Automatic adjusting method and system for enameled wire production process

Abstract

The invention relates to the technical field of varnished wire production, in particular to an automatic adjustment method and system for a varnished wire production process, and the method comprises the steps: obtaining the real-time linear speed data of a wire rod in a varnishing process of varnished wire production; determining wire motion characteristics according to the real-time linear velocity data; inputting the wire motion characteristics and the wire speed characteristics into a pre-constructed risk prediction model for prediction processing to obtain a wire motion risk coefficient; steepness of all the sequences is determined according to the real-time linear speed data, and an automatic adjustment strategy is generated according to the steepness output of all the sequences to obtain the wire operation complexity; and inputting the maximum steepness and the wire motion characteristics in the extracted sequence into a multi-dimensional learning model to obtain the wire pay-off speed regulation and control proportion. Under the condition that the safety and adjustability are qualified, the linear speed is automatically regulated by considering the change fluctuation, early warning, trend and complex comprehensive degree of the linear speed, so that the wire pay-off speed regulation and control proportion is greatly improved, and the wire pay-off speed regulation and control proportion is improved. And the uniform coating effect of the enameled wire is ensured.

Description

Technical Field

[0001] The invention relates to the technical field of enameled wire production, and in particular to an automated adjustment method and system for an enameled wire production process. Background Art

[0002] In the existing technology, especially in the painting process, when the wire is at a low speed, the paint liquid maintains a laminar state in the coating mold and flows steadily; when the wire speed exceeds a certain value, the paint liquid flow rate accelerates and the Reynolds number increases, which may cause turbulence, causing "fish scale patterns" or local accumulation on the surface of the paint film. When the high-speed wire passes through the paint liquid, air is easily drawn into the paint film to form bubbles or pinholes; if the paint layer is uneven, it may cause the insulation performance to fail to meet the standards, or breakdown problems may occur in subsequent use. Summary of the invention

[0003] The purpose of the present invention is to provide an automated adjustment method and system for the enameled wire production process. The technical problem solved by the present invention is that the line speed fluctuates during wire coating, resulting in uneven paint film on the wire surface, which affects the quality of the enameled wire.

[0004] The purpose of the present invention can be achieved through the following technical solutions: The method for automatically adjusting the production process of enameled wire comprises the following steps: S1. In the painting process of enameled wire production, obtaining the real-time linear velocity data of the wire; determining the movement characteristics of the wire according to the real-time linear velocity data; S2. Inputting the wire movement characteristics and wire speed characteristics into a pre-built risk prediction model for prediction processing to obtain a wire movement risk coefficient; and determining the wire speed risk level according to the wire movement risk coefficient; S3. Based on the determined line speed danger level, the steepness of all sequences is determined according to the real-time line speed data, and the wire running complexity obtained by the steepness output of all sequences is used to generate an automated adjustment strategy; S4. When receiving the adjustment instruction in the automatic adjustment strategy, the maximum steepness and wire movement characteristics in the sequence are extracted and input into the multi-dimensional learning model to obtain the wire pay-off speed control ratio.

[0005] As a further solution of the present invention: a method for determining the motion characteristics of a wire according to real-time wire speed data comprises: An analysis period is set, and the real-time wire speed data of the analysis period is analyzed in turn in terms of fluctuation degree, warning degree, and trend degree to obtain wire movement fluctuation characteristics, wire movement warning characteristics, and wire movement trend characteristics.

[0006] As a further solution of the present invention, the real-time line speed data of the analysis period is analyzed in turn for fluctuation degree, and the process of obtaining the fluctuation characteristics of the wire movement is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least squares method, extract all the peak points in the linear velocity curve, construct a time series sequence of the linear velocity peaks, perform fluctuation calculation on the time series sequence of the linear velocity peaks, and obtain the motion fluctuation characteristics of the wire.

[0007] As a further solution of the present invention: The process of performing early warning degree analysis on the real-time linear velocity data of the analysis period in sequence to obtain the motion early warning characteristics of the wire is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least squares method, extract all the peak points in the linear velocity curve, construct a time series sequence of the linear velocity peaks, extract the maximum value in the time series sequence of the linear velocity peaks, calculate the proximity degree between the maximum value and the preset risk value, and obtain the motion early warning characteristics of the wire.

[0008] As a further solution of the present invention: The process of performing trend degree analysis on the real-time linear velocity data of the analysis period in sequence to obtain the motion trend characteristics of the wire is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least squares method, extract all the peak points in the linear velocity curve, construct a time series sequence of the linear velocity peaks, and obtain the motion trend characteristics of the wire according to the distribution difference between the center linear velocity data of the initial sequence segment and other linear velocity data.

[0009] As a further solution of the present invention: The motion risk coefficient of the wire includes 2 levels, corresponding to 1 and 2 respectively. 1 indicates that the motion speed of the wire during the coating process is in a safe state, and 2 indicates that the motion speed of the wire during the coating process is in a dangerous state.

[0010] As a further solution of the present invention: The method for determining the motion complexity of the wire according to the real-time linear velocity data is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least squares method, extract all the peak segments, construct a time series sequence of the linear velocity peak segments, calculate the total steepness value of all the sequences, and obtain the motion complexity of the wire; Calculate the degree of the steepness of all the sequences combined with the steepness coefficient and frequency coefficient of each sequence.

[0011] As a further solution of the present invention: The method for generating an automatic adjustment strategy through the motion complexity of the wire output by the steepness of all the sequences is as follows: If the motion complexity of the wire exceeds the motion complexity threshold of the wire, a regulation instruction is generated; If the motion complexity of the wire does not exceed the motion complexity threshold of the wire, a stop instruction is generated.

[0012] As a further solution of the present invention: the input of the input layer is the maximum steepness, wire movement fluctuation characteristics, wire movement warning characteristics, and wire movement trend characteristics in the extraction sequence. The hidden layer contains 7 nodes, the activation function of the hidden layer is the ReLU activation function, the output layer contains 2 nodes, and the output of the output layer is the wire pay-off speed regulation ratio.

[0013] An automatic adjustment system for the enameled wire production process, the system includes: Analysis module: In the painting process of enameled wire production, obtain the real-time wire speed data of the wire; determine the wire movement characteristics according to the real-time wire speed data; Evaluation module: Input the wire movement characteristics and wire speed characteristics into a pre-constructed risk prediction model for prediction processing to obtain the wire movement risk coefficient; and determine the wire speed danger level according to the wire movement risk coefficient; Strategy module: Based on the determined wire speed danger level, determine the steepness of all sequences according to the real-time wire speed data, output the wire running complexity obtained by the steepness of all sequences, and generate an automatic adjustment strategy; Adjustment module: When receiving the adjustment instruction in the automatic adjustment strategy, input the maximum steepness and wire movement characteristics in the extraction sequence into a multi-dimensional learning model to obtain the wire pay-off speed regulation ratio.

[0014] Advantages of the present invention: (1) In the painting process of enameled wire production, the present invention obtains the real-time wire speed data of the wire; determines the wire movement characteristics according to the real-time wire speed data; inputs the wire movement characteristics and wire speed characteristics into a pre-constructed risk prediction model for prediction processing to obtain the wire movement risk coefficient; the present invention takes the wire speed of the enameled wire as the analysis object, comprehensively analyzes the wire speed from fluctuations, warnings, and trends, and comprehensively and accurately evaluates the risk that the current wire speed exceeds a certain value, so that when the wire movement speed is in a dangerous state, the tension of the coating die can be controlled in time; (2) Based on the determined wire speed danger level, the present invention determines the steepness of all sequences according to the real-time wire speed data, outputs the wire running complexity obtained by the steepness of all sequences, and generates an automatic adjustment strategy; when receiving the adjustment instruction in the automatic adjustment strategy, inputs the maximum steepness and wire movement characteristics in the extraction sequence into a multi-dimensional learning model to obtain the wire pay-off speed, and controls the wire speed in the painting process of the enameled wire; the present invention further evaluates the complex performance degree of the wire speed data, and under the condition that the safety and adjustability are qualified, automatically adjusts the wire speed by considering the change fluctuations, warnings, trends, and complexity of the wire speed to ensure the uniform coating effect of the enameled wire. Description of the Drawings

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0016] Figure 1 It is a flowchart of the automatic adjustment method for the enameled wire production process provided by the embodiments of the present invention; Figure 2 It is a schematic structural diagram of the automatic adjustment system for the enameled wire production process provided by the embodiments of the present invention. Detailed implementation manners

[0017] In order to enable those skilled in the art of the present technology to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0018] Embodiment 1 As Figure 1 shown, the automatic adjustment method for the enameled wire production process provided by the embodiments of the present invention, wherein the enameled wire production process mainly includes links such as the preparation of metal wire, painting, drying, curing, and detection; especially in the painting process, if the paint layer is uneven, it may lead to unqualified insulation performance or breakdown problems in subsequent use. Therefore, among the many processes of enameled wire, the specific painting process is selected for automatic adjustment to ensure the effect of uniform coating of the enameled wire. The automatic adjustment includes the following steps: S1. In the painting process of enameled wire production, obtain the real-time linear velocity data of the wire; determine the motion characteristics of the wire according to the real-time linear velocity data; In S1, specifically, the process of obtaining the real-time linear velocity data of the wire is as follows: Install a rotary encoder (incremental or absolute) on the driving roller or the take-up / reeling shaft of the production line, and convert the linear velocity by measuring the rotational speed and circumference of the roller, which has the advantages of high precision (±0.1%) and fast response; Line speed data refers to the speed at which the enameled wire moves during the coating and curing process, that is, the distance the wire moves in unit time, usually in meters per minute (m / min) or meters per second (m / s). This parameter directly affects production efficiency and coating quality. The specific reasons are: when the wire is at a low speed, the paint liquid maintains a laminar state in the coating mold and flows steadily; when the line speed exceeds a certain value, the paint liquid flow rate accelerates and the Reynolds number increases, which may cause turbulence, causing "fish scales" or local accumulation on the surface of the paint film. When high-speed wire passes through the paint liquid, it is easy to draw air into the paint film to form bubbles or pinholes. Therefore, analyzing the line speed of the wire and adjusting and controlling the enameled wire coating process can effectively improve the quality of the enameled wire. Second, specifically, a method for determining a wire motion characteristic based on real-time wire speed data includes: An analysis period is set, and the real-time wire speed data of the analysis period is analyzed in turn in terms of fluctuation degree, warning degree, and trend degree to obtain wire movement fluctuation characteristics, wire movement warning characteristics, and wire movement trend characteristics.

[0019] The real-time line speed data of the analysis period is analyzed in turn for fluctuation degree, and the process of obtaining the fluctuation characteristics of the wire movement is as follows: Acquire real-time line speed data of the analysis period, obtain a line speed curve by fitting the least square method, extract all peak points in the line speed curve, construct a line speed peak time series, perform fluctuation calculation on the line speed peak time series, and obtain the fluctuation characteristics of the wire movement. For example, the fluctuation calculation method can adopt a variance formula; The real-time wire speed data of the analysis period is analyzed in turn for early warning degree, and the process of obtaining the wire movement early warning characteristics is as follows: Acquire the real-time line speed data of the analysis period, obtain the line speed curve by least squares fitting, extract all the peak points in the line speed curve, construct the line speed peak time series, extract the maximum value in the line speed peak time series, calculate the closeness between the maximum value and the preset risk value, and obtain the wire movement warning characteristics; Exemplarily, the process of calculating the degree of proximity between the maximum value and the preset risk value is: Calculate the difference between the preset risk value and the maximum value, calculate the ratio of the difference to the preset risk value, and obtain the degree of proximity between the maximum value and the preset risk value, which is the wire movement warning feature; It should be explained that the preset risk value is pre-set by the technicians according to the enameled wire production process, that is, the certain value mentioned above. If the value is exceeded, the paint film surface will be abnormal. The real-time line speed data of the analysis period is analyzed in turn to obtain the trend characteristics of the wire movement: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least squares method, extract all the peak points in the linear velocity curve, construct the time series of the peak points of the linear velocity, and obtain the movement trend characteristics of the wire according to the distribution difference between the central linear velocity data of the initial sequence segment and other linear velocity data; Exemplarily, the specific process of the distribution difference between the central linear velocity data of the initial sequence segment and other linear velocity data is as follows:

[0020] In the formula, Y represents the movement trend characteristics of the wire; V1 represents the first linear velocity data in the time series of the peak points of the linear velocity; V2 represents the central linear velocity data in the time series of the peak points of the linear velocity; V3 represents the last linear velocity data in the time series of the peak points of the linear velocity; a represents a preset hyperparameter, and the preferred value of a is 0.01, which is used to prevent the denominator from being 0; | | represents the absolute value symbol; It can be understood that: the movement fluctuation characteristics of the wire reflect the stability of the linear velocity during wire coating within the analysis period. The greater the movement fluctuation characteristics of the wire, the more unstable the linear velocity of the wire, and the higher the risk of wire coating; the movement warning characteristics of the wire reflect the proximity of the linear velocity to the risk value during wire coating within the analysis period. The smaller the movement warning characteristics of the wire, the higher the warning degree of the linear velocity during wire coating, and the higher the risk of wire coating; the movement trend characteristics of the wire reflect the growth trend of the linear velocity during wire coating within the analysis period. The greater the movement trend characteristics of the wire, the more the linear velocity of the wire shows a growth trend, and the higher the risk of wire coating; S2. Input the movement characteristics of the wire and the speed characteristics of the wire into a pre-constructed risk prediction model for prediction processing to obtain the movement risk coefficient of the wire; and determine the danger degree of the linear velocity according to the movement risk coefficient of the wire; In S2, specifically, the construction of the risk index prediction model is based on the training of a Multi-Layer Perceptron (MLP) model. The input layer of this model can be set with multiple nodes to input data such as the mean wire speed, wire movement fluctuation characteristics, wire movement warning characteristics, and wire movement trend characteristics. To improve the model's expressive ability and generalization ability and reduce the complex calculation degree of parameters, the hidden layer can be divided into multiple hidden layer groups. Each hidden layer group can share a part of the data, which helps to reduce the model's computational burden and improve the efficiency of training and prediction. At the same time, multiple hidden layer groups can help the model better generalize to unseen data. Each group can focus on learning different types of data features, process data in different regions, and more effectively capture local patterns and signs in the dataset, thereby improving the model's adaptability to different data distributions and reducing the risk of overfitting. Preferably, 3 hidden layer groups can be selected to process the data of the mean wire speed, wire movement fluctuation characteristics, wire movement warning characteristics, and wire movement trend characteristics input by multiple input nodes in the input layer. During the entire model training process, the attention mechanism needs to be used to fuse the outputs of multiple hidden layer groups to obtain the wire movement risk coefficient; It should be explained that the wire movement risk coefficient includes 2 levels, corresponding to 1 and 2 respectively. 1 indicates that the movement speed of the wire during the coating process is in a safe state, and 2 indicates that the movement speed of the wire during the coating process is in a dangerous state; The technical solution of the embodiment of the present invention: In the painting process of enameled wire production, obtain the real-time wire speed data of the wire; determine the wire movement characteristics according to the real-time wire speed data; input the wire movement characteristics and wire speed characteristics into a pre-constructed risk prediction model for prediction processing to obtain the wire movement risk coefficient. The present invention takes the wire speed of enameled wire as the analysis object, comprehensively analyzes the wire speed from fluctuations, warnings, and trends, and comprehensively and accurately evaluates the risk that the current wire speed exceeds a certain value, so that when the movement speed of the wire is in a dangerous state, the tension control of the coating die can be timely performed.

[0021] Embodiment 2 Based on the above Embodiment 1, the automatic adjustment further includes the following steps: S3. Based on the determined wire speed danger level, determine the steepness of all sequences according to the real-time wire speed data, and generate an automatic adjustment strategy through the wire running complexity output by the steepness of all sequences; In S3, specifically, the method for determining the wire movement complexity according to the real-time wire speed data is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least squares method, extract all peak segments, construct the time series of the linear velocity peak segments, calculate the total steepness value of all sequences, and obtain the complexity of wire movement; It should be explained that: the peak segment refers to the curve segment connecting the peak point with the adjacent valley point obtained by taking any peak point in the linear velocity curve as the reference point and obtaining the adjacent valley point of the reference point, which is defined as the peak segment; Exemplarily, calculating the steepness of all sequences comprehensively considers the degree of the steepness coefficient and the frequency coefficient of each sequence. Specifically, the steepness coefficient and the frequency coefficient are multiplied to obtain the steepness; Among them, the calculation process of the steepness coefficient of each sequence is as follows: Taking any peak point as the reference point, determine the peak height H: , where y peak is the peak vertex value, and y baseline is the adjacent valley value; Calculate the half-height position y HM : ; Determine the left and right boundaries of the half-height width: the left boundary x left : Search left from the peak vertex to find the first position where the coordinate y i ≤y HM ; the right boundary x right : Search right from the peak vertex to find the first position where the coordinate y i ≤y HM ; Calculate the half-height width F WHM : ; The steepness coefficient is obtained by calculating the ratio of the half-height width F WHM to the peak height H; The calculation process of the frequency coefficient of each sequence is as follows: Taking any peak point as the reference point, determine the positions of the two adjacent peak points on the left and right of the reference point, and calculate the time length between the two adjacent peak points on the left and right; calculate the ratio of the time length to the analysis period to obtain the frequency coefficient; It can be understood that: the steepness coefficient reflects the proportional relationship between the height and the half-height width of each peak. The smaller the steepness coefficient, the higher the degree of change in the linear velocity of the peak segment; the frequency coefficient reflects the proportional relationship of each peak in time. The smaller the frequency coefficient, the higher the degree of change in time of the peak segment; therefore, the complexity of wire movement can effectively evaluate the complexity of the change in linear velocity within the analysis period. The smaller the complexity of wire movement, the higher the degree of change in linear velocity, which can not only characterize the urgency of the current change in linear velocity data, but also characterize the controllability of the subsequent automatic adjustment of the linear velocity; Second, specifically, the method for generating an automatic adjustment strategy based on the wire running complexity obtained from the steepness output of all sequences is as follows: Compare the wire running complexity with the wire running complexity threshold: If the wire running complexity exceeds the wire running complexity threshold, generate an adjustment instruction; when receiving the adjustment instruction, control the wire speed through the conveying device; If the wire running complexity does not exceed the wire running complexity threshold, generate a stop instruction; when receiving the stop instruction, stop the enameling process of the enameled wire and repair the relevant equipment involved in its enameling process; S4. When receiving the adjustment instruction in the automatic adjustment strategy, extract the maximum steepness and wire movement characteristics in the sequence, input them into the multi-dimensional learning model to obtain the wire pay-off speed regulation ratio, and control the wire speed in the enameling process of the enameled wire; In S4, the construction process of the multi-dimensional learning model is as follows: The multi-dimensional learning model is constructed based on the BP neural network. The multi-dimensional learning model includes 1 input layer, 1 hidden layer, and 1 output layer. The input layer contains 5 nodes. The inputs of the input layer are the maximum steepness, wire movement fluctuation characteristics, wire movement warning characteristics, and wire movement trend characteristics in the extraction sequence. The hidden layer contains 7 nodes. The activation function of the hidden layer is the ReLU activation function. The output layer contains 2 nodes. The output of the output layer is the wire pay-off speed regulation ratio. The activation function of the output layer is the linear activation function. The learning rate of the multi-parameter learning model (learning_rate) is 0.001, the loss function (loss) is the MSE mean square error, the optimization algorithm (optimizers) is Adam, the number of training epochs (epochs) is 500, and the batch size (batch_size) is 64; The process of controlling the wire speed in the enameling process of the enameled wire is as follows: Send the wire pay-off speed regulation ratio to the wire pay-off controller to adjust and control the wire pay-off speed; Technical solution of the embodiment of the present invention: Based on the determined risk degree of the linear velocity, determine the steepness of all sequences according to the real-time linear velocity data, generate an automatic adjustment strategy through the wire running complexity obtained from the steepness of all sequences; when receiving the adjustment instruction in the automatic adjustment strategy, extract the maximum steepness and wire movement characteristics in the sequence, input them into the multi-dimensional learning model to obtain the wire pay-off speed, and control the linear velocity in the painting process of the enameled wire; the present invention further evaluates the complexity degree of the linear velocity data, and under the condition that the safety and adjustability are qualified, automatically adjust the linear velocity by considering the change fluctuation, early warning, trend and comprehensive degree of complexity of the linear velocity to ensure the uniform coating effect of the enameled wire.

[0022] Embodiment 4 As Figure 2 shown, the automatic adjustment system for the enameled wire production process provided by the embodiment of the present invention includes: Analysis module: In the painting process of enameled wire production, obtain the real-time linear velocity data of the wire; determine the wire movement characteristics according to the real-time linear velocity data; In this embodiment, the process of obtaining the real-time linear velocity data of the wire is specifically as follows: Install a rotary encoder on the production line drive roller or the take-up / reeling shaft; The method for determining the wire movement characteristics according to the real-time linear velocity data includes: Set an analysis period, and analyze the real-time linear velocity data of the analysis period for the fluctuation degree, early warning degree, and trend degree in turn to obtain the wire movement fluctuation characteristics, wire movement early warning characteristics, and wire movement trend characteristics.

[0023] The process of analyzing the real-time linear velocity data of the analysis period for the fluctuation degree in turn to obtain the wire movement fluctuation characteristics is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least square method, extract all the peak points in the linear velocity curve, construct the linear velocity peak time series, and perform fluctuation calculation on the linear velocity peak time series to obtain the wire movement fluctuation characteristics; The process of analyzing the real-time linear velocity data of the analysis period for the early warning degree in turn to obtain the wire movement early warning characteristics is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least square method, extract all the peak points in the linear velocity curve, construct the linear velocity peak time series, extract the maximum value in the linear velocity peak time series, and calculate the proximity degree between the maximum value and the preset risk value to obtain the wire movement early warning characteristics; The process of analyzing the real-time linear velocity data of the analysis period for the trend degree in turn to obtain the wire movement trend characteristics is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least squares method, extract all the peak points in the linear velocity curve, construct the time series sequence of the linear velocity peaks, and obtain the movement trend characteristics of the wire according to the distribution difference between the central linear velocity data of the initial sequence segment and other linear velocity data.

[0024] Evaluation module: Input the movement characteristics of the wire and the velocity characteristics of the wire into a pre-constructed risk prediction model for prediction processing to obtain the movement risk coefficient of the wire; and determine the danger degree of the linear velocity according to the movement risk coefficient of the wire. In this embodiment, the risk index prediction model is constructed based on the training of a Multi-Layer Perceptron (MLP) model. The input layer of this model can be set with multiple nodes to input data such as the average wire velocity, the movement fluctuation characteristics of the wire, the movement warning characteristics of the wire, and the movement trend characteristics of the wire. In order to improve the expression ability of the model and improve the generalization ability of the model, and reduce the complex calculation degree of parameters, the hidden layer can be divided into multiple hidden layer groups, and each hidden layer group can share a part of the data, which helps to reduce the calculation burden of the model and improve the efficiency of training and prediction. At the same time, multiple hidden layer groups can help the model better generalize to unseen data. Each group can focus on learning different types of data characteristics, process data in different regions, and more effectively capture local patterns and signs in the data set, thereby improving the adaptability of the model to different data distributions and reducing the risk of overfitting; preferably, 3 hidden layer groups can be selected to process the data of the average wire velocity, the movement fluctuation characteristics of the wire, the movement warning characteristics of the wire, and the movement trend characteristics of the wire input by multiple input nodes of the input layer; during the entire model training process, the attention mechanism needs to be used to fuse the outputs of multiple hidden layer groups to obtain the movement risk coefficient of the wire. It should be explained that: the movement risk coefficient of the wire includes 2 levels, corresponding to 1 and 2 respectively. 1 means that the movement speed of the wire during the coating process is in a safe state, and 2 means that the movement speed of the wire during the coating process is in a dangerous state.

[0025] Strategy module: Based on the determined danger degree of the linear velocity, determine the steepness of all sequences according to the real-time linear velocity data, and generate an automatic adjustment strategy through the wire running complexity obtained from the steepness of all sequences. In this embodiment, the method for determining the movement complexity of the wire according to the real-time linear velocity data is as follows: Obtain the real-time linear velocity data of the analysis period, fit the linear velocity curve by the least squares method, extract all peak segments, construct the time series sequence of the linear velocity peak segments, calculate the total steepness value of all sequences, and obtain the movement complexity of the wire. It should be noted that: the peak segment refers to a curve segment defined by taking any peak point in the linear velocity curve as a reference point, obtaining the adjacent valley point of the reference point, and connecting the peak point and the adjacent valley point; Exemplarily, calculate the steepness of all sequences by comprehensively considering the degree of the steepness coefficient and the frequency coefficient of each sequence. Specifically, multiply the steepness coefficient and the frequency coefficient to obtain the steepness.

[0026] Adjustment module: When receiving the adjustment instruction in the automatic adjustment strategy, extract the maximum steepness and wire movement characteristics in the sequence, input them into the multi-dimensional learning model to obtain the wire pay-off speed regulation ratio, and control the linear velocity in the painting process of the enameled wire; In this embodiment, the construction process of the multi-dimensional learning model is as follows: The multi-dimensional learning model is constructed based on the BP neural network. The multi-dimensional learning model includes 1 input layer, 1 hidden layer, and 1 output layer. The input layer contains 5 nodes, and the inputs of the input layer are the maximum steepness, wire movement fluctuation characteristics, wire movement warning characteristics, and wire movement trend characteristics extracted from the sequence. The hidden layer contains 7 nodes, and the activation function of the hidden layer is the ReLU activation function. The output layer contains 2 nodes, and the output of the output layer is the wire pay-off speed regulation ratio. The activation function of the output layer is the linear activation function. The learning rate (learning_rate) of the multi-parameter learning model is 0.001, the loss function (loss) is the MSE mean square error, the optimization algorithm (optimizers) is Adam, the number of training epochs (epochs) is 500, and the batch size (batch_size) is 64; The process of controlling the linear velocity in the painting process of the enameled wire is as follows: Send the wire pay-off speed regulation ratio to the wire pay-off controller to adjust and control the speed of wire pay-off.

[0027] The above has described a detailed description of an embodiment of the present invention, but the content described is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. The method for automatically adjusting the production process of enameled wire is characterized in that: The following steps are involved: S1. In the painting process of enameled wire production, obtaining the real-time linear velocity data of the wire; determining the movement characteristics of the wire according to the real-time linear velocity data; S2. Inputting the wire movement characteristics and wire speed characteristics into a pre-built risk prediction model for prediction processing to obtain a wire movement risk coefficient; and determining the wire speed risk level according to the wire movement risk coefficient; S3. Based on the determined line speed danger level, the steepness of all sequences is determined according to the real-time line speed data, and the wire running complexity obtained by the steepness output of all sequences is used to generate an automated adjustment strategy; S4. When receiving the adjustment instruction in the automatic adjustment strategy, the maximum steepness and wire movement characteristics in the sequence are extracted and input into the multi-dimensional learning model to obtain the wire pay-off speed control ratio.

2. The method for automatically adjusting the enameled wire production process according to claim 1, characterized in that: The method for determining the wire motion characteristics based on real-time wire speed data includes: An analysis period is set, and the real-time wire speed data of the analysis period is analyzed in turn in terms of fluctuation degree, warning degree, and trend degree to obtain wire movement fluctuation characteristics, wire movement warning characteristics, and wire movement trend characteristics.

3. The method for automatically adjusting the enameled wire production process according to claim 2 is characterized in that: The real-time line speed data of the analysis period is analyzed in turn for fluctuation degree, and the process of obtaining the fluctuation characteristics of the wire movement is as follows: The real-time linear speed data of the analysis period is obtained, the linear speed curve is obtained by least square fitting, all the peak points in the linear speed curve are extracted, the linear speed peak timing sequence is constructed, the fluctuation of the linear speed peak timing sequence is calculated, and the fluctuation characteristics of the wire movement are obtained.

4. The method for automatically adjusting the enameled wire production process according to claim 2, characterized in that: The real-time wire speed data of the analysis period is analyzed in turn for early warning degree, and the process of obtaining the wire movement early warning characteristics is as follows: The real-time line speed data of the analysis period is obtained, and the line speed curve is obtained by least squares fitting. All the peak points in the line speed curve are extracted, and the line speed peak time series is constructed. The maximum value in the line speed peak time series is extracted, and the degree of proximity between the maximum value and the preset risk value is calculated to obtain the wire movement warning characteristics.

5. The method for automatically adjusting the enameled wire production process according to claim 2, characterized in that: The real-time line speed data of the analysis period is analyzed in turn to obtain the trend characteristics of the wire movement: The real-time linear speed data of the analysis period is obtained, and the linear speed curve is obtained by least square fitting. All the peak points in the linear speed curve are extracted, and the linear speed peak time series is constructed. According to the distribution difference between the central linear speed data of the initial sequence segment and other linear speed data, the wire movement trend characteristics are obtained.

6. The method for automatically adjusting the enameled wire production process according to claim 1, characterized in that: The wire movement risk coefficient includes two levels, corresponding to 1 and 2 respectively. 1 means that the wire movement speed is in a safe state during the coating process, and 2 means that the wire movement speed is in a dangerous state during the coating process.

7. The method for automatically adjusting the enameled wire production process according to claim 5, characterized in that: The method for determining the complexity of wire motion based on real-time wire speed data is as follows: The real-time linear velocity data of the analysis period is obtained, the linear velocity curve is obtained by least square fitting, all peak segments are extracted, the linear velocity peak segment time series is constructed, the total steepness value of all sequences is calculated, and the wire motion complexity is obtained; Calculate the steepness of all series by combining the steepness coefficient and frequency coefficient of each series.

8. The method for automatically adjusting the enameled wire production process according to claim 7, characterized in that: The method of generating an automatic adjustment strategy based on the wire running complexity obtained by the steepness output of all sequences is as follows: If the wire operation complexity exceeds the wire operation complexity threshold, an adjustment instruction is generated; If the wire operation complexity does not exceed the wire operation complexity threshold, a stop instruction is generated.

9. The method for automatically adjusting the enameled wire production process according to claim 8, characterized in that: The input of the input layer is the maximum steepness in the extraction sequence, the wire movement fluctuation characteristics, the wire movement warning characteristics, and the wire movement trend characteristics. The hidden layer contains 7 nodes, and the activation function of the hidden layer is the ReLU activation function. The output layer contains 2 nodes, and the output of the output layer is the wire pay-off speed control ratio.

10. The automatic adjustment system for the production process of enameled wire is characterized by: The system is used to execute the method described in any one of claims 1 to 9, and the system comprises: Analysis module: In the painting process of enameled wire production, the real-time linear velocity data of the wire is obtained; the movement characteristics of the wire are determined based on the real-time linear velocity data; Evaluation module: input the wire movement characteristics and wire speed characteristics into the pre-built risk prediction model for prediction processing to obtain the wire movement risk coefficient; and determine the wire speed danger level according to the wire movement risk coefficient; Strategy module: Based on the determined line speed danger level, the steepness of all sequences is determined according to the real-time line speed data, and the wire running complexity obtained by the steepness output of all sequences is used to generate an automated adjustment strategy; Adjustment module: When receiving the adjustment instruction in the automated adjustment strategy, it extracts the maximum steepness and wire movement characteristics in the sequence and inputs them into the multi-dimensional learning model to obtain the wire pay-off speed control ratio.

Images