Elasticizer roller performance health monitoring and early warning method and system

Abstract

The invention discloses an elasticizer roller performance health monitoring and early warning method, which belongs to the textile machinery fault diagnosis field, and comprises the following steps: firstly, extracting hidden parameters corresponding to hidden faults, analyzing corresponding monitoring frequency bands, and calculating spectrum entropy increments of hidden parameter signals in the frequency bands; judging whether to trigger a hidden parameter acquisition and storage operation or not; secondly, hidden parameter data are extracted, a complex-dimensional coordinate system is constructed to form an anomaly evaluation graph, an evaluation trajectory is constructed, and an anomaly evaluation value is calculated to judge whether the elasticizer roller has hidden faults or not; and finally, hidden parameter data corresponding to the hidden faults are extracted to construct a hidden parameter matrix, a fault diagnosis graph is constructed, whether the corresponding hidden faults occur or not is judged, a serious evaluation value and an early warning value are calculated, and an early warning level is output for corresponding early warning, so that the hidden faults of the elasticizer roller are effectively monitored, the accuracy and timeliness of fault diagnosis are improved, and the working efficiency is improved. And stable operation and production efficiency of equipment are guaranteed.

Description

Technical Field

[0001] This invention belongs to the field of textile machinery fault diagnosis technology, specifically relating to a method and system for monitoring and early warning of the performance health of texturing machine rollers. Background Technology

[0002] In the textile production process, the performance of texturing machine rollers has a significant impact on product quality and production efficiency. Existing technologies can effectively monitor and respond to explicit faults of texturing machine rollers, such as severe vibration and sudden temperature rise. However, for latent faults, such as micro-cracks on the roller surface and early pitting corrosion of the bearing inner ring, existing technologies have significant shortcomings.

[0003] Currently, the main technical challenges in monitoring latent faults are as follows: It is difficult to effectively monitor hidden faults in the texturing machine rollers; While high-frequency monitoring can improve the ability to capture subtle features of latent faults, it can lead to a surge in data volume and significantly increase the consumption of storage and computing resources, making it difficult for monitoring systems to balance efficiency and resources in practical applications. The lack of effective assessment and early warning methods for the severity of latent faults makes it difficult to take reasonable maintenance measures in a timely manner based on the actual situation of the fault, which may lead to further deterioration of the fault, increase maintenance costs and equipment downtime. To address this, we propose a method and system for monitoring and early warning of the performance health of texturing machine rollers. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for monitoring and early warning of the performance health of texturing machine rollers, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring and early warning of the performance health of a texturing machine roller, comprising the following steps: Step 1: Extract the latent parameters corresponding to the latent faults, analyze the lower effective frequency upper limit of the latent parameters, and construct the monitoring frequency band; calculate the spectral entropy increment corresponding to the latent parameter signal within the monitoring frequency band; and determine whether to trigger the latent parameter acquisition and storage operation based on this. Step 2: Extract the stored implicit parameter data, construct a combined coordinate system and form an anomaly evaluation map, determine the center coordinates of each anomaly evaluation map, and construct the evaluation trajectory line; calculate the trajectory anomaly number, trajectory average deviation value, trajectory bending frequency, and trajectory deviation value; and comprehensively analyze to obtain the anomaly evaluation value, thereby determining whether the texturing machine roller has a hidden fault. Step 3: Extract the latent parameter data corresponding to the latent fault and construct the latent parameter matrix; analyze the principal component data of the latent fault; construct the principal component multidimensional coordinate system to form a fault diagnosis map, and determine whether the corresponding latent fault has occurred; calculate the severity assessment value and warning value of the latent fault, output the warning level and issue the corresponding warning.

[0006] Preferably, in step one, the specific process of constructing the monitoring frequency band is as follows: Extract the latent parameters corresponding to historical latent faults; for each latent parameter, record the minimum effective frequency upper limit of the latent parameter signal corresponding to that type of latent fault; Under the same latent parameter, the maximum value of the lowest effective frequency upper limit corresponding to various latent faults is taken, and 5kHz is used as a fixed lower limit to form the frequency band for calculating the spectral entropy increment corresponding to the latent parameter, which is denoted as the monitoring frequency band.

[0007] Preferably, in step one, the specific process for determining whether the implicit parameter acquisition and storage operation is triggered is as follows: Construct a simulated spectral entropy detector to calculate the spectral entropy increment corresponding to the implicit parameter signal within the monitoring frequency band in real time; The spectral entropy increment at each time point is compared with the corresponding threshold. If the spectral entropy increment is greater than the corresponding threshold, the implicit parameter acquisition operation is triggered, and the acquired implicit parameters are stored in the memory. Otherwise, if the spectral entropy increment is less than or equal to the corresponding threshold, the spectral entropy increment corresponding to the monitoring frequency band is calculated in real time.

[0008] Preferably, the specific process for triggering the implicit parameter acquisition operation is as follows: When the implicit parameter acquisition operation is triggered, the implicit parameter data frames of the trigger period and the preset buffer before and after it are acquired; the acquired implicit parameter data frames are divided into several subframes. For each subframe, the environmental noise standard deviation SC and signal entropy value XS of that data segment are obtained and normalized. Then, the pruning threshold TH is obtained using the formula: TH=SC×a1+XS×a2, where a1 and a2 are preset weighting coefficients. For each subframe, implicit parameter data below the deletion threshold is removed, while implicit parameter data above the deletion threshold is stored in memory.

[0009] Preferably, in step two, the specific process of constructing the evaluation trajectory line is as follows: Extract the implicit parameter data stored in the memory; use each implicit parameter as a different dimension to form a combined-dimensional coordinate system; For multiple latent parameter values ​​collected at the same time, determine their coordinate points in the combined coordinate system, connect the different latent parameters corresponding to the same time in the order of preset parameters to form line segments, connect the ends of these line segments to construct a closed figure, which is called the anomaly assessment figure, and mark the center coordinates of the anomaly assessment figure. Connect the center coordinates of the anomaly assessment graph at each time point with line segments in chronological order to obtain the assessment trajectory line.

[0010] Preferably, the specific process for calculating the number of trajectory anomalies and the average deviation of the trajectory is as follows: The trajectory length is obtained by calculating the sum of the distances between two adjacent center trajectory points. The trajectory length is then divided by the unit time to obtain the rate of change of the trajectory length. The trajectory length change rate is compared with the corresponding threshold to determine whether it is an abnormal fluctuation trajectory, and the deviation of the abnormal fluctuation trajectory from the corresponding threshold is calculated and recorded as the trajectory deviation degree. The number of times abnormal fluctuation trajectories occur and the mean deviation of the trajectory corresponding to all abnormal fluctuation trajectories are calculated to obtain the number of trajectory anomalies GY and the mean deviation of the trajectory GP.

[0011] Preferably, in step two, the specific process of analyzing abnormal evaluation values ​​and determining whether the texturing machine roller has a latent fault is as follows: For any three consecutive adjacent points At1, Bt2, and Ct3 on the central trajectory line, calculate the trajectory angle at point Bt2; If the angle between the trajectory lines is greater than the corresponding threshold, a curve in the trajectory will be marked at point Bt2. The trajectory bending frequency GQ is obtained by counting the number of bends in the trajectory line and dividing by the time length of the trajectory line. Calculate the standard deviation of the deviation distance between each point on the current evaluation trajectory line and the corresponding benchmark position of the preset standard evaluation trajectory line to obtain the trajectory deviation value GB; After normalizing the number of trajectory anomalies GY, the average deviation of the trajectory GP, the trajectory bending frequency GQ, and the trajectory deviation value GB, the anomaly evaluation value YP is obtained using the formula: YP=GY×b1+GP×b2+GQ×b3+GB×b4, where b1, b2, b3, and b4 are preset weight coefficients. If the abnormal evaluation value is greater than the corresponding threshold, it is determined that the texturing machine roller has a hidden fault.

[0012] Preferably, in step three, the specific process for determining whether a corresponding latent fault has occurred is as follows: Extract the latent parameter data corresponding to the latent faults in the texturing machine rollers and construct a latent parameter matrix; For each type of latent fault, the latent parameter matrix is ​​analyzed to obtain the corresponding principal component data; Using each principal component as a coordinate axis, a corresponding multidimensional coordinate system for the principal components is constructed. The implicit parameter matrix is ​​then used to label the data points after dimensionality reduction in the corresponding coordinate system to obtain the coordinate points of each principal component. Following the preset principal component numerical connection order, the coordinate points of each principal component are connected sequentially to construct a closed graph and obtain the fault diagnosis map. The fault diagnosis diagram is matched with the corresponding template diagram for similarity. If the similarity value is greater than the preset threshold, it is determined that the texturing machine roller has a corresponding hidden fault.

[0013] Preferably, in step three, the specific process of calculating the severity assessment value and early warning value of the latent fault, outputting the early warning level, and issuing the corresponding early warning is as follows: For each factor affecting the fault, the principal components are divided into parameters with severe positive correlation and parameters with severe negative correlation. For each latent fault occurring in the texturing machine roller, the corresponding principal component values ​​are obtained and normalized, and then the formula is used: The severity assessment value YG was obtained; Where i is the label of the principal component parameter, and n is the total number of principal component parameters; Let i be the value of the severely positively correlated parameter. Preset weighting coefficients for the i-th severely positively correlated parameter. Let i be the value of the severely negatively correlated parameter. Preset weighting coefficients for the i-th severely negatively correlated parameter; The severity assessment values ​​corresponding to each latent fault were compiled and then used as a formula: The warning value YJ is obtained, where m is the label of the latent fault and M is the total number of latent faults. The preset weighting coefficient corresponds to the severity assessment value of the m-th latent fault. Set several warning value ranges, and pre-set each warning value range corresponds to a warning level; Match the warning value corresponding to the loading machine roller with all warning value ranges, output the corresponding warning level, and issue the corresponding warning based on the warning level.

[0014] Compared with the prior art, the beneficial effects of the present invention are: (1) The method and system for monitoring and early warning of the performance health of the texturing machine roller, by determining the minimum effective frequency upper limit of the latent parameter, constitutes the monitoring frequency band, so that the analog spectral entropy detector can accurately monitor the spectral entropy increment of the latent parameter signal within a specific frequency band and capture the occurrence of latent faults in a timely manner; after triggering the latent parameter acquisition operation, by dividing and processing the acquired latent parameter data frame signal, the signal is filtered by using the deletion threshold, invalid data below the threshold is removed, and valid data above the threshold is retained, thereby realizing efficient data compression and storage and improving the efficiency of data processing; so that the system can not only ensure the ability to capture the weak features of latent faults, but also avoid the surge in data volume.

[0015] (2) The texturing machine roller performance health monitoring and early warning method and system introduces a combined coordinate system, anomaly evaluation diagram, evaluation trajectory line and a variety of trajectory feature calculation methods to construct a comprehensive and detailed equipment health status evaluation system. This system can reflect the operating status of the texturing machine roller in multiple dimensions, obtain abnormal evaluation values ​​through comprehensive analysis, and thus accurately determine whether there are hidden faults. This greatly improves the accuracy and timeliness of fault diagnosis, helps to discover potential fault hazards as early as possible, prevents the occurrence and expansion of faults, and ensures the stable operation of equipment and the improvement of production efficiency.

[0016] (3) The method and system for monitoring and early warning of the performance health of texturing machine rollers, by performing dimensionality reduction processing on the latent parameter matrix and extracting key principal component data, not only reduces the amount of data processing, but also realizes intuitive display of equipment operating status changes and accurate identification of latent faults by constructing a multidimensional coordinate system of principal components and a fault diagnosis diagram; in addition, by calculating the severity assessment value and the early warning value, the severity of the fault can be quantitatively assessed from multiple dimensions, providing maintenance personnel with accurate decision-making basis, realizing timely and reasonable maintenance measures, reducing maintenance costs and equipment downtime, and effectively solving the problem of lack of effective assessment and early warning means for the severity of latent faults in the existing technology. Attached Figure Description

[0017] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1 Please see Figure 1This invention provides a method for monitoring and early warning of the performance health of texturing machine rollers, comprising the following steps: Step 1: Extract the latent parameters corresponding to the latent fault, analyze the lower effective frequency upper limit of the latent parameters, and construct the monitoring frequency band; calculate the spectral entropy increment corresponding to the latent parameter signal within the monitoring frequency band; based on this, determine whether to trigger the latent parameter acquisition and storage operation. The specific process is as follows: Extract hidden parameters corresponding to historical latent faults from the working log of the texturing machine roller; the latent parameters include: roller vibration acceleration, roller internal stress wave, roller drive device high-frequency current ripple, etc. For each latent parameter, offline excitation tests were conducted for each type of latent fault. The minimum effective frequency upper limit that can still maintain the recognition accuracy of ≥95% for the latent parameter signal corresponding to this type of latent fault was recorded. Latent faults include: microcracks, early pitting corrosion, surface micro-peeling, etc. Under the same latent parameter, the maximum value of the lowest effective frequency upper limit corresponding to various latent faults is taken, and 5kHz is used as a fixed lower limit to form the frequency band for calculating the spectral entropy increment corresponding to the latent parameter, which is denoted as the monitoring frequency band. A programmable Gm-C continuous-time filter is used to construct an analog spectral entropy detector, which calculates the spectral entropy increment corresponding to the implicit parameter signal within the monitoring frequency band in real time. A preset threshold for spectral entropy increment is set, and the spectral entropy increment at each time step is compared with the corresponding threshold. If the spectral entropy increment is greater than the corresponding threshold, the latent parameter acquisition operation is triggered, and the acquired latent parameters are stored in the memory. If the spectral entropy increment is greater than the corresponding threshold, it indicates that the signal complexity and uncertainty in the latent parameter signal have increased significantly. This change usually means that there may be abnormal features or latent fault features in the signal. Conversely, if the spectral entropy increment is less than or equal to the corresponding threshold, the spectral entropy increment corresponding to the monitoring frequency band will continue to be calculated in real time at the microwatt level. The specific process for triggering the implicit parameter acquisition operation is as follows: When the implicit parameter acquisition operation is triggered, the implicit parameter data frames of the trigger period and the preset buffer before and after it are acquired; the acquired implicit parameter data frames are divided into several subframes. For each subframe, the environmental noise standard deviation SC and signal entropy value XS of that data segment are obtained and normalized. Then, the pruning threshold TH is obtained using the formula: TH=SC×a1+XS×a2, where a1 and a2 are preset weighting coefficients. For each subframe, the latent parameter data is processed according to the corresponding deletion threshold TH. Latent parameter data below TH is removed, and latent parameter data above TH is stored in memory.

[0020] It should be noted that offline excitation tests are conducted for various latent faults under each latent parameter, and the corresponding minimum effective frequency upper limit is determined to form a monitoring frequency band. This enables the analog spectral entropy detector to accurately monitor the spectral entropy increment of the latent parameter signal within a specific frequency band and promptly detect the occurrence of latent faults. After triggering the latent parameter acquisition operation, the acquired latent parameter data frame signals are divided and processed. The signals are filtered using a reduction threshold to remove invalid data below the threshold and retain valid data above the threshold. This achieves efficient data compression and storage, improving data processing efficiency. This ensures that the system can capture the weak features of latent faults while avoiding a surge in data volume. When the spectral entropy increment is less than or equal to the threshold, the system maintains microwatt-level power consumption to continue real-time calculations, effectively reducing energy consumption and the waste of computing resources.

[0021] Step Two: Extract the stored implicit parameter data, construct a combined coordinate system and form an anomaly evaluation map, determine the center coordinates of each anomaly evaluation map, and construct the evaluation trajectory line; calculate the number of trajectory anomalies, the average trajectory deviation, the trajectory bending frequency, and the trajectory deviation value; and comprehensively analyze the anomaly evaluation values ​​to determine whether the texturing machine roller has a hidden fault. The specific process is as follows: Extract all implicit parameter data generated by the current texturing machine roller from the memory; use each implicit parameter as a different dimension to form a combined-dimensional coordinate system; For multiple latent parameter values ​​collected at the same time, determine their coordinate points in the combined coordinate system, connect the different latent parameters corresponding to the same time in the order of preset parameters to form line segments, connect the ends of these line segments to construct a closed figure, which is called the anomaly assessment figure, and mark the center coordinates of the anomaly assessment figure. Connect the center coordinates of the anomaly assessment map at each time point with line segments in chronological order to obtain the assessment trajectory line; The trajectory length is obtained by calculating the sum of the distances between two adjacent center trajectory points. The trajectory length change rate is obtained by dividing the trajectory length by the unit time. A preset trajectory length change rate threshold is set. If the trajectory length change rate is greater than the corresponding threshold, it is marked as an abnormal fluctuation trajectory. The deviation of the abnormal fluctuation trajectory from the corresponding threshold is calculated and recorded as the trajectory deviation degree. The number of occurrences of abnormal fluctuation trajectories is counted and the mean deviation of the trajectory corresponding to all abnormal fluctuation trajectories is calculated to obtain the number of trajectory anomalies GY and the mean deviation of the trajectory GP; where the number of trajectory anomalies reflects the frequency of abnormal fluctuations, and the mean deviation of the trajectory reflects the severity of abnormal fluctuations. For any three consecutive adjacent points At1, Bt2, and Ct3 on the central trajectory line, calculate the vector. and Using the formula: The angle between the trajectories is obtained. ; A preset trajectory angle threshold is set. If the trajectory angle is greater than the corresponding threshold, a trajectory bend will be marked at point Bt2. The number of trajectory bends in the evaluation trajectory line is counted and divided by the time length of the evaluation trajectory line to obtain the trajectory bend frequency GQ. The trajectory bend frequency provides information about the frequency of changes in the equipment's operating status, which helps to identify whether the equipment is subject to frequent interference or has intermittent faults. The evaluation trajectory line is compared with the corresponding preset standard evaluation trajectory line under the same working conditions. The standard deviation of the deviation distance between each point of the current evaluation trajectory line and the corresponding reference position of the preset standard evaluation trajectory line is calculated to obtain the trajectory deviation value GB. The trajectory deviation value provides the degree of deviation between the equipment's operating status and the standard status, which helps to identify whether the equipment has long-term health problems or systemic failures. After normalizing the number of trajectory anomalies GY, the average deviation of the trajectory GP, the trajectory bending frequency GQ, and the trajectory deviation value GB, the anomaly evaluation value YP is obtained using the formula: YP=GY×b1+GP×b2+GQ×b3+GB×b4, where b1, b2, b3, and b4 are preset weight coefficients. A preset threshold for abnormal evaluation values ​​is set. If the abnormal evaluation value is greater than the corresponding threshold, it is determined that the texturing machine roller has a hidden fault.

[0022] It should be noted that the process involves extracting the stored implicit parameter data, constructing a combined coordinate system and forming an anomaly evaluation map, determining the center coordinates of each anomaly evaluation map, and constructing the evaluation trajectory line; calculating the number of trajectory anomalies, the average trajectory deviation, the trajectory bending frequency, and the trajectory deviation value; and comprehensively analyzing the results to obtain the anomaly evaluation value, thereby determining whether the texturing machine roller has a hidden fault. This step enables multi-dimensional monitoring and analysis of the texturing machine roller's operating status. By constructing a combined coordinate system and evaluation trajectory lines, and combining the calculation of various trajectory features, the health status of the equipment can be comprehensively and meticulously assessed. By calculating abnormal evaluation values, it is possible to accurately determine whether there are hidden faults, thereby improving the accuracy and timeliness of fault diagnosis, effectively preventing the occurrence and expansion of faults, and ensuring the stable operation and production efficiency of the equipment.

[0023] Step 3: Extract the latent parameter data corresponding to the latent faults and construct the latent parameter matrix; analyze the principal component data of the latent faults; construct a multidimensional coordinate system of principal components to form a fault diagnosis map, and determine whether the corresponding latent fault has occurred; calculate the severity assessment value and warning value of the latent faults, output the warning level and issue corresponding warnings. The specific process is as follows: Extract latent parameter data corresponding to the latent fault of the texturing machine roller from the storage, organize the extracted data, and construct a latent parameter matrix; where the rows of the matrix represent different sampling times and the columns represent different latent parameters; For each type of latent fault, the latent parameter matrix is ​​analyzed using the principal component analysis (PCA) algorithm to obtain the corresponding principal component data; Using each principal component as a coordinate axis, a corresponding multidimensional coordinate system for the principal components is constructed. The implicit parameter matrix is ​​then used to label the data points after dimensionality reduction in the corresponding coordinate system to obtain the coordinate points of each principal component. Following the preset principal component numerical connection order, the coordinate points of each principal component are connected sequentially to construct a closed graph and obtain the fault diagnosis map. A hidden fault diagnosis template is preset. The fault diagnosis diagram is matched with the corresponding template diagram. If the similarity value is greater than the preset threshold, it is determined that the texturing machine roller has a corresponding hidden fault. For each influencing fault, a correlation analysis-based method is used to analyze the relationship between each principal component and the severity of the corresponding latent fault; based on the positive or negative correlation, each principal component is divided into severely positively correlated parameters and severely negatively correlated parameters. For each latent fault occurring in the texturing machine roller, the corresponding principal component values ​​are obtained and normalized, and then the formula is used: The severity assessment value YG was obtained; Where i is the label of the principal component parameter, and n is the total number of principal component parameters; Let i be the value of the severely positively correlated parameter. Preset weighting coefficients for the i-th severely positively correlated parameter. Let i be the value of the severely negatively correlated parameter. Preset weighting coefficients for the i-th severely negatively correlated parameter; The severity assessment values ​​corresponding to each latent fault were compiled and then used as a formula: The warning value YJ is obtained, where m is the label of the latent fault and M is the total number of latent faults. The preset weighting coefficient corresponds to the severity assessment value of the m-th latent fault. Several warning value ranges are set, and each warning value range corresponds to a warning level. By matching the warning value corresponding to the loading machine roller with all warning value ranges, the corresponding warning level is output, and a corresponding warning is issued based on the corresponding warning level.

[0024] It should be noted that by constructing a latent parameter matrix and using the PCA algorithm to analyze and extract principal component data, the data dimensionality is reduced while retaining key information, effectively reducing the amount of data processing and improving the efficiency of fault diagnosis. Construct a principal component multidimensional coordinate system and fault diagnosis diagram to intuitively display changes in equipment operating status. Combine this with a preset template diagram for similarity matching to achieve accurate identification and judgment of hidden faults. Correlation analysis was used to classify the types of principal component parameters, and they were processed according to positive and negative correlations. Normalization and weighted summation were combined to calculate the severity assessment value and warning value, and the severity of the fault was quantitatively assessed from multiple dimensions, providing an accurate basis for taking targeted measures. Set warning value ranges and corresponding levels, quickly match and output warning levels, promptly remind maintenance personnel to take appropriate measures, effectively prevent the fault from developing further, and ensure reliable equipment operation.

[0025] A performance health monitoring and early warning system for texturing machine rollers includes: Data monitoring and acquisition module: Extracts latent parameters corresponding to latent faults, analyzes the minimum effective frequency upper limit of latent parameters, and constitutes a monitoring frequency band; calculates the spectral entropy increment corresponding to the latent parameter signal within the monitoring frequency band; and determines whether to trigger the latent parameter acquisition and storage operation based on this. Health assessment module: Extracts stored implicit parameter data, constructs a combined coordinate system and forms an anomaly assessment map, determines the center coordinates of each anomaly assessment map, and constructs assessment trajectory lines; calculates the number of trajectory anomalies, trajectory average deviation, trajectory bending frequency, and trajectory deviation value; and comprehensively analyzes to obtain anomaly assessment values, thereby determining whether the texturing machine roller has a hidden fault. Fault Diagnosis and Early Warning Module: Extracts latent parameter data corresponding to latent faults and constructs a latent parameter matrix; analyzes the principal component data of latent faults; constructs a multidimensional coordinate system of principal components to form a fault diagnosis map, and determines whether the corresponding latent fault has occurred; calculates the severity assessment value and early warning value of latent faults, outputs the early warning level and issues corresponding warnings.

[0026] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring and early warning of the performance health of texturing machine rollers, characterized in that: Includes the following steps: Step 1: Extract the latent parameters corresponding to the latent faults, analyze the minimum effective frequency upper limit of the latent parameters, and construct the monitoring frequency band; Calculate the spectral entropy increment corresponding to the latent parameter signal within the monitoring frequency band; determine whether to trigger the latent parameter acquisition and storage operation based on this. Step 2: Extract the stored implicit parameter data, construct a combined coordinate system and form an anomaly evaluation map, determine the center coordinates of each anomaly evaluation map, and construct the evaluation trajectory line; Calculate the number of trajectory anomalies, the average trajectory deviation, the trajectory curvature frequency, and the trajectory offset value; The abnormal evaluation values ​​were obtained through comprehensive analysis, and it was determined whether there was a hidden fault in the texturing machine roller. Step 3: Extract the latent parameter data corresponding to the latent faults and construct the latent parameter matrix; Analyze the principal component data of latent faults; Construct a multidimensional coordinate system for principal components to form a fault diagnosis map, and determine whether a corresponding hidden fault has occurred. Calculate the severity assessment value and early warning value of latent faults, output the early warning level and issue corresponding warnings.

2. The method for monitoring and early warning of the performance health of a texturing machine roller according to claim 1, characterized in that: In step one, the specific process of constructing the monitoring frequency band is as follows: Extract the latent parameters corresponding to historical latent faults; for each latent parameter, record the minimum effective frequency upper limit of the latent parameter signal corresponding to that type of latent fault; Under the same latent parameter, the maximum value of the lowest effective frequency upper limit corresponding to various latent faults is taken, and 5kHz is used as a fixed lower limit to form the frequency band for calculating the spectral entropy increment corresponding to the latent parameter, which is denoted as the monitoring frequency band.

3. The method for monitoring and early warning of the performance health of a texturing machine roller according to claim 2, characterized in that: In step one, the specific process for determining whether the implicit parameter acquisition and storage operation has been triggered is as follows: Construct a simulated spectral entropy detector to calculate the spectral entropy increment corresponding to the implicit parameter signal within the monitoring frequency band in real time; The spectral entropy increment at each time point is compared with the corresponding threshold. If the spectral entropy increment is greater than the corresponding threshold, the implicit parameter acquisition operation is triggered, and the acquired implicit parameters are stored in the memory. Otherwise, if the spectral entropy increment is less than or equal to the corresponding threshold, the spectral entropy increment corresponding to the monitoring frequency band is calculated in real time.

4. The method for monitoring and early warning of the performance health of a texturing machine roller according to claim 3, characterized in that: The specific process for triggering the implicit parameter acquisition operation is as follows: When the implicit parameter acquisition operation is triggered, the implicit parameter data frames of the trigger period and the preset buffer before and after it are acquired; the acquired implicit parameter data frames are divided into several subframes. For each subframe, the environmental noise standard deviation SC and signal entropy value XS of that data segment are obtained and normalized. Then, the pruning threshold TH is obtained using the formula: TH=SC×a1+XS×a2, where a1 and a2 are preset weighting coefficients. For each subframe, implicit parameter data below the deletion threshold is removed, while implicit parameter data above the deletion threshold is stored in memory.

5. The method for monitoring and early warning of the performance health of a texturing machine roller according to claim 4, characterized in that: In step two, the specific process of constructing the evaluation trajectory line is as follows: Extract the implicit parameter data stored in the memory; use each implicit parameter as a different dimension to form a combined-dimensional coordinate system; For multiple latent parameter values ​​collected at the same time, determine their coordinate points in the combined coordinate system, connect the different latent parameters corresponding to the same time in the order of preset parameters to form line segments, connect the ends of these line segments to construct a closed figure, which is called the anomaly assessment figure, and mark the center coordinates of the anomaly assessment figure. Connect the center coordinates of the anomaly assessment graph at each time point with line segments in chronological order to obtain the assessment trajectory line.

6. The method for monitoring and early warning of the performance health of a texturing machine roller according to claim 5, characterized in that: The specific process for calculating the number of trajectory anomalies and the average trajectory bias is as follows: The trajectory length is obtained by calculating the sum of the distances between two adjacent center trajectory points. The trajectory length is then divided by the unit time to obtain the rate of change of the trajectory length. The trajectory length change rate is compared with the corresponding threshold to determine whether it is an abnormal fluctuation trajectory, and the deviation of the abnormal fluctuation trajectory from the corresponding threshold is calculated and recorded as the trajectory deviation degree. The number of times abnormal fluctuation trajectories occur and the mean deviation of the trajectory corresponding to all abnormal fluctuation trajectories are calculated to obtain the number of trajectory anomalies GY and the mean deviation of the trajectory GP.

7. The method for monitoring and early warning of the performance health of a texturing machine roller according to claim 6, characterized in that: In step two, the specific process of analyzing abnormal evaluation values ​​and determining whether the texturing machine roller has a hidden fault is as follows: For any three consecutive adjacent points At1, Bt2, and Ct3 on the central trajectory line, calculate the trajectory angle at point Bt2; If the angle between the trajectory lines is greater than the corresponding threshold, a curve in the trajectory will be marked at point Bt2. The trajectory bending frequency GQ is obtained by counting the number of bends in the trajectory line and dividing by the time length of the trajectory line. Calculate the standard deviation of the deviation distance between each point on the current evaluation trajectory line and the corresponding benchmark position of the preset standard evaluation trajectory line to obtain the trajectory deviation value GB; After normalizing the number of trajectory anomalies GY, the average deviation of the trajectory GP, the trajectory bending frequency GQ, and the trajectory deviation value GB, the anomaly evaluation value YP is obtained using the formula: YP=GY×b1+GP×b2+GQ×b3+GB×b4, where b1, b2, b3, and b4 are preset weight coefficients. If the abnormal evaluation value is greater than the corresponding threshold, it is determined that the texturing machine roller has a hidden fault.

8. The method for monitoring and early warning of the performance health of a texturing machine roller according to claim 7, characterized in that: In step three, the specific process for determining whether a corresponding latent fault has occurred is as follows: Extract the latent parameter data corresponding to the latent faults in the texturing machine rollers and construct a latent parameter matrix; For each type of latent fault, the latent parameter matrix is ​​analyzed to obtain the corresponding principal component data; Using each principal component as a coordinate axis, a corresponding multidimensional coordinate system for the principal components is constructed. The implicit parameter matrix is ​​then used to label the data points after dimensionality reduction in the corresponding coordinate system to obtain the coordinate points of each principal component. Following the preset principal component numerical connection order, the coordinate points of each principal component are connected sequentially to construct a closed graph and obtain the fault diagnosis map. The fault diagnosis diagram is matched with the corresponding template diagram for similarity. If the similarity value is greater than the preset threshold, it is determined that the texturing machine roller has a corresponding hidden fault.

9. The method for monitoring and early warning of the performance health of a texturing machine roller according to claim 8, characterized in that: In step three, the specific process of calculating the severity assessment value and early warning value of the latent fault, outputting the early warning level, and issuing the corresponding early warning is as follows: For each factor affecting the fault, the principal components are divided into parameters with severe positive correlation and parameters with severe negative correlation. For each latent fault occurring in the texturing machine roller, the corresponding principal component values ​​are obtained and normalized, and then the formula is used: The severity assessment value YG was obtained; Where i is the label of the principal component parameter, and n is the total number of principal component parameters; Let i be the value of the severely positively correlated parameter. Preset weighting coefficients for the i-th severely positively correlated parameter. Let i be the value of the severely negatively correlated parameter. Preset weighting coefficients for the i-th severely negatively correlated parameter; The severity assessment values ​​corresponding to each latent fault were compiled and then used as a formula: The warning value YJ is obtained, where m is the label of the latent fault and M is the total number of latent faults. The preset weighting coefficient corresponds to the severity assessment value of the m-th latent fault. Set several warning value ranges, and pre-set each warning value range corresponds to a warning level; Match the warning value corresponding to the loading machine roller with all warning value ranges, output the corresponding warning level, and issue the corresponding warning based on the warning level.

10. A texturing machine roller performance health monitoring and early warning system, applied to the texturing machine roller performance health monitoring and early warning method proposed in any one of claims 1-9, characterized in that: Data monitoring and acquisition module: Extracts latent parameters corresponding to latent faults, analyzes the minimum effective frequency upper limit of latent parameters, and constitutes a monitoring frequency band; calculates the spectral entropy increment corresponding to the latent parameter signal within the monitoring frequency band; and determines whether to trigger the latent parameter acquisition and storage operation based on this. Health assessment module: Extract stored implicit parameter data, construct a combined coordinate system and form anomaly assessment map, determine the center coordinates of each anomaly assessment map, and construct the assessment trajectory line; Calculate the number of trajectory anomalies, the average trajectory deviation, the trajectory curvature frequency, and the trajectory offset value; The abnormal evaluation values ​​were obtained through comprehensive analysis, and it was determined whether there was a hidden fault in the texturing machine roller. Fault diagnosis and early warning module: Extract the latent parameter data corresponding to latent faults and construct a latent parameter matrix; Analyze the principal component data of latent faults; Construct a multidimensional coordinate system for principal components to form a fault diagnosis map, and determine whether a corresponding hidden fault has occurred. Calculate the severity assessment value and early warning value of latent faults, output the early warning level and issue corresponding warnings.

Images