifically, key parameters characterizing the current control responsiveness are extracted from the updated tension regulation model, including the tension adjustment coefficient, feedback error tolerance, and time response characteristic index. These parameters are used as reference parameters and combined with the current feedback tension value, adjustment delay time, and error deviation data contained in the real-time tension regulation results. Numerical fusion is performed using a weighted linear combination method. The fusion results are normalized and mapped according to the parameter dimensions to form state fusion input information that comprehensively expresses the current tension state and the model's responsiveness. S504: Based on the state fusion input information, image sequence information is acquired, and the state fusion input information and the image sequence information are input into a path trend prediction model to obtain wire movement trend information.

[0068] In this embodiment, image sequence information refers to a sequence of image frames of the target wire during the winding process, continuously captured by a high-speed industrial camera or visual acquisition device. The image frames are arranged in chronological order and are used to record the wire's spatial motion trajectory and deformation, which will be subsequently used for target tracking, edge extraction, and path trend prediction. The path trend prediction model is a mathematical model used to predict the trend of changes in the target wire's motion path over a future time period based on input features. The model takes state fusion input information and image sequence information as input, combines historical trajectory offset samples for fitting calculations, and outputs the motion trend prediction result at the current moment. The model is constructed using a long-short-term memory network.

[0069] Specifically, a set image acquisition device is used to collect multiple frames of continuous images of the target wire during the winding process, and each frame of the image covers the actual position and deformation state of the wire in space. The frames of the image are sorted in time sequence through a timestamp synchronization mechanism to form image sequence information; an image preprocessing algorithm is used to perform noise suppression and contour enhancement on the image sequence, and image feature data containing the edge contour and posture information of the wire is extracted; the tension feedback information and the reference parameters in the state fusion input information are jointly mapped into a force state vector, which is aligned with the image feature data in the time dimension, and the aligned time series data is input as a joint input into the trained path trend prediction model. Through the spatiotemporal feature extraction and trend fitting structure in the model, the motion trend information of the target wire is output, which is used to characterize the potential trajectory evolution direction and offset trend of the wire in the subsequent time period under the current winding state.

[0070] S505: Generate path pre-adjustment instruction information according to matching of the wire movement trend information with a preset offset identification rule, and fuse the restricted adjustment instruction information with the path pre-adjustment instruction information to form joint control output information.

[0071] Specifically, based on the trajectory offset vector, position change rate and directionality index in the wire motion trend information, the motion parameters related to the spatial stability of the wire are extracted and compared with the pre-set offset identification rules, which include the maximum offset threshold, minimum rebound radius and corresponding spatial distribution pattern allowed under different wire diameters and tension conditions; when any motion parameter exceeds the corresponding threshold, the required path compensation angle and feed speed adjustment amount are calculated based on the trend change direction and the offset degree to form the path pre-adjustment instruction information; then the path pre-adjustment instruction information and the restricted adjustment instruction information are matched at the information level according to the preset fusion mapping strategy, and a temporal linkage relationship between the tension adjustment command and the path control compensation is established. The instruction fusion is completed and the joint control output information is generated to simultaneously guide the tension adjustment and path correction operations to improve the stability and accuracy of the winding process.

[0072] In one embodiment, if Figure 7 As shown, in step S504, the state fusion input information and the image sequence information are input into the path trend prediction model to obtain the wire movement trend information, including: S5041: Constructing a wire force state vector based on the tension change characteristics and reference parameter characteristics in the state fusion input information.

[0073] Specifically, the numerical feature dimensions of the tension change features and reference parameter features in the state fusion input information are extracted respectively, where the tension change features include the tension measurement values, fluctuation frequency and maximum change gradient of multiple continuous time points, and the reference parameter features include the historical response coefficient, adjustment delay time and model error weight coefficient in the tension regulation model. After the numerical scale is unified by normalization, the tension data and reference parameter values ​​corresponding to each moment are spliced ​​in a time-series synchronization manner to form a multi-dimensional feature matrix with consistent structure. Subsequently, the vector mapping algorithm is used to map the feature matrix into a fixed-length linear vector to obtain the force state vector of the wire during the winding process.

[0074] S5042: performing target tracking and edge contour extraction between consecutive frames on the image sequence information, extracting spatial position information of the target wire at each moment, and forming an actual motion trajectory sequence.

[0075] Specifically, a time-sequential input method of continuous image frames is adopted, and target tracking processing based on the optical flow method and feature point matching is performed for each frame image contained in the image sequence information. First, the position and edge information of the wire in the initial frame are identified through image grayscale distribution analysis and background modeling methods. Then, the KLT tracking algorithm or correlation filtering method is used in subsequent frames to update the position of the initial target area to obtain the displacement path of the target wire between consecutive frames. At the same time, an edge detection algorithm, such as the Canny or Sobel operator, is executed in each frame image to extract the edge contour of the wire. Combined with the target center point trajectory and edge curvature characteristics, the center coordinates, boundary shape and orientation information of the target wire in each frame image are recorded, and the spatial position information in all frames is arranged in chronological order to form a complete actual motion trajectory sequence.

[0076] S5043: Perform feature alignment processing on the wire force state vector and the actual motion trajectory sequence to obtain path-related feature information.

[0077] Specifically, for the constructed wire stress state vector and actual motion trajectory sequence, the time axes of the two are first synchronized to ensure that each set of stress state data corresponds one-to-one to the trajectory point at the corresponding moment. On this basis, the interpolation method is used to uniformly adjust the sampling frequency of data with inconsistent time resolution. Then, a joint feature mapping matrix is ​​constructed based on the position coordinate information and the tension state index. The response relationship between the tension magnitude and spatial displacement at each time point is calculated. After eliminating the dimensional difference through normalization, the corresponding change trend, turning point and extreme value information are extracted. The dynamic time warping (DTW) algorithm is further used to align the two sequences in the time domain and feature space to obtain the associated matching pairs between the tension state and the motion path corresponding to each moment, which are finally combined to form the path-related feature information that characterizes the motion behavior of the wire under the influence of tension.

[0078] S5044: Input the path association feature information into the path trend prediction model, and output the wire movement trend information by fitting and analyzing the association trend between the current wire state and historical offset samples.

[0079] Specifically, the input feature vector group is first extracted according to the tension response features and spatial trajectory features contained in the path-associated feature information, and the training vectors of the corresponding dimensions are extracted based on the training sample set constructed in the historical offset samples. The long short-term memory network (LSTM) is adopted as the path trend prediction model. Its advantages in time series modeling are utilized to perform fitting analysis on the nonlinear temporal relationship between the current input feature vector group and the historical samples. During the fitting process, error backpropagation and gradient update are performed to optimize the model parameters, thereby outputting the wire motion trend information representing the movement direction, speed change and spatial offset trend of the target wire in a predetermined time domain.

[0080] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0081] In one embodiment, an electromagnetic coil winding tension control device is provided, and the electromagnetic coil winding tension control device corresponds to the electromagnetic coil winding tension control method in the above embodiment. Figure 8 As shown, the electromagnetic coil winding tension control device includes a state acquisition module, a type identification module, a parameter generation module, a tension execution module and a result determination module. The functional modules are described in detail as follows: A status acquisition module is used to obtain real-time status information of the target wire during the winding process; a type identification module, configured to input the real-time status information into a wire type identification model to obtain wire type information, and call a corresponding tension adjustment model from a tension control model library based on the wire type information; a parameter generation module, configured to input the real-time state information into the tension adjustment model to obtain tension adjustment parameter information, and compare the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information; a tension execution module, configured to execute a tension adjustment operation according to the limited adjustment instruction information, obtain feedback tension information, and compare and analyze the feedback tension information with the limited adjustment instruction information to obtain a real-time tension adjustment result; The result determination module is configured to maintain the current tension adjustment state if the real-time tension adjustment result satisfies the target tension range.

[0082] Optional, type identification module, including: a tension feature extraction submodule, configured to perform sliding window processing on the tension change value in the real-time state information to obtain tension feature information; a bending feature fitting submodule, configured to perform segment fitting processing on the bending state information in the real-time state information to obtain bending feature information; a feature fusion submodule, configured to combine the tension feature information and the bending feature information to form a wire behavior feature data set, and input the wire behavior feature data set into the wire type recognition model; The type identification submodule is configured to perform multi-dimensional feature clustering and type mapping analysis on the wire behavior feature data group using the wire type identification model, and output the wire type information corresponding to the real-time status information.

[0083] Optionally, the type identification module also includes: a model matching submodule, configured to use the material type parameter, wire diameter parameter, and flexibility parameter in the wire type information as matching conditions, and sequentially match the index information of the candidate tension adjustment models in the tension control model library; a scoring and selection submodule, configured to perform correlation scoring processing on the candidate tension adjustment models that meet the matching conditions, obtain a scoring result, and determine a target tension adjustment model according to the scoring result; The model retrieval submodule is used to retrieve and extract the model structure and parameter configuration corresponding to the target tension adjustment model from the tension control model library according to the identification information of the target tension adjustment model, so as to obtain the tension adjustment model.

[0084] Optional, parameter generation module, including; A feature extraction submodule is used to perform data segmentation processing and value normalization processing on the real-time status information, and extract the value change and change rate of the real-time status information in multiple time series intervals; The parameter generation submodule is used to input the numerical value change and the change rate as input data into the tension adjustment model to obtain the tension adjustment parameter information.

[0085] Optional, parameter generation module, including; a boundary judgment submodule, configured to perform range judgment on the tension target value in the tension adjustment parameter information, and compare whether the tension target value is between the tension upper limit value and the tension lower limit value defined by the tension safety boundary information; a parameter correction submodule, configured to correct the tension adjustment parameter information based on a difference between the tension target value and the tension safety margin information and generate the limited adjustment instruction when the tension target value exceeds the tension upper limit value or is lower than the tension lower limit value; The instruction confirmation module is configured to use the tension adjustment parameter information as the restricted adjustment instruction information when the tension target value is between the tension upper limit value and the tension lower limit value defined by the tension safety boundary information.

[0086] Optionally, the result determination module includes: a record acquisition submodule, configured to acquire historical tension adjustment records if the real-time tension adjustment result does not meet the target tension range; A model updating submodule, configured to update the tension adjustment model using the historical tension adjustment records to obtain an updated tension adjustment model; a parameter fusion submodule, configured to obtain reference parameter information from the updated tension adjustment model, and fuse the real-time tension adjustment result with the reference parameter information to form state fusion input information; a trend prediction submodule, configured to obtain image sequence information based on the state fusion input information, and input the state fusion input information and the image sequence information into a path trend prediction model to obtain wire movement trend information; The path control submodule is used to match the wire movement trend information with the preset offset recognition rules to generate path pre-adjustment instruction information, and fuse the limited adjustment instruction information with the path pre-adjustment instruction information to form joint control output information.

[0087] Optional, path control module, including: A state vector construction module is used to construct a wire force state vector based on the tension change characteristics and reference parameter characteristics in the state fusion input information; a trajectory extraction module, configured to perform target tracking and edge contour extraction between consecutive frames of the image sequence information, extract the spatial position information of the target wire at each moment, and form an actual motion trajectory sequence; A feature alignment module is used to perform feature alignment processing on the wire force state vector and the actual motion trajectory sequence to obtain path association feature information; The trend prediction module is used to input the path correlation feature information into the path trend prediction model, and output the wire movement trend information by fitting and analyzing the correlation trend between the current wire state and the historical offset samples.

[0088] The specific definition of an electromagnetic coil winding tension control device can be found in the definition of an electromagnetic coil winding tension control method described above and will not be further elaborated here. Each module in the aforementioned electromagnetic coil winding tension control device can be implemented in whole or in part via software, hardware, or a combination thereof. Each of these modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device memory in software form, allowing the processor to call and execute the corresponding operations of each module.

[0089] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 9 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, it implements a method for controlling the tension of an electromagnetic coil winding.

[0090] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed: Obtain real-time status information of the target wire during the winding process; Input the real-time status information into the wire type recognition model to obtain the wire type information, and call the corresponding tension adjustment model from the tension control model library based on the wire type information; Inputting real-time state information into the tension adjustment model to obtain tension adjustment parameter information, and comparing the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information; Perform tension adjustment operations according to the limited adjustment instruction information to obtain feedback tension information, and compare and analyze the feedback tension information with the limited adjustment instruction information to obtain real-time tension adjustment results; If the real-time tension adjustment result meets the target tension range, the current tension adjustment state is maintained.

[0091] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: Obtain real-time status information of the target wire during the winding process; Input the real-time status information into the wire type recognition model to obtain the wire type information, and call the corresponding tension adjustment model from the tension control model library based on the wire type information; Inputting real-time state information into the tension adjustment model to obtain tension adjustment parameter information, and comparing the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information; Perform tension adjustment operations according to the limited adjustment instruction information to obtain feedback tension information, and compare and analyze the feedback tension information with the limited adjustment instruction information to obtain real-time tension adjustment results; If the real-time tension adjustment result meets the target tension range, the current tension adjustment state is maintained.

[0092] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0093] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0094] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for controlling the tension of an electromagnetic coil winding, characterized in that: The electromagnetic coil winding tension control method comprises: Obtain real-time status information of the target wire during the winding process; Inputting the real-time status information into a wire type recognition model to obtain wire type information, and calling a corresponding tension adjustment model from a tension control model library based on the wire type information; Inputting the real-time state information into the tension adjustment model to obtain tension adjustment parameter information, and comparing the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information; performing a tension adjustment operation according to the limited adjustment instruction information to obtain feedback tension information, and comparing and analyzing the feedback tension information with the limited adjustment instruction information to obtain a real-time tension adjustment result; If the real-time tension adjustment result meets the target tension range, the current tension adjustment state is maintained.

2. The electromagnetic coil winding tension control method according to claim 1, characterized in that: Inputting the real-time status information into a wire type recognition model to obtain wire type information includes: Performing sliding window processing on the tension change value in the real-time status information to obtain tension feature information; Performing section fitting processing on the bending state information in the real-time state information to obtain bending feature information; Combining the tension characteristic information with the bending characteristic information to form a wire behavior characteristic data set, and inputting the wire behavior characteristic data set into the wire type recognition model; The wire type recognition model performs multi-dimensional feature clustering and type mapping analysis on the wire behavior feature data group, and outputs the wire type information corresponding to the real-time status information.

3. The electromagnetic coil winding tension control method according to claim 1, characterized in that: The calling a corresponding tension adjustment model from a tension control model library based on the wire type information includes: Using the material type parameter, wire diameter parameter, and flexibility parameter in the wire type information as matching conditions, the matching is performed in sequence with the index information of the candidate tension adjustment models in the tension control model library; performing correlation scoring processing on the candidate tension adjustment models that meet the matching condition to obtain a scoring result, and determining a target tension adjustment model according to the scoring result; According to the identification information of the target tension adjustment model, the model structure and parameter configuration corresponding to the target tension adjustment model are retrieved and extracted from the tension control model library to obtain the tension adjustment model.

4. The electromagnetic coil winding tension control method according to claim 1, characterized in that: The step of inputting the real-time status information into the tension adjustment model to obtain tension adjustment parameter information includes: Performing data segmentation processing and value normalization processing on the real-time status information, and extracting the value change and change rate of the real-time status information in multiple time series intervals; The numerical change amount and the change rate are used as input data and input into the tension adjustment model to obtain the tension adjustment parameter information.

5. The electromagnetic coil winding tension control method according to claim 1, characterized in that: The comparing the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information includes: Performing a range judgment on the tension target value in the tension adjustment parameter information, and comparing whether the tension target value is between the tension upper limit value and the tension lower limit value defined by the tension safety boundary information; When the tension target value exceeds the tension upper limit or is lower than the tension lower limit, based on the difference between the tension target value and the tension safety margin information, the tension adjustment parameter information is modified to generate the limited adjustment instruction; When the tension target value is between the tension upper limit value and the tension lower limit value defined by the tension safety margin information, the tension adjustment parameter information is used as the limited adjustment instruction information.

6. The electromagnetic coil winding tension control method according to claim 1, characterized in that: If the real-time tension adjustment result meets the target tension range, maintaining the current tension adjustment state further includes: If the real-time tension adjustment result does not meet the target tension range, obtaining historical tension adjustment records; updating the tension adjustment model using the historical tension adjustment records to obtain an updated tension adjustment model; Acquiring reference parameter information from the updated tension adjustment model, and fusing the real-time tension adjustment result with the reference parameter information to form state fusion input information; Based on the state fusion input information, image sequence information is acquired, and the state fusion input information and the image sequence information are input into a path trend prediction model to obtain wire movement trend information; According to the matching of the wire movement trend information with the preset offset identification rule, path pre-adjustment instruction information is generated, and the limited adjustment instruction information is fused with the path pre-adjustment instruction information to form joint control output information.

7. The electromagnetic coil winding tension control method according to claim 6, characterized in that: The step of inputting the state fusion input information and the image sequence information into a path trend prediction model to obtain wire movement trend information includes: Constructing a wire force state vector based on the tension change characteristics and reference parameter characteristics in the state fusion input information; Performing target tracking and edge contour extraction between consecutive frames on the image sequence information, extracting the spatial position information of the target wire at each moment, and forming an actual motion trajectory sequence; Performing feature alignment processing on the wire force state vector and the actual motion trajectory sequence to obtain path-related feature information; The path correlation feature information is input into the path trend prediction model, and the correlation trend between the current wire state and the historical offset samples is analyzed by fitting to output the wire movement trend information.

8. An electromagnetic coil winding tension control device, characterized in that: The electromagnetic coil winding tension control device comprises: A status acquisition module is used to obtain real-time status information of the target wire during the winding process; a type identification module, configured to input the real-time status information into a wire type identification model to obtain wire type information, and call a corresponding tension adjustment model from a tension control model library based on the wire type information; a parameter generation module, configured to input the real-time state information into the tension adjustment model to obtain tension adjustment parameter information, and compare the tension adjustment parameter information with preset tension safety boundary information to generate limited adjustment instruction information; a tension execution module, configured to execute a tension adjustment operation according to the limited adjustment instruction information, obtain feedback tension information, and compare and analyze the feedback tension information with the limited adjustment instruction information to obtain a real-time tension adjustment result; The result determination module is configured to maintain the current tension adjustment state if the real-time tension adjustment result satisfies the target tension range.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the electromagnetic coil winding tension control method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the electromagnetic coil winding tension control method according to any one of claims 1 to 7 are implemented.

Images