Intelligent yarn winding system based on intelligent control

Abstract

The invention discloses an intelligent yarn winding system based on intelligent control, and relates to the technical field of yarn winding, a monitoring acquisition module is arranged to acquire real-time image acquisition data and monitoring acquisition data of target yarn, and an identification control module is arranged to identify yarn broken ends and correspondingly adjust a plurality of operation parameters. The setting analysis optimization unit analyzes a representation data set of a plurality of reference parameters capable of representing the deep cause every time when a manager analyzes the deep cause, and analyzes an optimization reference table of the plurality of reference parameters with linear strong correlation in the representation data set; the backtracking determination unit can correspondingly carry out further interception in continuous monitoring values in a period of time before each broken end based on the corresponding optimization reference table, directly intercepts data which are related to deep reasons and have strong linear correlation, and avoids waste of time and labor cost caused by manual analysis of excessive redundant information.

Description

Technical Field

[0001] The present invention relates to the technical field of yarn winding, and in particular to an intelligent yarn winding system based on intelligent control. Background Art

[0002] Yarn winding is a key process in spinning production. Its operational stability and winding quality have a decisive impact on subsequent weaving, dyeing and other process flows as well as the quality of the final product. In recent years, with the advancement of industrial intelligence, the intelligent yarn winding system based on intelligent control has gradually become the core equipment for improving production efficiency and product quality. By integrating sensors, data processing units and actuators, it realizes real-time monitoring and dynamic control of key parameters such as yarn tension, speed, and forming state during the winding process. In spinning production, yarn breakage is one of the main problems affecting winding efficiency and quality. In the existing technology, when a yarn breakage is detected during operation of a yarn winding system based on intelligent control, a parameter backtracking mechanism is usually triggered: that is, key operating parameters (such as yarn tension, winding speed, ambient temperature and humidity, etc.) in the period before the breakage occurs are retrieved, and these parameters are compared with preset normal thresholds or historical averages to identify the surface causes of the breakage (such as excessive instantaneous tension, winding speed fluctuations, etc.). The system then adjusts the relevant parameters in real time according to the identified surface causes to reduce the occurrence of subsequent breakages. However, this approach only scratches the surface of yarn breakage analysis. Yarn breaks are often the result of the long-term cumulative effects of multiple factors. Some underlying causes (such as aging of key equipment components, gradual changes in raw material properties, and hidden mechanical failures) are not directly reflected by parameter fluctuations before a single break, but are instead indirectly reflected through long-term trend changes in characteristics such as break frequency and location. It is worth noting that some of these underlying causes are characterized by strong correlations in the monitored values ​​of certain baseline parameters over time. This type of data has clear patterns and distinct characteristics, allowing technicians to quickly identify them manually without relying on intelligent identification methods. Even so, current analysis of these underlying causes still relies on manual retrieval of historical data within a fixed lookback period. Because fixed lookback periods are primarily designed to ensure data integrity and accuracy, they often cover large amounts of data over a long period of time. This means that manual analysis still requires filtering out highly relevant, accurate parameter data from redundant information. This not only increases the complexity of data screening and correlation, but also significantly reduces the efficiency of analyzing underlying causes, extending the time from when a break occurs to locating the root cause. In order to solve the above problems, the present invention proposes a solution. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent yarn winding system based on intelligent control, in order to solve the problems raised in the above background technology.

[0004] The present invention provides an intelligent yarn winding system based on intelligent control, comprising: Monitoring and acquisition module, used to acquire real-time image acquisition data and monitoring acquisition data of target yarn; An identification control module is configured to perform end-break identification on the target yarn after receiving image acquisition data of the target yarn at each moment. If the identification determines that the yarn has broken at the said moment, the numerical values ​​of several operating parameters are adjusted according to preset comparison and control steps. During the control process, an end-break log information of the said moment is generated. The end-break log information includes several reference parameters and their monitoring values. A backtracking analysis and optimization module is used to determine and backtrack the amount of stored broken log information and optimize it. The backtracking analysis and optimization module includes a backtracking determination unit and an analysis and optimization unit; Whenever a broken log message at a certain moment is received, the backtracking determination unit uses the moment as the backtracking starting point, backtracks forward by Z1 time, counts the total number of broken log messages received within the backtracked Z1 time, and selects and generates, based on the statistical result, if the total number is equal to P1, analysis benchmark information of an analysis cycle is generated according to a preset first backtracking determination rule, and deep characterization data of the analysis cycle is determined according to the analysis benchmark information, where the deep characterization data includes several characterization data sets including deep causes and several benchmark parameters; For any deep cause, when the amount of deep characterization data containing the deep cause stored in the analysis and optimization unit reaches a preset fixed amount, all deep characterization data containing the deep cause are analyzed to obtain an optimized reference table of several benchmark parameters of the deep cause and transmit the optimized reference table to the retrospective determination unit; The backtracking determination unit receives and stores the broken log information at each moment with several benchmark parameters. The backtracking determination unit uses the moment as the backtracking starting point, backtracks forward Z1 time, and counts the total number of broken log information received within the backtracked Z1 time. If the total number is equal to P1, the analysis benchmark information of an analysis cycle is generated according to the preset second backtracking determination rule, and the deep characterization data of the analysis cycle is determined according to the analysis benchmark information.

[0005] Furthermore, the first backtracking determination rule is as follows: If the total number is equal to P1, then for each moment corresponding to the broken log information received within the backtracked Z1 time, backtrack forward to Z2 time, obtain all monitoring values ​​of each benchmark parameter in the broken log information within the backtracked Z2 time and generate the backtracking analysis data of the moment based on them, Z1 and Z2 are the preset first standard backtracking time and second standard backtracking time respectively; The analysis benchmark information of an analysis cycle is generated based on the backtracking analysis data at the moment corresponding to each broken log information received within the backtracking Z1 time.

[0006] Furthermore, the steps for determining the deep characterization data of the analysis period according to the analysis benchmark information are as follows: The analysis benchmark data of the analysis period is displayed to the management personnel for review. The management personnel determines the deep reason why the number of breakages in the analysis period reaches P1 times by reviewing the changes in the monitoring values ​​of each benchmark parameter in the retrospective analysis data table at each moment. During the analysis process, the management personnel will extract some data that can determine the deep reason from the analysis benchmark data to obtain the deep characterization data of the analysis period.

[0007] Furthermore, among the several characterization data sets of several benchmark parameters contained in a deep characterization data, one characterization data set corresponds to one benchmark parameter, one benchmark parameter corresponds to several characterization data sets, and one characterization data set contains continuous monitoring values ​​of the corresponding benchmark parameter at several acquisition moments.

[0008] Furthermore, the analysis steps for obtaining an optimized reference table of several benchmark parameters of the underlying cause are as follows: S11: The deep characterization data containing the deep cause stored in the analysis and optimization unit are marked as A1, A2, ..., Aa, a ≥ 1, the deep characterization data A1, A2, ..., Aa are traversed, all the benchmark parameters are extracted and deduplicated, and all the remaining benchmark parameters after deduplication are marked as B1, B2, ..., Bb, b ≥ 1; S12: Extract all representation data sets corresponding to the benchmark parameters B1 from the deep standard data A1, and mark them as C1, C2, ..., Cc, where c≥1; S13: creating a linear record variable H1 of the baseline parameter B1 relative to the underlying cause, wherein the initial value of the linear record variable H1 is 0; Determine whether the detection value and the acquisition time have a strong linear correlation characteristic in combination with the characterization data set C1. If it is determined that the linear correlation characteristic is strong, calculate and obtain the first characterization value, the second characterization value, and the characterization duration of the characterization data set C1. Otherwise, no processing is performed. S14: According to S13, the detection values ​​and the acquisition time are sequentially combined with the characterization data sets C2, C3, ..., Cc to determine whether they have a strong linear correlation characteristic. After the determination is completed, the value of the linear record variable H1 at this time is obtained. If the value of H1 is greater than or equal to H, it is determined that the benchmark parameter B1 is qualified for optimization, and an optimization data table for the benchmark parameter B1 is created, where H1 is the critical threshold for strong linear correlation determination; if the value of H1 is less than H, it is determined that the benchmark parameter B1 is not qualified for optimization; S15: If the optimized data table of the benchmark parameter B1 is created in S14, all characterization data sets corresponding to the benchmark parameter B1 are extracted from the benchmark parameters B2, B3, ..., Bb in sequence according to S11 to S14 and several optimized data tables of the benchmark parameter B1 are obtained according to S13 to S14; Merge the data of each row of all optimization data tables of the benchmark parameter B1 into one optimization data table to obtain an optimization reference table of the benchmark parameter B1; S16: determining whether the benchmark parameters B2, B3, ..., Bb have optimization qualifications in sequence according to S11 to S15, and obtaining an optimization reference table of all benchmark parameters that have optimization qualifications based on the determination results; The analysis and optimization unit optimizes a reference table of several benchmark parameters of the deep-seated cause.

[0009] Furthermore, in S13, the determination content is as follows: SS21: Extract all monitoring values ​​D1, D2, ..., Dd from the characterization data set C1 in the order of collection time, where d ≥ 1. The collection time of the monitoring values ​​D1, D2, ..., Dd is marked as T1, T2, ..., Td respectively. SS22: Calculate the linear correlation coefficient E1 between the collection time and the monitoring value based on the monitoring values ​​D1, D2, ..., Dd and the collection time T1, T2, ..., Td. The linear correlation coefficient E1 is used to measure the correlation between the collection time and the monitoring value. The linear correlation coefficient E1 has positive and negative values, ranging from -1 to 1. A positive value indicates positive correlation, and a negative value indicates negative correlation. The closer the absolute value is to 1, the stronger the linear relationship. SS23: Compare |E1| and E. If |E1| ≥ E, it is determined that the monitoring value and the acquisition time in the characterization data set C1 have a strong linear correlation characteristic; At this time, the acquisition times T1, T2, ..., Td are assigned to x in sequence, and the monitoring values ​​D1, D2, ..., Dd are assigned to y accordingly. Substituting them into the linear function y = kx + b, we get d groups of linear equations with one variable. Solving any two of these linear equations together, we get d(d-1) / 2 groups of linear equations with two variables. Solving these linear equations with two variables, we get d(d-1) / 2 values ​​of k and b. Use a discrete point filtering algorithm to process the obtained d(d-1) / 2 k values, and calculate the average value of all k values ​​remaining after data processing, and calibrate the average value as the first characterization quantity of the characterization data set C1. Similarly, obtain the second characterization quantity of the characterization data set C1; use d as the characterization duration of the characterization data set C1.

[0010] Furthermore, the second retrospective determination rule for generating the analysis benchmark information of an analysis cycle is as follows: If the total number is equal to P1, then for each moment corresponding to the broken log information received within the backtracked time T1, backtrack forward to time T2, and for any benchmark parameter in the broken log information within time T2, if the backtracking determination unit does not store an optimization reference table for the benchmark parameter, then obtain all monitoring values ​​of the benchmark parameter within the backtracked time T2 and use them as the backtracking data of the benchmark parameter at the moment; If an optimization reference table of the benchmark parameter is stored, all monitored values ​​of the benchmark parameter within the retrospective time T2 are intercepted to obtain a number of retrospective data of the benchmark parameter at the time; Generate the retrospective analysis data of the moment according to the retrospective data of all the benchmark parameters in the obtained broken log information; generate the analysis benchmark information of an analysis cycle according to the retrospective analysis data of the moment corresponding to each broken log information received within the retrospective Z1 time.

[0011] Compared with the existing technology, it has the following beneficial effects: The present invention provides a monitoring and acquisition module to collect real-time image acquisition data and monitoring acquisition data of the target yarn, and provides an identification control module to identify yarn breakage based on the image acquisition data. Based on the monitoring values ​​of various parameters in the monitoring and acquisition data at the time of yarn breakage, a number of operating parameters are adjusted accordingly. In this way, the real-time control accuracy of the yarn winding process can be improved, and the impact of yarn breakage on production can be reduced. The present invention sets a retrospective determination unit. When the number of head breaks reaches a preset amount within a fixed time, the continuous monitoring values ​​of several benchmark parameters in the period before each head break are retrospectively analyzed and provided to management personnel for analysis to determine the underlying cause. This avoids the disadvantage of manually blindly screening information from massive historical data and reduces the interference of redundant data on analysis. The present invention sets up an analysis and optimization unit to analyze a characterization data set of several benchmark parameters that can characterize the deep-seated cause each time the management personnel analyze the deep-seated cause, and analyzes an optimization reference table for several benchmark parameters with strong linear correlation. The backtracking determination unit can further intercept the continuous monitoring values ​​for a period of time before each break based on the corresponding optimization reference table. In this way, for data with strong correlation, the backtracking determination unit can directly intercept data related to the deep-seated cause and with strong linear correlation from the massive data of a fixed backtracking period, avoiding the waste of time and labor costs caused by manual analysis of too much redundant information, and greatly shortening the time from data acquisition to locating the deep-seated cause. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a system block diagram of the present invention. DETAILED DESCRIPTION

[0013] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0014] See also Figure 1 , this application provides an intelligent yarn winding system based on intelligent control, including a monitoring and acquisition module, an identification and control module, and a backtracking analysis and optimization module; The monitoring and acquisition module is used to monitor the winding process of the target yarn in real time and collect data. The monitoring and acquisition module includes a first acquisition unit and a second acquisition unit; The first acquisition unit acquires images of the winding process of the target yarn in real time to obtain real-time image acquisition data of the target yarn, and transmits the image acquisition data to the recognition control module; The second collection unit collects the monitoring values ​​of several monitoring parameters during the winding process of the target yarn in real time to obtain real-time monitoring data of the target yarn, wherein the monitoring parameters are parameters selected by the management personnel that can directly or indirectly affect the end breakage; In this application, the winding process of the target yarn depends on the yarn winding equipment; In this application, the monitored parameters include but are not limited to yarn tension, yarn linear speed, yarn diameter, yarn particle size, yarn hairiness, winding roller spindle speed, yarn guide position, moving speed, winding diameter, equipment vibration / noise, motor current, power, ambient temperature and ambient humidity; The monitoring parameters can be collected in real time by sensors, which can serve as the basic data for subsequent analysis of breakage causes and trigger parameter adjustments. The sensors include but are not limited to tension sensors, laser diameter gauges, encoders, temperature and humidity sensors, vibration sensors, etc. The monitoring and acquisition module transmits the real-time monitoring and acquisition data of the target yarn to the identification and control module; The identification control module is used to identify and control ends broken during the winding process of the target yarn. The identification control module pre-stores an end-break identification threshold and comparison control information of several reference parameters. The comparison control information includes an over-limit value and correction values ​​of several operating parameters. The reference parameters are several monitoring parameters selected by the management personnel from all monitoring parameters and can be used to directly determine whether an end break has occurred through numerical comparison; The function of the over-limit value is to compare the values ​​to determine whether the end break has occurred. It is set by the management personnel. The operating parameters refer to the equipment control parameters that can be dynamically adjusted during the winding process to maintain the stability of the winding process. When an end break is detected, it is necessary to adjust these operating parameters to eliminate the cause and prevent the end break again. The correction value is a preset parameter adjustment target value used to correct the cause of the end break. The preset basis can be that the management personnel summarize the mapping relationship between "parameter abnormality-end breakage type" through long-term production data and preset correction plans, for example: If the monitoring parameter: yarn tension>18cN triggers the end break, the operating parameter: the tension setting value is corrected to 12cN; After receiving the image acquisition data and monitoring acquisition data of the target yarn at each moment, the image acquisition data is firstly subjected to end breakage recognition. If it is determined that an end breakage occurs at the said moment, the values ​​of several operating parameters are adjusted according to the preset comparison and control steps. During the control process, the end breakage log information of the said moment is generated. The end breakage log information includes several reference parameters and their monitoring values. The comparison and control steps are as follows: Extracting the monitored values ​​of all the benchmark parameters at the time from the monitored collected data; comparing the monitored value with the over-limit value of any one of the extracted benchmark parameters; if the monitored value is greater than or equal to the over-limit value, extracting all the operating parameters and their correction values ​​from the comparison control information of the benchmark parameters; adjusting the value of each extracted operating parameter to the corresponding correction value; otherwise, performing no processing; After the comparison between the monitored values ​​of all extracted benchmark parameters and the corresponding over-limit values ​​is completed, all recorded benchmark parameters and their monitored values ​​are obtained and the broken log information at the time is generated based on them, and the broken log information is transmitted to the retrospective analysis and optimization module for temporary storage; If the identification determines that no end breakage occurs at the time, no processing is performed; During the recognition process, the target yarn profile in the image acquisition data is first extracted. If there is a broken area in the target yarn profile that exceeds the preset breakage recognition threshold, it is determined that the breakage occurs at that moment. Otherwise, it is determined that no breakage occurs at that moment. In this application, the preset breakage recognition threshold is 5, which means that 5 consecutive pixels are missing. Transmitting the broken log information to the backtracking analysis optimization module for temporary storage; A backtracking analysis and optimization module is used to periodically determine and backtrack the number of stored broken log information and optimize it. The backtracking analysis and optimization module includes a backtracking determination unit and an analysis and optimization unit; Every time a broken log message is received at a certain moment, the backtracking determination unit takes the moment as the backtracking starting point, backtracks forward by Z1 time, and counts the total number of broken log messages received within the backtracked Z1 time. If the total number is equal to P1, then for each moment corresponding to a broken log message received within the backtracked Z1 time, backtracks forward by Z2 time, obtains all monitoring values ​​of each benchmark parameter in the broken log message within the backtracked Z2 time within the Z2 time, and generates the backtracking analysis data of the moment based on the value, Z1 and Z2 are the preset first standard backtracking time and second standard backtracking time respectively, and P1 is the preset standard analysis quantity threshold; Generate analysis benchmark information for an analysis period based on the generated backtracking analysis data corresponding to each broken log information received within the backtracking Z1 time, and determine deep representation data for the analysis period based on the analysis benchmark information; The following has been determined: The analysis benchmark data of the analysis period is displayed to the management personnel for review. The management personnel determines the deep-seated cause of the number of breakages in the analysis period reaching P1 times by reviewing the changes in the monitored values ​​of each benchmark parameter in the retrospective analysis data table at each moment. During the analysis process, the management personnel extracts a number of data that can determine the deep-seated cause from the analysis benchmark data to obtain deep-seated characterization data of the analysis period, and transmits the deep-seated characterization data to the analysis optimization unit for storage; The deep characterization data includes several characterization data sets of several benchmark parameters, one characterization data set corresponds to one benchmark parameter, one benchmark parameter corresponds to several characterization data sets, and one characterization data set includes several continuous monitoring values ​​of the corresponding benchmark parameter at the acquisition time; For any deep cause, when the amount of deep characterization data containing the deep cause stored in the analysis and optimization unit reaches a preset fixed amount, all deep characterization data containing the deep cause are analyzed, where the fixed amount is preset by the management personnel. The preset basis is sufficient to support effective analysis and ensure that the analysis process does not rely on fragmented information, but is based on sufficient and complete characterization data, thereby improving the depth of analysis of the deep cause and the reliability of the conclusion, providing an accurate basis for subsequent optimization; The analysis steps are as follows: S11: The deep characterization data containing the deep cause stored in the analysis and optimization unit are marked as A1, A2, ..., Aa, a ≥ 1, the deep characterization data A1, A2, ..., Aa are traversed, all the benchmark parameters are extracted and deduplicated, and all the remaining benchmark parameters after deduplication are marked as B1, B2, ..., Bb, b ≥ 1; S12: Extract all representation data sets corresponding to the benchmark parameters B1 from the deep standard data A1, and mark them as C1, C2, ..., Cc, where c≥1; S13: creating a linear record variable H1 of the baseline parameter B1 relative to the underlying cause, wherein the initial value of the linear record variable H1 is 0; Combined with the characterization data set C1, the detection value and the acquisition time are judged to determine whether they have a strong linear correlation characteristic. If it is determined that they have a strong linear correlation characteristic, the first characterization value, the second characterization value and the characterization time length in the characterization data set C1 are calculated and obtained. The judgment content is as follows: SS21: Extract all monitoring values ​​D1, D2, ..., Dd from the characterization data set C1 in the order of collection time, where d ≥ 1. The collection time of the monitoring values ​​D1, D2, ..., Dd is marked as T1, T2, ..., Td respectively. SS22: Calculate the linear correlation coefficient E1 between the collection time and the monitoring value based on the monitoring values ​​D1, D2, ..., Dd and the collection time T1, T2, ..., Td. The linear correlation coefficient E1 is used to measure the correlation between the collection time and the monitoring value. The linear correlation coefficient E1 has positive and negative values, ranging from -1 to 1. A positive value indicates positive correlation, and a negative value indicates negative correlation. The closer the absolute value is to 1, the stronger the linear relationship. SS23: Compare |E1| and E. If |E1| ≥ E, it is determined that the monitoring value and the acquisition time in the characterization data set C1 have a linear strong correlation characteristic. E is a preset strong correlation critical threshold, preferably 0.7. At this time, the acquisition times T1, T2, ..., Td are assigned to x in sequence, and the monitoring values ​​D1, D2, ..., Dd are assigned to y accordingly. Substituting them into the linear function y = kx + b, we get d groups of linear equations with one variable. Solving any two of these linear equations together, we get d(d-1) / 2 groups of linear equations with two variables. Solving these linear equations with two variables, we get d(d-1) / 2 values ​​of k and b. Using a discrete point filtering algorithm to process the obtained d(d-1) / 2 k values, and calculating the average value of all k values ​​remaining after the data processing, the average value is calibrated as the first characterization quantity of the characterization data set C1. Similarly, using a discrete point filtering algorithm to process the obtained d(d-1) / 2 b values, and calculating the average value of all b values ​​remaining after the data processing, the average value is calibrated as the second characterization quantity of the characterization data set C1. In this application, the discrete point filtering algorithm can be any one of the Z-score filtering algorithm, the IQR filtering algorithm, and the density filtering algorithm; Using d as the characterization duration of the characterization dataset C1; S14: According to S13, the characterization data sets C2, C3, ..., Cc are sequentially combined to determine whether the detection value and the collection time have a strong linear correlation characteristic. After the determination is completed, the value of the linear record variable H1 at this time is obtained. If the value of H1 is greater than or equal to H, it is determined that the benchmark parameter B1 has optimization qualifications, and an optimization data table for the benchmark parameter B1 is created. The optimization data table contains three fields: a first characterization quantity, a second characterization quantity, and a characterization duration. After the characterization data set Cc is combined to determine whether the detection value and the collection time have a strong linear correlation characteristic, all the first characterization quantities, second characterization quantities, and characterization durations obtained are sequentially filled into the optimization data table, wherein H1 is a critical threshold for determining linear strong correlation, and H is a preset standard quantity for determining optimization qualifications; It should be noted here that the first characterization quantity, the second characterization quantity, and the characterization duration calculated based on the same characterization data set correspond to each other, and are represented as data in the same row in the optimization data table; If the value of H1 is less than H, the benchmark parameter B1 is judged to be unqualified for optimization; S15: If the optimized data table of the benchmark parameter B1 is created in S14, all characterization data sets corresponding to the benchmark parameter B1 are extracted from the benchmark parameters B2, B3, ..., Bb in sequence according to S11 to S14 and several optimized data tables of the benchmark parameter B1 are obtained according to S13 to S14; Merge the data of each row of all optimized data tables of benchmark parameter B1 into one optimized data table to obtain the optimized reference table of benchmark parameter B1. It should be noted that if there are several rows of data that are identical during the merging process, only one row of data will be merged into the optimized reference table. Some rows of data are identical, indicating that the first characterization quantity, the second characterization quantity, and the characterization duration are all the same; S16: determining whether the benchmark parameters B2, B3, ..., Bb have optimization qualifications in sequence according to S11 to S15, and obtaining an optimization reference table of all benchmark parameters that have optimization qualifications based on the determination results; The analysis and optimization unit transmits the optimization reference table of several benchmark parameters of the deep-seated cause to the retrospective determination unit for storage; After storing the optimization reference tables of several benchmark parameters of several deep-seated causes, each time a broken log message at a moment is received, the backtracking determination unit uses the moment as the backtracking starting point, backtracks forward to T1 time, and counts the total number of broken log messages received within the backtracked T1 time. If the total number is equal to P1, then for each moment corresponding to the broken log message received within the backtracked T1 time, backtrack forward to T2 time. For any benchmark parameter in the broken log message within T2 time, if the backtracking determination unit does not store the optimization reference table of the benchmark parameter, then all monitoring values ​​of the benchmark parameter within the backtracked T2 time are obtained and used as the backtracking data of the benchmark parameter at the moment; If an optimization reference table of the benchmark parameters is stored, all monitored values ​​of the benchmark parameters within the retrospective time T2 are intercepted to obtain several sets of retrospective data of the benchmark parameters at the time. The interception process is as follows: S21: First, obtain all the characterization durations contained in the reference table from the optimization of the benchmark parameters and mark them as Q1, Q2, ..., Qq in ascending order, q≥1; S22: Then, in descending order of the collection time from the current time, the collection time of each monitoring value of the benchmark parameter within the retrospective time T2 is selected as the starting time; S23: Setting a capture duration Z3, where the values ​​of Z3 are Q1, Q2, ..., Qq in sequence. Each time Z3 is taken, extracting from the optimization reference table the first characterization value and the second characterization value corresponding to the same characterization duration as the value of Z3, and using the extracted first characterization value and the second characterization value as a comparison set for this value taking; It should be noted here that if there are multiple characterization durations with the same value as Z3 in the optimization reference table, all the corresponding first characterization quantities and second characterization quantities are extracted to obtain several groups of comparison sets; Acquire monitoring values ​​of the reference parameter corresponding to all acquisition moments within a time period Z3 starting from the starting time, where the values ​​of Z3 are Q1, Q2, ..., Qq respectively; Assign all the collected time values ​​obtained in the order of the collection time to x, assign the corresponding monitoring value at the corresponding collection time to y, substitute it into the linear function y=kx+b to obtain several groups of linear equations with one variable, combine any two of the linear equations with one variable to obtain several groups of linear equations with two variables, solve these linear equations with two variables to obtain several values ​​of k and b; Performing data processing on the obtained values ​​of k using a discrete point filtering algorithm, and calculating the average value of all k values ​​remaining after the data processing, and calibrating the average value as the first characterizing quantity of the reference parameter at time Z3; similarly, performing data processing on the obtained values ​​of b using a discrete point filtering algorithm, and calculating the average value of all b values ​​remaining after the data processing, and calibrating the average value as the second characterizing quantity of the reference parameter at time Z3; Perform a difference comparison between the first characterization value and the second characterization value of the benchmark parameter and the first comparison characterization value and the second comparison characterization value in the comparison set respectively. If the preset interception condition is met: the difference between the first characterization value and the first comparison characterization value is less than or equal to P2, and the difference between the second characterization value and the second comparison characterization value is less than or equal to P3, then stop taking the value of Z3, and intercept the monitoring values ​​of the benchmark parameter corresponding to all acquisition moments within the Z3 time from all the monitoring values ​​of the benchmark parameter at the backtracked time T2 as a set of analysis data of the benchmark parameter at the time, P2 and P3 are the preset first interception judgment value and second interception judgment value respectively; It should be noted here that if there are several comparison sets, only one comparison set needs to meet the preset interception condition when performing difference comparison with the first characterization value and the second characterization value of the reference parameter, that is, the corresponding data is intercepted to obtain the backtracking data of the reference parameter at the said moment; It should be noted here that the collection time corresponding to the intercepted monitoring value will no longer be selected as the starting time; When the total number of remaining acquisition moments is less than P4, the interception process is stopped to obtain all the backtracking data of the benchmark parameters at this moment, where P4 is the preset standard stop operation amount; The backtracking analysis data at the moment is generated based on the acquired backtracking data of all the benchmark parameters in the broken log information and stored.

[0015] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0016] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. The intelligent yarn winding system based on intelligent control is characterized by: include: Monitoring and acquisition module, used to acquire real-time image acquisition data and monitoring acquisition data of target yarn; An identification control module is configured to perform end-break identification on the target yarn after receiving image acquisition data of the target yarn at each moment. If the identification determines that the yarn has broken at the said moment, the numerical values ​​of several operating parameters are adjusted according to preset comparison and control steps. During the control process, an end-break log information of the said moment is generated. The end-break log information includes several reference parameters and their monitoring values. A backtracking analysis and optimization module is used to determine and backtrack the amount of stored broken log information and optimize it. The backtracking analysis and optimization module includes a backtracking determination unit and an analysis and optimization unit; Whenever a broken log message at a certain moment is received, the backtracking determination unit uses the moment as the backtracking starting point, backtracks forward by Z1 time, counts the total number of broken log messages received within the backtracked Z1 time, and selects and generates, based on the statistical result, if the total number is equal to P1, analysis benchmark information of an analysis cycle is generated according to a preset first backtracking determination rule, and deep characterization data of the analysis cycle is determined according to the analysis benchmark information, where the deep characterization data includes several characterization data sets including deep causes and several benchmark parameters; For any deep cause, when the amount of deep characterization data containing the deep cause stored in the analysis and optimization unit reaches a preset fixed amount, all deep characterization data containing the deep cause are analyzed to obtain an optimized reference table of several benchmark parameters of the deep cause and transmit the optimized reference table to the retrospective determination unit; The backtracking determination unit receives and stores the broken log information at each moment with several benchmark parameters. The backtracking determination unit uses the moment as the backtracking starting point, backtracks forward Z1 time, and counts the total number of broken log information received within the backtracked Z1 time. If the total number is equal to P1, the analysis benchmark information of an analysis cycle is generated according to the preset second backtracking determination rule, and the deep characterization data of the analysis cycle is determined according to the analysis benchmark information.

2. The intelligent yarn winding system based on intelligent control according to claim 1 is characterized in that: The first retrospective determination rule is as follows: If the total number is equal to P1, then for each moment corresponding to the broken log information received within the backtracked Z1 time, backtrack forward to Z2 time, obtain all monitoring values ​​of each benchmark parameter in the broken log information within the backtracked Z2 time and generate the backtracking analysis data of the moment based on them, Z1 and Z2 are the preset first standard backtracking time and second standard backtracking time respectively; The analysis benchmark information of an analysis cycle is generated based on the backtracking analysis data at the moment corresponding to each broken log information received within the backtracking Z1 time.

3. The intelligent yarn winding system based on intelligent control according to claim 2 is characterized in that: The steps for determining the deep characterization data of the analysis period according to the analysis benchmark information are as follows: The analysis benchmark data of the analysis period is displayed to the management personnel for review. The management personnel determines the deep reason why the number of breakages in the analysis period reaches P1 times by reviewing the changes in the monitoring values ​​of each benchmark parameter in the retrospective analysis data table at each moment. During the analysis process, the management personnel will extract some data that can determine the deep reason from the analysis benchmark data to obtain the deep characterization data of the analysis period.

4. The intelligent yarn winding system based on intelligent control according to claim 1 is characterized in that: Among the several characterization data sets of several benchmark parameters contained in a deep characterization data, one characterization data set corresponds to one benchmark parameter, one benchmark parameter corresponds to several characterization data sets, and one characterization data set contains continuous monitoring values ​​of the corresponding benchmark parameter at several collection moments.

5. The intelligent yarn winding system based on intelligent control according to claim 1 is characterized in that: The analysis steps for obtaining an optimized reference table of several benchmark parameters for the underlying causes are as follows: S11: The deep characterization data containing the deep cause stored in the analysis and optimization unit are marked as A1, A2, ..., Aa, a ≥ 1, the deep characterization data A1, A2, ..., Aa are traversed, all the benchmark parameters are extracted and deduplicated, and all the remaining benchmark parameters after deduplication are marked as B1, B2, ..., Bb, b ≥ 1; S12: Extract all representation data sets corresponding to the benchmark parameters B1 from the deep standard data A1, and mark them as C1, C2, ..., Cc, where c≥1; S13: creating a linear record variable H1 of the baseline parameter B1 relative to the underlying cause, wherein the initial value of the linear record variable H1 is 0; Determine whether the detection value and the acquisition time have a strong linear correlation characteristic in combination with the characterization data set C1. If it is determined that the linear correlation characteristic is strong, calculate and obtain the first characterization value, the second characterization value, and the characterization duration of the characterization data set C1. Otherwise, no processing is performed. S14: According to S13, the detection values ​​and the acquisition time are sequentially combined with the characterization data sets C2, C3, ..., Cc to determine whether they have a strong linear correlation characteristic. After the determination is completed, the value of the linear record variable H1 at this time is obtained. If the value of H1 is greater than or equal to H, it is determined that the benchmark parameter B1 is qualified for optimization, and an optimization data table for the benchmark parameter B1 is created, where H1 is the critical threshold for strong linear correlation determination; if the value of H1 is less than H, it is determined that the benchmark parameter B1 is not qualified for optimization; S15: If the optimized data table of the benchmark parameter B1 is created in S14, all characterization data sets corresponding to the benchmark parameter B1 are extracted from the benchmark parameters B2, B3, ..., Bb in sequence according to S11 to S14 and several optimized data tables of the benchmark parameter B1 are obtained according to S13 to S14; Merge the data of each row of all optimization data tables of the benchmark parameter B1 into one optimization data table to obtain an optimization reference table of the benchmark parameter B1; S16: determining whether the benchmark parameters B2, B3, ..., Bb have optimization qualifications in sequence according to S11 to S15, and obtaining an optimization reference table of all benchmark parameters that have optimization qualifications based on the determination results; The analysis and optimization unit optimizes a reference table of several benchmark parameters of the deep-seated cause.

6. The intelligent yarn winding system based on intelligent control according to claim 5, characterized in that: S13, the determination content is as follows: SS21: Extract all monitoring values ​​D1, D2, ..., Dd from the characterization data set C1 in the order of collection time, where d ≥ 1. The collection time of the monitoring values ​​D1, D2, ..., Dd is marked as T1, T2, ..., Td respectively. SS22: Calculate the linear correlation coefficient E1 between the acquisition time and the monitoring value based on the monitoring values ​​D1, D2, ..., Dd and the acquisition time T1, T2, ..., Td. The linear correlation coefficient E1 is used to measure the correlation between the acquisition time and the monitoring value. The linear correlation coefficient E1 has positive and negative values, and its value ranges from -1 to 1. A positive value indicates positive correlation, and a negative value indicates negative correlation. The closer the absolute value is to 1, the stronger the linear relationship. SS23: Compare |E1| and E. If |E1| ≥ E, it is determined that the monitoring value and the acquisition time in the characterization data set C1 have a strong linear correlation characteristic; At this time, the acquisition times T1, T2, ..., Td are assigned to x in sequence, and the monitoring values ​​D1, D2, ..., Dd are assigned to y accordingly. Substituting them into the linear function y = kx + b, we get d groups of linear equations with one variable. Solving any two of these linear equations together, we get d(d-1) / 2 groups of linear equations with two variables. Solving these linear equations with two variables, we get d(d-1) / 2 values ​​of k and b. Use a discrete point filtering algorithm to process the obtained d(d-1) / 2 k values, and calculate the average value of all k values ​​remaining after data processing, and calibrate the average value as the first characterization quantity of the characterization data set C1. Similarly, obtain the second characterization quantity of the characterization data set C1; use d as the characterization duration of the characterization data set C1.

7. The intelligent yarn winding system based on intelligent control according to claim 2, characterized in that: The second retrospective determination rule for generating analysis benchmark information for one analysis cycle is as follows: If the total number is equal to P1, then for each moment corresponding to the broken log information received within the backtracked time T1, backtrack forward to time T2, and for any benchmark parameter in the broken log information within time T2, if the backtracking determination unit does not store an optimization reference table for the benchmark parameter, then obtain all monitoring values ​​of the benchmark parameter within the backtracked time T2 and use them as the backtracking data of the benchmark parameter at the moment; If an optimization reference table of the benchmark parameter is stored, all monitored values ​​of the benchmark parameter within the retrospective time T2 are intercepted to obtain a number of retrospective data of the benchmark parameter at the time; Generate the back-tracing analysis data at the moment according to the obtained back-tracing data of all the benchmark parameters in the broken log information; The analysis benchmark information of an analysis cycle is generated based on the retrospective analysis data at the moment corresponding to each broken log information received within the retrospective Z1 time.

Images