A working process monitoring method and system for winding machine

Abstract

The present application relates to the field of data processing technology, and more specifically to a method and system for monitoring the working process of a winding machine. The method comprises: acquiring and preprocessing the winding machine motor speed data to obtain standardized speed data; smoothing the standardized speed data using an improved EWMA algorithm to obtain smoothed data; and performing anomaly detection on the smoothed data to obtain anomaly results. Specifically, the solution of the present application can effectively improve the accuracy of the collected winding machine motor speed data, thereby enabling accurate anomaly detection.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and system for monitoring the working process of a winding machine. Background Art

[0002] In industrial manufacturing, wrapping machines are key equipment used in processes such as packaging and composite molding. Their operational stability directly impacts production efficiency and product quality. As the core drive component of wrapping machines, real-time speed monitoring and anomaly detection are crucial technologies for ensuring equipment reliability.

[0003] Traditional methods typically use fixed threshold alarms or data smoothing techniques based on simple moving averages (SMAs) for speed monitoring. However, these methods lack adaptability under dynamic operating conditions, making it difficult to effectively capture gradual anomalies and subject to detection lag. Real-time anomaly detection of motor speed data is crucial for monitoring the working process of winding machines, as motor speed is a key indicator of their operating performance. Real-time detection of motor speed data can promptly identify and address speed anomalies, ensuring the stable operation of the winding machine and avoiding downtime due to faults, thereby improving production efficiency. The EWMA (Exponentially Weighted Moving Average) algorithm is a powerful and flexible time series smoothing technique that captures the latest data trends by assigning different weights to observations at different time points. This algorithm can effectively improve the accuracy of collected winding machine motor speed data, enabling accurate anomaly detection.

[0004] In the related technology, such as the Chinese patent application document with application publication number CN115045807A, entitled A Method for Detecting Abnormal Generator Speed of a Wind Turbine Generator Set, it is disclosed that by utilizing the variance relationship before and after abnormality detection, the abnormality detection threshold can be scientifically and reasonably selected. The above method performs abnormality detection within the entire generator speed range, writes the rules and determined thresholds fitted from historical samples into the controller, realizes online abnormality detection, and provides timely warnings in advance to ensure safe and stable operation of the unit. However, the above method may cause the data to exhibit different noise levels, resulting in its inability to effectively balance the sensitivity to new data and dependence on historical data, that is, it will be affected by changes in noise levels, thereby affecting the accuracy of abnormality detection.

[0005] Therefore, it is urgent to develop a motor speed anomaly detection method with adaptive threshold adjustment and real-time streaming processing to improve the operation reliability and production efficiency of the winding machine. Summary of the Invention

[0006] The purpose of this application is to propose a working process monitoring method and system for a winding machine, so as to solve the problem that the fixed smoothing factor is used in the prior art, resulting in the data showing different noise levels, making it impossible to effectively balance the sensitivity to new data and the dependence on historical data, that is, it will be affected by the change of noise level, thereby affecting the accuracy of anomaly detection.

[0007] In a first aspect, the present application provides a method for monitoring a working process of a winding machine, comprising:

[0008] Obtain the winding machine motor speed data and preprocess it to obtain standardized speed data;

[0009] Adaptively smoothing the normalized speed data using an improved EWMA algorithm to obtain smoothed data;

[0010] Performing anomaly detection on the smoothed data to obtain an anomaly result;

[0011] Among them, the smoothing factor in the improved EWMA algorithm is positively correlated with the noise level and the initial smoothing factor. The noise performance level is calculated by the noise performance degree of the speed data point. The noise performance degree is positively correlated with the fluctuation frequency of the data points within the target range. The noise performance degree is negatively correlated with the distance between adjacent extreme points. The data point fluctuation frequency is the ratio of the number of extreme points in the target range to the length of the target range.

[0012] The above scheme obtains the winding machine motor speed data and preprocesses it to obtain standardized speed data, uses the improved EWMA algorithm to smooth the standardized speed data to obtain smoothed data, performs anomaly detection on the smoothed data to obtain abnormal results. The present application calculates the noise performance level by analyzing the changing characteristics of the winding machine motor speed data, and adaptively adjusts the smoothing factor of the EWMA algorithm according to the noise performance level to improve the noise resistance of the EWMA algorithm, thereby improving the accuracy of anomaly detection of the winding machine motor speed data.

[0013] Optionally, the noise performance level of the motor speed data point specifically includes:

[0014] Starting from the smoothing point, each speed data point is traversed forward in turn. If the degree of belonging of the currently traversed data point is greater than the set threshold, the data point is used as the target point. The traversal is stopped until M consecutive data points do not meet the joining conditions. All target points are obtained to constitute the target data, and the mean of the noise performance degree of all data points in the target data is used as the noise performance level of the target data.

[0015] The above scheme traverses each speed data point in sequence until M consecutive data points do not meet the inclusion conditions, then stops traversing, and takes the average of the noise performance levels of all data points in the target data as the noise performance level of the target data, thereby selecting appropriate data points to add to the target data to calculate the accurate noise performance level of the motor speed data point.

[0016] Optionally, the adaptive smoothing process is performed on the standardized speed data, specifically:

[0017] ;

[0018] Where s is the adjusted smoothing factor, s0 is the initial smoothing factor, and P is the noise performance level of the target data.

[0019] Optionally, the calculation formula of the attribution degree is:

[0020] ;

[0021] in, Indicates the degree of belonging of the currently traversed data point b, Indicates the degree of noise performance of data point b, represents the mean value of the noise performance in the target data, then It is the difference between the current traversed data point b and the mean of the noise performance of all data points in the target data. Its negative exponential power can represent the degree of consistency between the current traversed data point b and the noise interference level of the data in the target data, that is, the degree of belonging, where exp() is an exponential function.

[0022] In the above scheme, the attribution degree is calculated to reduce the randomness of the calculation of the noise performance degree of individual data points, which leads to the problem of insufficient target data set, thereby improving accuracy.

[0023] Optionally, the extreme point is obtained in the following manner:

[0024] If z i,j is the jth speed data point within the target range, and the adjacent data point is z i,j-1 z i,j+1 , if z i,j Satisfy z i,j-1 >z i,j And z i,j+1 ≥z i,j or z i,j-1 ≥z i,j And z i,j+1 >z i,j , then z i,j is identified as a minimum point if z i,j Satisfy z i,j-1≤z i,j And z i,j+1 <z i,j or z i,j-1 <z i,j And z i,j+1 ≤z i,j , then z i,j Identified as a maximum point.

[0025] The above scheme monitors data changes and obtains data fluctuations by satisfying the condition that one of the two adjacent data points is greater than or not less than the current minimum point or maximum point, thereby determining the data fluctuation frequency.

[0026] Optionally, the noise performance level of the motor speed data point is specifically:

[0027] ;

[0028] in, Indicates the speed data point The noise performance level; n i Indicates the speed data point The number of extreme points in the target range, N i is the speed data point z i The number of data points in the target range, D i Indicates the speed data point z i The variance of the distances between all adjacent extreme points within the target range, norm() is a linear normalization function.

[0029] Optionally, the abnormality detection of the winding machine motor speed data includes:

[0030] The smoothed data is restored to speed data using the inverse normalization formula of Z-socre and compared with the motor speed data range set by the winding machine to detect abnormalities in the motor speed data of the winding machine.

[0031] The above scheme uses the smoothed data to restore it to speed data using the inverse normalization formula of Z-socre, and compares it with the motor speed data range set by the winding machine to accurately detect abnormalities in the winding machine motor speed data.

[0032] Optionally, the acquiring and preprocessing of the winding machine motor speed data further includes: performing data cleaning and data denoising on the winding machine motor speed data.

[0033] In a second aspect, the present application provides a working process monitoring system for a winding machine, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned working process monitoring method for a winding machine when executing the computer program.

[0034] The beneficial effects of this application are:

[0035] The solution of the present application can detect anomalies in the motor speed data of the winding machine through the EWMA algorithm, and can well meet the real-time requirements of anomaly detection of the motor speed data of the winding machine. Because the EWMA algorithm is calculated through a recursive formula, it only needs to store the EWMA value of the previous time point and the current observation value, which greatly saves computing resources and storage space. It can calculate the noise performance level by analyzing the change characteristics of the motor speed data of the winding machine, and thus adaptively adjust the smoothing factor of the EWMA algorithm according to the noise performance level, and can better adapt to the changes in the motor speed data and the noise level during the operation of the winding machine. When the noise level is high, the algorithm can automatically increase its dependence on new data to reduce the impact of noise on anomaly detection. When the noise level is low, the EWMA algorithm can rely more on historical data to provide more stable anomaly detection results, thereby avoiding the fixed smoothing factor that makes the EWMA algorithm difficult to adapt to changes in motor speed data. The adaptive smoothing factor in the present application improves the accuracy of anomaly detection of the motor speed data of the winding machine, and also improves the efficiency of monitoring during the operation of the winding machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A flowchart schematically illustrates a method for monitoring a working process of a winding machine in this embodiment;

[0037] Figure 2 The structural block diagram of a working process monitoring system for a winding machine in this embodiment is schematically shown. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0039] It should be understood that the size of the serial numbers of the steps in the following 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.

[0040] The present application is directed to a monitoring solution for the working process of a winding machine, which requires performing abnormality detection on the speed data of the motor of the winding machine by using an EWMA algorithm.

[0041] Therefore, based on the above content, the present application provides a method and system for monitoring the working process of a winding machine to accurately detect abnormalities in the speed data of the winding machine motor.

[0042] Specifically, if Figure 1 As shown, a working process monitoring method for a winding machine in this embodiment includes the following steps:

[0043] Step S1, obtaining the winding machine motor speed data and performing preprocessing.

[0044] Specifically, in this embodiment, a speed sensor is used to collect motor speed data for a wrapping machine at a 1-second interval. The collected motor speed data is then acquired. Because the speed data has a wide dimensional range, the collected motor speed data is standardized to facilitate subsequent calculations. This standardized data can then be used for subsequent analysis and calculations. The wrapping machine is an automated packaging device used to protect items from damage and contamination, thereby improving transportation efficiency and safety. The motor speed refers to the speed at which the wrapping machine motor rotates within a fixed time period while the wrapping machine is operating. In this application, the wrapping machine motor speed data is acquired and preprocessed.

[0045] Optionally, the method further includes performing data cleaning and data denoising on the speed data of the winding machine motor. The time interval for collecting the speed of the winding machine in this embodiment is for illustration only and should not be construed as limiting the present application.

[0046] In this embodiment, by collecting the winding machine motor speed data and normalizing the winding machine motor speed data to obtain standardized data, the EWMA algorithm is used to perform abnormality detection on the winding machine motor speed data, providing data that can better adapt to the changes in motor speed data and noise levels during the working process of the winding machine.

[0047] Step S2: Adaptively adjust the smoothing factor of the EWMA algorithm by analyzing the change characteristics of the winding machine motor speed data, thereby smoothing the motor speed data.

[0048] In this embodiment, since the speed sensor is easily affected by the surrounding environmental noise, such as the operating noise of other electrical equipment in the surrounding environment, when collecting the speed data of the winding machine motor, the collected speed data may deviate from the actual speed data. Therefore, abnormality detection based on such speed data is definitely inaccurate, which may lead to false alarms or omissions of abnormal working of the winding machine motor.

[0049] Therefore, in this embodiment, the collected rotation speed data is smoothed by the EWMA algorithm to eliminate the influence of noise thereon, thereby improving the accuracy of abnormality detection.

[0050] Specifically, when using the existing EWMA algorithm for data smoothing, a fixed smoothing factor is often used to smooth the data. However, due to the working environment of the winding machine, the data collected in different time periods are subject to different degrees of noise interference. In this case, the fixed smoothing factor may not be able to change according to the degree of noise interference, resulting in poor smoothing effect, which in turn affects the accuracy of anomaly detection. Therefore, in this embodiment, the noise performance level is calculated by analyzing the changing characteristics of the winding machine motor speed data, and the smoothing factor of the EWMA algorithm is adaptively adjusted according to the noise performance level to improve the adaptability of the EWMA algorithm to noise and further improve the accuracy of the speed data. The specific process is as follows:

[0051] Step S21: Calculate the noise performance level of the data point based on the variation characteristics of the winding machine motor speed data.

[0052] Specifically, when the speed sensor collects the speed data of the winding machine motor, it will be affected by the operating noise of other electrical equipment in the surrounding environment. Such an influence will cause the speed data to fluctuate up and down with high frequency and irregularity. Therefore, the noise performance degree can be calculated by using the fluctuation frequency and fluctuation regularity of the speed data points within the target range. The higher the fluctuation frequency and the worse the fluctuation regularity, the higher the noise performance degree.

[0053] It should be noted that the fluctuation frequency of the speed data points within the target range can be expressed by the ratio of the number of extreme points within the target range to the length of the target range, because the appearance of extreme points means that the data sequence has fluctuated. Therefore, the number of extreme points indicates the number of data fluctuations, and the ratio to the length of the target range indicates the number of data fluctuations per unit length, that is, the fluctuation frequency, and the ratio of the number of extreme points to the length of the target range is positively correlated with the fluctuation frequency; the regularity of fluctuations can be expressed by the consistency of the distances between all adjacent extreme points within the target range of data points, because the distances between adjacent extreme points reflect the time interval and fluctuation amplitude between adjacent extreme points, then the consistency of the distances of all adjacent extreme points can comprehensively express the regularity of its fluctuations from the time dimension and the fluctuation amplitude dimension, and consistency is negatively correlated with regularity.

[0054] Specifically, in the present application, the target range of the i-th data point in the winding machine speed data sequence is: a range of 2m+1 data points in length centered on the current speed data point, for example: m=11, and for the edge data points of the data sequence whose positions cannot be reached in the target range, they are discarded accordingly. The specific data described in this embodiment is only for illustration and should not be construed as a limitation to the present application.

[0055] At the speed data point Identify the extreme point within the target range. In this embodiment, the method for identifying the extreme point is as follows: Let z i,j is the speed data point The jth data point in the target range has z adjacent data points. i,j-1 z i,j+1 , if z i,j Satisfy z i,j-1 >z i,j And z i,j+1 ≥z i,j or z i,j-1 ≥z i,j And z i,j+1 >z i,j , then z i,j is identified as a minimum point if z i,j Satisfy z i,j-1 ≤z i,j And z i,j+1 <z i,j or z i,j-1 <z i,j And z i,j+1 ≤z i,j , then z i,j The speed data point can be identified by the working process monitoring method of the winding machine in this application. All extreme points within the target range are recorded as ,in, Indicates the speed data point The jth extreme point in the target range, n i Indicates the speed data point The number of extreme points within the target range. It should be noted that extreme point identification can also be performed using existing methods, such as the differential method. The difference is that the extreme points identified by the differential method meet the requirement that the two adjacent data points of the extreme point are both greater than or less than the current point. In this embodiment, data changes can also be regarded as a data fluctuation.

[0056] After the extreme point is identified, the speed data point can be used The number of extreme points and speed data points within the target range The ratio of the number of data points within the target range represents its fluctuation frequency. The distance between adjacent extreme points can be expressed using the Euclidean distance. The variance of the distances between all adjacent extreme points within the target range can represent the consistency. The larger the variance, the worse the consistency. So the speed data point The calculation formula for the noise performance degree is:

[0057] ;

[0058] in, Indicates the speed data point The noise performance level; n i Indicates the number of extreme points in the speed data point, N i is the speed data point The number of data points within the target range; D i It represents the variance of the distance between all adjacent extreme points in the speed data point. The variance is negatively correlated with consistency, and the consistency is negatively correlated with the regularity of fluctuation. The variance is positively correlated with the regularity of fluctuation. norm() is a linear normalization function. The role of the norm() linear normalization function is to unify the dimensions of the fluctuation frequency and the regularity of fluctuation. This is to limit the noise level to the range of 0-1.

[0059] In an embodiment of the present application, the noise performance degree of the data point is calculated by the change characteristics of the winding machine motor speed data, and the noise performance degree is used to characterize the noise performance level of the winding machine, and the smoothing factor of the EWMA algorithm is adaptively adjusted according to the noise performance level, which can better adapt to the changes in the motor speed data and the noise level during the working process of the winding machine.

[0060] Step S22: Calculating the noise performance level of the target data according to the noise performance degree of the speed data point.

[0061] In the embodiment of the present application, since the noise to which the speed data points of the winding machine are subject is stage-specific and the noise performance level is different at different stages, the noise performance level of the target data can be used to comprehensively judge the situation in which the target data is interfered with by noise, thereby avoiding the randomness of using only the noise performance degree of a single data point. At the same time, by using the noise performance level of the target data to comprehensively judge the situation in which the target data is interfered with by noise, the accuracy of the subsequent adaptive adjustment of the smoothing factor of the EWMA algorithm can be improved.

[0062] It should be noted that the target data is defined as follows: first, starting from the current smoothing point, traverse each speed data point in sequence. If the degree of belonging of the currently traversed data point to the target data is greater than the set threshold, the point is added to the target data. For example, if it is set to be greater than or equal to 0.76, the target data is added. The degree of belonging refers to the degree of consistency between the current traversed data point and the data in the target data. The more consistent the level of noise interference, the greater the degree of belonging. Specifically, let b be the currently traversed data point, and the calculation formula for its degree of belonging is:

[0063] ;

[0064] in, Indicates the degree of belonging of the currently traversed data point b, Indicates the degree of noise performance of data point b, represents the mean value of the noise performance in the target data, then It is the difference between the currently traversed data point b and the mean of the noise performance of all data points in the target data. In this embodiment, the negative exponential power can represent the degree of consistency between the currently traversed data point b and the noise interference level of the data in the target data, that is, the degree of belonging, where exp() is an exponential function.

[0065] Through the above-mentioned degree of attribution, each speed data point can be traversed forward in turn until there are M consecutive data points that do not meet the joining conditions, then the traversal is stopped. The purpose of this is to prevent the randomness of the calculation of the degree of noise performance of individual data points, which leads to insufficient target data sets, improve accuracy, and facilitate the determination of target data. For example: if M=2, the traversal is stopped. The specific data described in this embodiment is only for illustration and should not be a limitation of this application.

[0066] Then, the mean of the noise performance of all data points in the target data is used as the noise performance level of the target data, that is:

[0067] ;

[0068] Where P represents the noise level of the target data, N represents the number of data points in the target data, and z q Indicates the degree of noise performance of the qth data point in the target data.

[0069] In this embodiment, the noise performance level of the winding machine motor is calculated based on the noise performance level of the target data, the number of data points in the target data, and the noise performance degree of each data point, thereby ensuring that the noise performance level of the target data is dynamically calculated according to the changing characteristics of the winding machine motor speed data.

[0070] Step S23: Adaptively adjust the smoothing factor of the EWMA algorithm according to the noise performance level of the target data.

[0071] Specifically, the EWMA algorithm smoothing factor can be adaptively adjusted based on the noise performance level of the target data. The greater the noise performance level of the target data, the greater the noise interference during this period. In this case, the reliance on new data can be automatically increased to reduce the impact of historical data on the algorithm. That is, the larger the smoothing factor, the more the EWMA algorithm can rely on historical data to provide more stable data smoothing when the noise level is low. The adaptive adjustment process is as follows: set the initial smoothing factor s0 to a smaller value. In this application, s0 is only adjusted upward. The adjustment process for each smoothing is: , where s is the adjusted smoothing factor, s0 is the initial smoothing factor, and P is the noise performance level of the target data. For example, the initial smoothing factor s0 = 0.3, and the smoothing factor is dynamically adjusted according to the noise performance level of the target data.

[0072] In an embodiment of the present application, by utilizing the EWMA algorithm with an adaptive smoothing factor, the collected winding machine motor speed data is smoothed in real time, so that anomaly detection is performed based on the smoothing value, thereby avoiding the EWMA algorithm with a fixed smoothing factor from being difficult to adapt to changes in motor speed data, improving the anomaly detection effect, and further improving the efficiency of the monitoring method during the working process of the winding machine.

[0073] Step S3: performing abnormality detection on the winding machine motor speed data.

[0074] Specifically, in the embodiment of the present application, after the winding machine motor speed data collected in real time is smoothed according to the EWMA algorithm, abnormality detection can be performed on the winding machine motor speed data.

[0075] Specifically, the smoothed data is restored to speed data using the inverse normalization formula of Z-socre and then compared with the motor speed data range set by the winding machine. For example, the setting range is related to the operator's needs, that is, if the operator sets the speed of the winding machine to R (r / min), then the setting range can be In this embodiment, t=50 is preset in the example, and the t value can also be adjusted as needed. When the motor speed data exceeds the set range, the abnormal alarm mechanism is triggered to notify relevant personnel to handle it.

[0076] This application also provides a working process monitoring system for a winding machine. Figure 2As shown, the system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a working process monitoring method for a winding machine according to the present application is implemented.

[0077] The system further includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and thus will not be described in detail here.

[0078] In this application, the aforementioned memory can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, the computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium can be part of, accessible to, or connectable to a device. Any application or module described in this application can be implemented by computer-readable / executable instructions stored or otherwise retained by such a computer-readable medium.

Claims

1. A method for monitoring the working process of a winding machine, characterized in that: include: Obtain the winding machine motor speed data and preprocess it to obtain standardized speed data; The improved EWMA algorithm is used to adaptively smooth the standardized speed data to obtain smoothed data. The smoothing factor in the improved EWMA algorithm is positively correlated with the noise level and the initial smoothing factor. Smoothing is achieved by dynamically adjusting the noise performance level, which is calculated by the noise performance degree of the speed data points. The noise performance degree is positively correlated with the data point fluctuation frequency within the target range and negatively correlated with the distance between adjacent extreme value points. The data point fluctuation frequency is the ratio of the number of extreme value points within the target range to the length of the target range. Noise performance levels, including: Starting from the smoothing point, each speed data point is traversed forward in sequence. If the degree of belonging of the currently traversed data point is greater than the set threshold, the data point is taken as the target point. The traversal is stopped until M consecutive data points do not meet the inclusion conditions. All target points are obtained to form the target data, and the mean noise performance degree of all data points in the target data is taken as the noise performance level of the target data. Noise performance level, specifically: ； in, Indicates the speed data point The noise performance level; n i Indicates the number of extreme points in the speed data point, N i is the speed data point The number of data points within the target range; D i It represents the variance of the distance between all adjacent extreme points in the speed data point, and norm() is a linear normalization function; Anomaly detection is performed on the smoothed data to obtain abnormal results.

2. A method for monitoring the working process of a winding machine according to claim 1, characterized in that: The standardized speed data is subjected to adaptive smoothing processing, specifically: ； Where s is the adjusted smoothing factor, s0 is the initial smoothing factor, and P is the noise performance level of the target data.

3. The method for monitoring the working process of a winding machine according to claim 1, characterized in that: The calculation formula of the attribution degree is: ； in, Indicates the degree of belonging of the currently traversed data point b, Indicates the degree of noise performance of data point b, It represents the mean value of the noise performance in the target data, and exp() is the exponential function.

4. The method for monitoring the working process of a winding machine according to claim 1, characterized in that: The method for obtaining the extreme point is: If z i,j is the jth speed data point within the target range, and the adjacent data point is z i,j-1 、z i,j+1 , if z i,j Satisfy z i,j-1 >z i,j And z i,j+1 ≥z i,j or z i,j-1 ≥z i,j And z i,j+1 >z i,j , then z i,j is identified as a minimum point if z i,j Satisfy z i,j-1 ≤z i,j And z i,j+1 <z i,j or z i,j-1 <z i,j And z i,j+1 ≤z i,j , then z i,j Identified as a maximum point.

5. The method for monitoring the working process of a winding machine according to claim 1, characterized in that: The target range is a continuous data interval defined when analyzing the fluctuation frequency of the data points.

6. A method for monitoring the working process of a winding machine according to claim 1, characterized in that: The anomaly detection includes: The smoothed data is restored to speed data using the inverse normalization formula of Z-socre, and compared with the motor speed data range set by the winding machine to perform abnormality detection on the motor speed data of the winding machine; wherein, when the motor speed data of the winding machine is within the set motor speed data range, the abnormal alarm mechanism is not triggered; when the motor speed data of the winding machine exceeds the set motor speed data range, the abnormal alarm mechanism is triggered.

7. The method for monitoring the working process of a winding machine according to claim 1, characterized in that: The method of obtaining the winding machine motor speed data and performing preprocessing further includes: performing data cleaning and data denoising on the winding machine motor speed data.

8. A working process monitoring system for a winding machine, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the working process monitoring method for a winding machine according to any one of claims 1 to 7 is implemented.

Images