Method for process data time correction using visibility graph algorithm for process management in continuous production process

Abstract

The present invention relates to a method for correcting process data time using a Visibility Graph Algorithm (VGA) for process control in a continuous production process, and more specifically, to a method for correcting process data time using a Visibility Graph Algorithm (VGA) for process control in a continuous production process, which is performed in a continuous product manufacturing system including a cable process for analyzing collected process data, verifying the presence or absence of equipment abnormalities, and quality monitoring, comprising: (1) a step of designating an analysis target location for process data analysis, preparing Y data which is quality data measured at a final quality inspection corresponding to the designated analysis target location, and loading parameters required for process data analysis; (2) a step of loading X data which constitutes the analysis data measured at the analysis target location designated in step (1) for a defined length; (3) a step of generating a Visibility Graph of Y data to check the time lag using the data prepared in step (2); (4) a step of generating a visibility graph by moving the X data prepared in step (2) by a defined interval, and calculating the adjacency matrix and Frobenius norm with the Y data to check the time delay of each X data; (5) a step of correcting the X data using the time delay confirmed in step (4); (6) a step of checking the distortion value of the X data confirmed through step (5) and, if it differs from a predefined value, generating a user alarm to warn; (7) a step of statistically comparing the correction value of the time delay of the X data with the previous value through step (5) and, if an anomaly occurs, generating an alarm to warn; and (8) a step of conducting an analysis using the X data and the time-corrected X data. According to the process data time correction method using a Visibility Graph Algorithm (VGA) for process control in a continuous production process proposed in this invention, the method comprises: designating an analysis target location for process data analysis; preparing Y data, which is quality data measured during the final quality inspection corresponding to the designated analysis target location, and loading parameters necessary for process data analysis; loading X data, which constitutes the analysis data measured at the designated analysis target location, by a defined length; generating a visibility graph of the Y data; generating a visibility graph while moving the X data by a defined interval, and calculating the adjacency matrix and Frobenius Norm with respect to the Y data to check the time delay of each X data; correcting the X data using the confirmed time delay; checking the confirmed distortion value of the X data and, if it differs from a predefined value, generating a user alarm to issue a warning; statistically comparing the corrected value of the time delay of the X data with the previous value and, if an anomaly occurs, generating an alarm to issue a warning; and X data and the time-corrected By configuring the system to include a step of conducting analysis using X data, it is possible to use collected process data for analysis in a continuous production process that produces long, continuous products such as wire products or submarine cables, while using a visibility graph algorithm to compensate for data time delays caused by process characteristics. In addition, according to the method for correcting process data time using a Visibility Graph Algorithm (VGA) for process management in a continuous production process of the present invention, collected process data for analysis is used in a continuous production process that produces long and continuous products such as wire products or submarine cables, and the time delay of data occurring due to process characteristics is corrected by using a Visibility Graph Algorithm. By using the Visibility Graph Algorithm (VGA) instead of the existing Cross-correlation method for correcting data time distortion, more accurate time correction of data is performed, thereby enabling more accurate analysis management when analyzing, monitoring, and predicting process data in a continuous process.

Description

Technology Field

[0001] The present invention relates to a method for correcting process data time using a Visibility Graph Algorithm (VGA) for process management in a continuous production process. More specifically, the invention relates to a method for correcting process data time using a VGA for process management in a continuous production process that produces long and continuous products, such as wire products or submarine cables, by using collected process data for analysis and employing a visibility graph algorithm to correct time delays in data occurring due to process characteristics. Background Technology

[0002] The content described in this section merely provides background information regarding an embodiment of the present invention and does not constitute prior art.

[0004] In general, maintaining high product quality and ensuring the trouble-free operation and maintenance of production equipment are critical in the manufacturing process. In particular, intermediate or final quality in production cannot be considered separately from the operating status of the equipment; they are closely interrelated. If quality control of products or the detection of equipment anomalies is missed at the appropriate time, the delivery of defective products can lead to a decline in the supplier's credibility. Furthermore, even if the issue is confirmed during the final inspection, the processing of defective products continues from the affected stage to the next, resulting in a waste of productivity in the intermediate process.

[0006] To prevent this, many production sites implement production process monitoring. This monitoring utilizes quality inspections during intermediate stages and collects production parameters output from manufacturing equipment. When an anomaly is detected during this quality and equipment monitoring, on-site workers or quality managers immediately verify the information and take corrective action. This prevents the spread of defects and stops the issue from escalating into a more serious equipment malfunction, thereby preventing greater financial losses.

[0008] In particular, unlike intermittent processes, continuous production processes such as cable manufacturing experience distortions—such as time lag or forwarding—in data collected from production and measurement equipment for analysis, prediction, and monitoring, depending on the specific characteristics of the product's manufacturing method. Specifically, the reasons for this temporal distortion in data collected during continuous processes are diverse, including the joining of workpieces (e.g., twisting of copper wires), heating for sheath formation and strength development, sheath joining for insulation (heating), cooling during the transition to the next process (shrinkage of the workpiece), and the accumulation of errors in length sensors used to verify the length of the workpiece during the continuous process. Generally, these errors stem from the expansion of the product (increase in length) caused by heating and firing, the contraction (reduction in length) caused by cooling, and the accumulation of errors or noise components in the data collection sensors.

[0010] Furthermore, analysis and monitoring are conducted using the collected data. Taking the monitoring or analysis of the final process as an example, quality data (Y data) measured during the final quality inspection is combined with data (X data) from each production or measurement facility prior to the final process to perform an analysis (XY analysis). At this stage, since the equipment process data or measurement data constituting the X data are subject to time distortion (time shifting or time delay), time correction is performed. The corrected X data is then combined with the Y data to form the analysis data, after which the analysis is conducted. In other words, when correcting the time of the X data, a cross-correlation with the Y data is generally calculated. The position is then aligned by either shifting the X data to the location with the largest value (t) (moving upward relative to the data table when the position with the largest cross-correlation value is t+n) or shifting the X data (moving downward relative to the data table when the position with the largest cross-correlation value is tn). Due to the limitations of the cross-correlation used in this process, a more efficient method is required to obtain more accurate analysis results. One disadvantage of cross-correlation is that it is difficult to evaluate the mutual relationship between strongly correlated data. Figure 3 shows an example of calculating cross-correlation in strongly correlated data. In Figure 3, the point with the highest cross-correlation is (T+2), (210), but due to the strong mutual relationship between the data, there is no significant difference compared to the surrounding points (220), (230), indicating low discriminability.

[0012] As such, in continuous production processes at industrial sites where long, continuous processed products such as wire products or submarine cables are produced, there is a need to correct for time lag or time advance caused by the characteristics of the continuous process in order to obtain more accurate results in the process of quality analysis and prediction of the final product, quality prediction or monitoring of quality abnormalities during intermediate production processes, or monitoring of equipment abnormalities or analysis or prediction of quality abnormalities caused by equipment abnormalities. However, in the case of existing cross-correlation, there was a problem in that accurate time correction was difficult because it was not robust to various environments in the continuous production process, and the accuracy of process data analysis and quality control was reduced due to the resulting time distortion of the collected data. Korean Patent Publication No. 10-2019-0040162 is disclosed as a prior art document.

[0014] The aforementioned background technology is technical information that the inventor possessed for the derivation of the present invention or acquired during the process of deriving the present invention, and it cannot be considered as publicly known technology disclosed to the general public prior to the filing of the present invention. The problem to be solved

[0015] The present invention is proposed to solve the aforementioned problems of previously proposed methods, and comprises the steps of: designating an analysis target location for process data analysis; preparing Y data, which is quality data measured during the final quality inspection corresponding to the designated analysis target location, and loading parameters necessary for process data analysis; loading X data, which constitutes the analysis data measured at the designated analysis target location, by a defined length; generating a visibility graph of the Y data; generating a visibility graph while moving the X data by a defined interval, and calculating the adjacency matrix and Frobenius Norm with respect to the Y data to check the time delay of each X data; correcting the X data using the confirmed time delay; checking the confirmed distortion value of the X data and, if it differs from a predefined value, generating a user alarm to warn; statistically comparing the corrected value of the time delay of the X data with the previous value and, if an anomaly occurs, generating an alarm to warn; and conducting analysis using the X data and the time-corrected X data, thereby enabling wire products or The purpose is to provide a method for correcting process data time using a Visibility Graph Algorithm (VGA) for process control in a continuous production process, which uses collected process data for analysis in a continuous production process that produces long, continuous products such as submarine cables, and uses a visibility graph algorithm to correct for data time delays caused by process characteristics.

[0017] In addition, another objective of the present invention is to provide a method for correcting process data time using a Visibility Graph Algorithm (VGA) for process management in a continuous production process, which uses collected process data for analysis in a continuous production process that produces long and continuous products such as wire products or submarine cables, and uses a Visibility Graph Algorithm to correct time delays in data that occur due to process characteristics, thereby performing more accurate time correction of data by using a Visibility Graph Algorithm (VGA) instead of the existing Cross-correlation to correct data time distortions, so that more accurate analysis management is possible when analyzing, monitoring, and predicting process data in a continuous process.

[0019] However, the technical problem that the present invention aims to solve is not limited to the technical problem described above, and other technical problems may exist. means of solving the problem

[0020] A process data time correction method using a Visibility Graph Algorithm (VGA) for process management in a continuous production process according to the features of the present invention for achieving the above-mentioned purpose is,

[0021] As a process data time correction method using VGA (Visibility Graph Algorithm) for process control in a continuous production process, it is performed in a continuous product manufacturing system including a cable process for analyzing collected process data, verifying the presence or absence of equipment abnormalities, and quality monitoring.

[0022] (1) A step of designating a target location for process data analysis, preparing Y data which is quality data measured in the final quality inspection corresponding to the designated target location for analysis, and loading parameters required for process data analysis;

[0023] (2) A step of loading X data, which constitutes the analysis data and is measured at the analysis target location specified in step (1) above, for a defined length;

[0024] (3) A step of generating a visibility graph of Y data to check the time lag using the data prepared in step (2) above;

[0025] (4) A step of generating a visibility graph by moving the X data prepared in step (2) by a defined interval, and calculating the adjacency matrix and Frobenius norm with the Y data to determine the time delay of each X data;

[0026] (5) A step of correcting X data using the time delay identified in step (4) above;

[0027] (6) A step of verifying the confirmed distortion value of X data through the above step (5), and if it differs from a predefined value, generating a user alarm to warn;

[0028] (7) A step of statistically comparing the correction value of the time delay of X data with the previous value through the above step (5) to generate an alarm and issue a warning when an abnormality occurs; and

[0029] (8) Its structural feature is that it includes a step of conducting analysis using X data and time-corrected X data.

[0031] Preferably, in step (1),

[0032] A location for analysis targeting process data analysis is designated, and Y data, which is quality data measured in the final quality inspection corresponding to the designated location for analysis targeting analysis, is prepared, and parameters required for process data analysis are loaded, wherein the parameters required for process data analysis can be defined as data lengths for detecting the occurrence of lag to verify time lag.

[0034] Preferably, in step (2),

[0035] X data constituting analysis data is loaded for a defined length by measuring at the analysis target location specified in step (1) above, and the X data may consist of data collected by measuring at the analysis target location from measurement equipment each placed in the continuous production process.

[0037] Preferably, in step (3),

[0038] In order to check the time lag with the data prepared in step (2) above, a visibility graph of Y data is generated, and this can be done by a process of loading Y data, a process of generating a visibility graph of the loaded Y data, and a process of generating an adjacency matrix with the generated visibility graph.

[0040] Preferably, in step (4),

[0041] The above step (2) can be performed by generating a visibility graph while moving the X data prepared in that step by a defined interval, and calculating the adjacency matrix and Frobenius norm with the Y data to check the time delay of each X data, and for all X data to be analyzed, the process of generating a visibility graph, generating an adjacency matrix, calculating the difference between the two matrices using the visibility graph and the adjacency matrix, and then calculating the Frobenius norm, the process of identifying the location with the smallest Frobenius norm value for each X data, and the process of taking the data from the location with the identified minimum Frobenius norm value for each X data and analyzing the Y data and the analysis data.

[0043] Preferably, in step (6),

[0044] Through the above step (5), the confirmed distortion value of the X data is checked, and if it differs from the predefined value, a user alarm is generated to warn. However, if the confirmed time lag is greater or smaller than the parameter value defined by the user using the time lag confirmed in the collected process data, or if it is outside the range or included in a specific range, it is determined that there is an abnormal situation in the process and an alarm may be generated.

[0046] Preferably, in step (7),

[0047] Through the above step (5), the correction value of the time delay of the X data is statistically compared with the previous value to generate an alarm and warn when an abnormality occurs, and an alarm can be generated if the confirmed time delay is greater than 2 sigma (σ) compared to the previously confirmed time delay. Effects of the invention

[0048] According to the process data time correction method using a Visibility Graph Algorithm (VGA) for process control in a continuous production process proposed in this invention, the method comprises: designating an analysis target location for process data analysis; preparing Y data, which is quality data measured during the final quality inspection corresponding to the designated analysis target location, and loading parameters necessary for process data analysis; loading X data, which constitutes the analysis data measured at the designated analysis target location, by a defined length; generating a visibility graph of the Y data; generating a visibility graph while moving the X data by a defined interval for each, and calculating the adjacency matrix and Frobenius Norm with respect to the Y data to check the time delay of each X data; correcting the X data using the confirmed time delay; checking the confirmed distortion value of the X data and, if it differs from a predefined value, generating a user alarm to warn; statistically comparing the corrected value of the time delay of the X data with the previous value and, if an anomaly occurs, generating an alarm to warn; and X data and the time-corrected By configuring the system to include a step of conducting analysis using X data, it is possible to use collected process data for analysis in a continuous production process that produces long, continuous products such as wire products or submarine cables, while using a visibility graph algorithm to compensate for data time delays caused by process characteristics.

[0050] In addition, according to the method for correcting process data time using a Visibility Graph Algorithm (VGA) for process management in a continuous production process of the present invention, collected process data for analysis is used in a continuous production process that produces long and continuous products such as wire products or submarine cables, and the time delay of data occurring due to process characteristics is corrected by using a Visibility Graph Algorithm. By using the Visibility Graph Algorithm (VGA) instead of the existing Cross-correlation method for correcting data time distortion, more accurate time correction of data is performed, thereby enabling more accurate analysis management when analyzing, monitoring, and predicting process data in a continuous process.

[0052] Furthermore, the various and beneficial advantages and effects of the present invention are not limited to those described above and may be more easily understood in the process of explaining specific embodiments of the present invention. Brief explanation of the drawing

[0053] FIG. 1 is a diagram illustrating the configuration of a continuous product manufacturing system in functional blocks for performing a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. FIG. 2 is a diagram illustrating the flow of a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. Figure 3 is a diagram illustrating an example explaining the disadvantage that discriminative power is lower when there is a strong correlation in an analysis using correlation in data analysis. FIG. 4 is a diagram illustrating an example in which the VGA has robust characteristics against the environment even in situations where various distortions occur, such that the same graph is drawn when the Visibility Graph is drawn even in a state with various distortions. FIG. 5 is a diagram showing a cable and the position of the cable used as an example to explain a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. FIG. 6 is a drawing showing the process equipment used as an example to explain a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention, the location of the process inspection and measurement equipment, and the location of the quality measurement equipment. FIG. 7 is a diagram showing data to be prepared according to parameters used for time delay verification to verify a time delay occurring in a continuous process, such as cable production, which is applied to a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention, data length, and a first location (location where data must be retrieved) for calculating the time delay in, for example, X1 data. FIG. 8 is a diagram illustrating the flow of a detailed process of step S130 in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. FIG. 9 is a diagram illustrating the process of determining the time delay of X data by calculating the visibility graph, adjacency matrix, and Frobenius norm of X data during the process of step S140 in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. FIG. 10 is a diagram illustrating the process of checking the time delay of all X data during step S140 in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. FIG. 11 is a diagram showing the identified time delay of each X data in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. FIG. 12 is a diagram illustrating an example of a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention, wherein data is retrieved for analysis and configured into XY data using the identified time delay of each X data in the process of step S150. FIG. 13 is a diagram showing an example of generating a visibility graph in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. FIG. 14 is a diagram showing an example of generating an adjacency matrix using a visibility graph in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. Specific details for implementing the invention

[0054] Embodiments of the present invention are described below with reference to the attached drawings so that those skilled in the art can easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals.

[0056] Throughout the specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected" but also cases where they are "indirectly connected" with other elements interposed between them. Furthermore, when a part is described as "including" a component, this means that, unless specifically stated otherwise, it does not exclude other components but rather allows for the inclusion of additional components; it should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0058] The following examples are detailed descriptions to aid in understanding the present invention and are not intended to limit the scope of the present invention. Accordingly, inventions within the same scope that perform the same function as the present invention will also fall within the scope of the present invention.

[0060] In addition, each component, process, procedure, or method included in each embodiment of the present invention may be shared within a scope that is not technically contradictory to one another.

[0062] FIG. 1 is a diagram illustrating the configuration of a continuous production product manufacturing system as functional blocks for performing a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. As illustrated in FIG. 1, a continuous production product manufacturing system (10) for performing a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention comprises: a parameter setting unit (11) for designating an analysis target location for process data analysis, which loads Y data, which is quality data measured in a final quality inspection corresponding to the designated analysis target location, and parameters required for process data analysis, and loads X data, which constitutes analysis data measured at the designated analysis target location, by a defined length; a measurement unit (12) composed of measurement equipment provided in plurality on the continuous production process, which measures X data and Y data at the analysis target location designated through the parameter setting unit (11); a VGA generation module unit (13) for generating a visibility graph of Y data and X datat; a time delay correction unit (14) for checking the time delay of X data and correcting the time delay of X data accordingly; an alarm unit (15) for checking the confirmed distortion value of X data and warning the user with an alarm when an abnormality occurs; and an analysis unit (16) for performing analysis using X data and time-corrected X data. It can be configured to include.

[0064] FIG. 2 is a diagram illustrating the flow of a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. As illustrated in FIG. 2, a method for correcting process data time using VGA for process control in a continuous production process according to an embodiment of the present invention comprises: a step of designating an analysis target location for process data analysis; a step of preparing Y data, which is quality data measured in a final quality inspection corresponding to the designated analysis target location, and loading parameters necessary for process data analysis (S110); a step of loading X data, which constitutes the analysis data measured at the designated analysis target location, by a defined length (S120); a step of generating a visibility graph of Y data (S130); a step of generating a visibility graph while moving X data by a defined interval, and calculating an adjacency matrix and a Frobenius norm with respect to Y data to check the time delay of each X data (S140); a step of correcting X data using the confirmed time delay (S150); a step of checking the confirmed distortion value of X data and, if it differs from a predefined value, generating a user alarm to warn (S160); and a step of transferring the correction value of the time delay of X data The method may be implemented by including a step (S170) of generating a warning by statistically comparing the value and when an abnormality occurs, and a step (S180) of conducting an analysis using X data and time-corrected X data. In this embodiment of the present invention, a process data time correction method using a Visibility Graph Algorithm (VGA) for process control in a continuous production process may be performed in a continuous production system (10) that includes a cable process for analyzing collected process data, verifying the presence or absence of abnormalities in equipment, and quality monitoring.

[0066] FIG. 3 is a diagram illustrating an example explaining the disadvantage that the discriminative power is lowered when there is a strong correlation in an analysis using correlation in data analysis; FIG. 4 is a diagram illustrating an example showing that VGA has robust characteristics against the environment even in situations where various distortions occur, as the same graph is drawn when a Visibility Graph is drawn even in a state with various distortions; FIG. 5 is a diagram showing a cable and the location of the cable used as an example to explain a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention; FIG. 6 is a diagram showing process equipment, the location of process inspection equipment, and the location of quality measurement equipment used as an example to explain a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention.

[0068] Additionally, FIG. 7 is a diagram showing the data to be prepared according to the parameters used for time delay verification to verify a time delay occurring in a continuous process, such as cable production, which is applied to the process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention, the data length, and the first location (location where data must be retrieved) for calculating the time delay in the example X1 data; FIG. 8 is a diagram illustrating the flow of the detailed process of step S130 in the process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention; FIG. 9 is a diagram illustrating the process of verifying the time delay of X data by calculating the visibility graph, adjacency matrix, and Frobenius norm of X data during the process of step S140 in the process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention; and FIG. 10 is a diagram showing all X during the process of step S140 in the process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention This is a diagram illustrating the process of verifying the time delay of the data.

[0070] In addition, FIG. 11 is a diagram showing the identified time delay of each X data in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention, FIG. 12 is a diagram explaining an example of bringing in data for analysis and configuring it into XY data for analysis using the identified time delay of each X data in the process of step S150 in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention, FIG. 13 is a diagram showing an example of generating a visibility graph in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention, and FIG. 14 is a diagram showing an example of generating an adjacency matrix using a visibility graph in a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention.

[0072] A method for correcting process data time using VGA for process control in a continuous production process according to an embodiment of the present invention relates to a method for correcting time distortions, such as time delays or time advances, in data collected from process equipment used for more accurate analysis, prediction, and monitoring when analyzing and predicting processes, such as final quality or monitoring abnormal conditions of quality equipment at intermediate stages in a continuous production process for producing cables or long products, by using VGA (Visibility Graph Algorithm). In other words, in continuous production processes at industrial sites where long, continuous processed products such as wire products or submarine cables are produced, a process is required to correct for time lag (a range from - to + in terms of time) or time advance caused by the characteristics of the continuous process in order to obtain more accurate results in the process of quality analysis and prediction of the final product, quality prediction or monitoring of quality abnormalities during the intermediate production process, monitoring of equipment abnormalities, or analysis or prediction of quality abnormalities caused by equipment abnormalities. The present invention provides a method for correcting time lag and time advance of data occurring due to process characteristics when using collected process data for analysis in a continuous production process, and includes monitoring the detection of process abnormalities in the continuous process using the sequence and the value of the confirmed time lag.

[0074] FIG. 3 illustrates an example explaining the disadvantage that discrimination power is low when there is a strong correlation in an analysis using correlation in data analysis. FIG. 4 illustrates an example showing that VGA has robust characteristics against the environment even in situations where various distortions occur, as the same graph is drawn when a Visibility Graph is plotted even in a state with various distortions. FIG. 5 shows a cable and the position of the cable used as an example to explain a process data time correction method using VGA for process control in a continuous production process according to an embodiment of the present invention. First, FIG. 3 is an example of calculating cross-correlation in data with strong correlation. Although the point with the highest cross-correlation is (T+2), (210), it can be seen that the discrimination power is low because there is no significant difference compared to the surrounding (220), (230) due to the strong inter-correlation between the data. To solve this, the present invention applies VGA (Visibility Graph Algorithm) instead of the conventionally used cross-correlation to resolve the parts that are difficult to evaluate.

[0076] In other words, the advantage of using VGA compared to cross-correlation calculation based on correlation calculation is that it has low sensitivity to outliers and is resistant to data noise or data distortion. This robustness against various types of noise or data distortion is illustrated in FIG. 4. In FIG. 4, the basic data is (310), the data with shifted data is (320), the data with a scaled level of values ​​is (330), the data with a scaled length (range) of data is (340), and the data with a trending data is (350). When the visibility graph from (310) to (350) is displayed, it takes the form of (360). In this case, since the same visibility graph is drawn even when various types of distortion occur, it can be said to be resistant to noise or distortion. That is, when the cross-correlation with fixed data of the same length from (310) to (350) in Fig. 4 is calculated, the results are all calculated as different values, indicating that the results are calculated differently depending on the situation and that the characteristics are not robust to the environment. In correcting the time distortion of data for analysis, the application of cross-correlation calculations or calculations based on correlation in continuous processes requires improvement, and the present invention presents an improved method applying the Visibility Graph Algorithm (VGA).

[0078] Additionally, FIG. 5 shows a cable as an example of a product from a continuous process such as cable production, the cable length is 1,000 m, and the point for explaining the example is Point_A (410). The explanation of the process for verifying the time delay in the final quality inspection or intermediate inspection is based entirely on Point_A (410). FIG. 6 is a schematic diagram illustrating a continuous process such as cable production, showing process equipment, intermediate inspection and measurement equipment, and process expansion and contraction, which is one of the major causes of time delays occurring during the progress of the process. At this time, FIG. 6 only shows the contraction and expansion of the product during the process for schematic explanation purposes; however, in an actual production process, there are various factors that cause time delays, such as errors in length measuring instruments and the accumulation of errors. At this time, the main content is data correction and anomaly detection when verifying and analyzing the time delay at Point_A (410) in FIG. 5. Point_A (410) passes through Process_A (520) and then passes through the inspection and measurement equipment, Inspection_A (530), after the process, and the inspection and measurement data is transmitted to and stored in the system. After that, Point_A (410) passes through Process_B (540) and then passes through the inspection and measurement equipment, Inspection_B (550), and the inspection and measurement data is transmitted to and stored in the system. Finally, Point_A (410) passes through the final quality measurement process, Quality Measurement_A (580), and the final quality data for Point_A (410) is collected. During this process, when passing through the process and proceeding to the next process, expansion (521, 541) and cooling (contraction) (561) of the product occur.

[0080] In step S110, a location for analysis targeting process data analysis is designated, Y data, which is quality data measured in the final quality inspection corresponding to the designated location for analysis targeting is prepared, and parameters required for process data analysis are loaded. In this step S110, a location for analysis targeting process data analysis is designated, Y data, which is quality data measured in the final quality inspection corresponding to the designated location for analysis targeting is prepared, and parameters required for process data analysis are loaded, wherein the parameters required for process data analysis are defined as data lengths for detecting the occurrence of lag to verify time lag.

[0082] Additionally, the process in step S110 is a process of preparing for data analysis of point_A (410) in FIG. 5, and a process of preparing the measured values ​​when point_A (410) passes through quality measurement_A (580) and preparing the parameters required for analysis. A fixed amount of Y data of a specified data length required for analysis is prepared. The length used for explanation in the present invention is 20, and thus, data of length 20 is prepared. In addition, the parameters required for analysis include the time “-” (minus area, data pulled forward based on Y data) to “+” (plus area, data pushed backward based on Y data) when represented as data length for detecting the occurrence of delay to check time lag. The user specifies a value from “-” to “+” to check time lag, and the time lag parameter value stored in the system is loaded during the process of saving. At this time, the parameter value of the time delay is set to 20 for the “-” range and 20 for the “+” range, and the detection range for the total time delay (Time lag) can be set to -20 to +20.

[0084] In step S120, X data constituting the analysis data, which is measured at the analysis target location specified in step S110, is loaded for a defined length. In this step S120, X data constituting the analysis data, which is measured at the analysis target location specified in step S110, is loaded for a defined length, and the X data may consist of data collected by measuring at the analysis target location from measurement equipment each placed in the continuous production process. Here, the process in step S120 is a process of preparing data collected while passing through measurement_A (530), measurement_B (540), and measurement_C (550), which are measurement equipment collected at point_A (410) for analysis, and FIG. 7 shows the data prepared for analysis, and the length of data used for the set analysis is indicated. That is, if the data length used for the analysis described in the analysis system is 20, then 20 pieces of Y data are prepared (640). If the time lag detection parameter set for the analysis has a setting value of 20 (-20) in the “-” (minus) area and a setting value of 20 (+20) in the “+” (plus) area, then the data length that must be prepared to check the analysis and time lag in the data measured as point_A (410) passes through measurement_A (530), measurement_B (540), and measurement_C (550) is 60 pieces. At this time, X datas is represented as data X1 (610), data X2 (620), and data X3 (630) in FIG. 7, and the data collected when point_A (410) passes through quality measurement_A (580) in FIG. 6 is represented as data Y (640) in FIG. 7.

[0086] In step S130, a visibility graph of Y data is generated to check the time lag using the data prepared in step S120. This step S130 may consist of a process of generating a visibility graph of Y data to check the time lag using the data prepared in step S120, a process of loading Y data, a process of generating a visibility graph of the loaded Y data, and a process of generating an adjacency matrix using the generated visibility graph.

[0088] In addition, the process in step S130 is a process for checking the time lag using the data prepared in step S120, and proceeds with the data (640) of FIG. 7. Here, FIG. 8 shows the details of the process in step S130. That is, in FIG. 8, step S131 is a process in which the data (640) of FIG. 7 is prepared or input in the VGA generation module (13) of the entire system. In addition, step S132 is a process of generating a visibility graph of Y data using the data prepared in step S131. At this time, the process of drawing the visibility graph can be represented by FIG. 13 and [Equation 1] below.

[0090]

[0092] In (1210) of FIG. 13, when there are three consecutive data points t1, t2, and t3, the process of checking the visibility of data t1 and t3 is to denote data t1 as tq, t2 as tr, and t3 as ts, and substitute each of the three data points into [Equation 1]; if the equation is satisfied, visibility is confirmed. Here, visibility is checked for all data in a defined interval as a single data point. Assuming there are 20 data points in a single data interval and the data numbers are from t1 to t20, the visibility between t1 and the remaining data points following t1 is checked first. Then, the visibility between t2 and the remaining data points following t2 is checked, and this process is carried out up to t18. (1220) of FIG. 13 shows the result of checking the visibility graph from t1 to t9. After generating the visibility graph of Y Data in step S132 of FIG. 7, the process of generating an adjacency matrix using the generated visibility graph is performed in step S133. FIG. 14 shows the adjacency matrix generated in step S133. Here, the adjacency matrix is ​​a matrix representing connectivity between adjacent nodes in graph theory, and (1310) in FIG. 14 represents the input visibility graph, and (1320) represents an example of an adjacency matrix generated based on the visibility graph.

[0094] In step S140, a visibility graph is generated by moving the X data prepared in step S120 by a defined interval, and the adjacency matrix and Frobenius norm with respect to the Y data are calculated to check the time delay of each X data. This step S140 involves generating a visibility graph by moving the X data prepared in step S120 by a defined interval and calculating the adjacency matrix and Frobenius norm with respect to the Y data to check the time delay of each X data. It may consist of the following steps for all X data to be analyzed: generating a visibility graph, generating an adjacency matrix, calculating the difference between the two matrices using the visibility graph and the adjacency matrix, and then calculating the Frobenius norm; identifying the location with the smallest Frobenius norm value for each X data; and retrieving the data at the location with the identified minimum Frobenius norm value for each X data to analyze the Y data and the analysis data.

[0096] Additionally, in step S140, once the adjacency matrix is ​​generated in step S130, as illustrated in FIG. 9, the process in step S141-1 loads data from the prepared X1 data. The prepared X1 data is data reflecting time lag parameters to find time lag in the system; in the example above, the time lag parameter is -20 to +20, and since the length of the Y data is 20, there are 60 parameters reflecting the data length and parameters. The first data to be loaded are the first 20 of the 60 data points reflecting the parameter -20 based on the Y data. This is position (650) in FIG. 7, and 20 data points are loaded at position (650). In step S141-2, after loading, a visibility graph is generated for the Y data in the same manner as the process. Subsequently, in step S141-3, after generating the visibility graph, an adjacency matrix is ​​generated in the same way as for the Y data. The difference between the two matrices is calculated: the adjacency matrix of the Y data generated in step S133 of Fig. 8 and the adjacency matrix of the X1 data generated in step S141-3 of Fig. 9. Here, the method for calculating the difference is to determine the difference in the values ​​of the elements at the same position in each matrix. The matrix of differences between the two matrices is subjected to the calculation of the Frobenius Norm. This Frobenius Norm calculation is performed using [Equation 2] below.

[0098]

[0100] After calculating the Frobenius Norm for data with a time lag of -20 in step S141-4 of Fig. 9, proceed to step S141-5 (850). Step S141-5 prepares 20 data points with a time lag starting at -19, increasing the time lag by 1 compared to the data in step S141-4. After step S141-5, proceed to step S141-2 with the prepared data, and repeat step S141-2 back to step S141-5. This iterative process proceeds from the time lag parameter -20 to +20, and once the Frobenius Norm is calculated up to the data at the +20 position, the calculation for X1 data is terminated. When the process of Fig. 9 for the X1 data is finished, the X2 data is prepared again and the process for the X2 data is performed, and this process is performed for all X data. This corresponds to step S141 of Fig. 10.

[0102] Additionally, when the calculation of all prepared X data is completed by proceeding to step S141 in Fig. 10, the location having the minimum Frobenius Norm value is identified for each X data. At this time, the location having the minimum value is the time-lag location in that X data. Step S142 is the process of identifying the time-lag location of each X data for all X. Afterward, when step S142 is completed, step S143 is performed to retrieve the identified time-lag data and configure the data to be analyzed with the Y data. Step S143 is illustrated as an example in Fig. 11. When the time lag value identified in the X1 data in Fig. 11 is +5, the time lag value identified in the X2 data is +10, and the time lag value identified in the X3 data is -5, as shown in Fig. 12, 10 data points corresponding to the length of the Y data with the corresponding time lag can be taken to construct analysis data. The data constructed in Fig. 12 is (1150). After performing step S143 of Fig. 10, the time lag location of all X data can be identified, and the identified time lag values ​​can be used to monitor for abnormalities in the process.

[0104] In step S150, the X data are corrected using the time delay identified in step S140. In this step S150, as illustrated in FIGS. 11 and FIGS. 12 respectively, the X data are corrected using the identified time delay, and the time delay data is pulled or pushed.

[0106] In step S160, the distortion values ​​of the X data identified through step S150 are checked, and if they differ from predefined values, a user alarm is generated to provide a warning. In this step S160, the distortion values ​​of the X data identified through step S150 are checked, and if they differ from predefined values, a user alarm is generated to provide a warning. Additionally, using the time lag identified in the collected process data, if the identified time lag is greater or smaller than the user-defined parameter value, falls outside a range, or falls within a specific range, it is determined to be an abnormal situation in the process, and an alarm may be triggered. Here, the identified distortion values ​​of the X data can function to be used for monitoring the presence of abnormalities during the continuous process. In the process of step S160, if the calculated time lag differs from the usual during the continuous process, it can be considered a process abnormality or a measurement abnormality. Since such process abnormalities lead directly to quality abnormalities, rapid abnormality detection and subsequent corrective measures are also important in quality control.

[0108] Additionally, the process in step S160 can function to determine a process anomaly and generate an alarm if the time lag value of each X data is greater or less than a value specified by the user. In this case, the user-specified value is prepared in the process of step S110 as an anomaly detection parameter of the system. For example, if the user-specified value is between -5 and +5 and the confirmed time lag is +8, an alarm is generated. Furthermore, a range such as -5 to +5 can be set, or if it is set in only one direction to generate an alarm if it is +8 or greater, the system can operate such that an alarm is generated only when the time lag is greater than +8. In this way, the time lag value can be understood as being user-specified for each continuous process and being variable.

[0110] In step S170, the correction value of the time lag of the X data is statistically compared with the previous value through step S150, and an alarm is generated to warn if an anomaly occurs. In this step S170, the correction value of the time lag of the X data is statistically compared with the previous value through step S150 to generate an alarm to warn if an anomaly occurs, but an alarm may be generated if the confirmed time lag is greater than 2 sigma (σ) compared to the previously confirmed time lag. Here, unlike step S160, step S170 can be used as a process to check for process anomalies by applying a statistical method to the time lag of each confirmed X data relative to the existing data. For example, it can function to monitor by generating an alarm if the confirmed time lag is statistically greater than 2 sigma (σ) compared to the previously confirmed time lag. This statistical anomaly detection level can be loaded and used in the process of step S110 as a parameter specified by the user.

[0112] In step S180, an analysis is performed using X data and time-corrected X data. In this step S180, after the process of step S170 is completed, the necessary analysis can be performed using the time-lag-corrected X data and Y data. That is, in step S180, in a continuous production process including a cable process, analysis of collected process data and analysis of equipment and quality monitoring to check for abnormalities can be performed.

[0114] Thus, a process using a Visibility Graph Algorithm (VGA) for process control in a continuous production process comprises the steps of: designating an analysis target location for process data analysis; preparing Y data, which is quality data measured during the final quality inspection corresponding to the designated analysis target location, and loading parameters necessary for process data analysis; loading X data, which constitutes the analysis data measured at the designated analysis target location, by a defined length; generating a Visibility Graph of the Y data; generating a Visibility Graph while moving the X data by a defined interval, and calculating the Adjacency Matrix and Frobenius Norm with respect to the Y data to identify the time delay of each X data; correcting the X data using the identified time delay; checking the identified distortion value of the X data and generating a user alarm to warn if it differs from a predefined value; statistically comparing the corrected value of the time delay of the X data with the previous value and generating an alarm to warn if an anomaly occurs; and conducting analysis using the X data and the time-corrected X data. The data time correction method enables quality improvement, equipment management, and cost reduction by performing analysis and prediction using data collected from production and measurement equipment during a continuous production process, such as analyzing critical factors or monitoring anomaly detection. In other words, when analyzing and predicting processes—such as final quality or monitoring abnormal conditions of quality equipment at intermediate stages—in a continuous process for producing cables or long products, the present invention functions to correct time distortions in data collected from process equipment, such as time delays or time advances, in order to achieve more accurate analysis, prediction, and monitoring, thereby improving the accuracy of data analysis.

[0116] As described above, a method for correcting process data time using a Visibility Graph Algorithm (VGA) for process control in a continuous production process according to an embodiment of the present invention comprises: designating an analysis target location for process data analysis; preparing Y data, which is quality data measured at a final quality inspection corresponding to the designated analysis target location, and loading parameters necessary for process data analysis; loading X data, which constitutes the analysis data measured at the designated analysis target location, by a defined length; generating a visibility graph of Y data; generating a visibility graph while moving X data by a defined interval, and calculating an adjacency matrix and a Frobenius norm with respect to Y data to check the time delay of each X data; correcting X data using the confirmed time delay; checking the confirmed distortion value of X data and, if it differs from a predefined value, generating a user alarm to warn; and statistically comparing the corrected value of the time delay of X data with the previous value to generate an alarm to warn when an anomaly occurs. By configuring the method to include a step of performing analysis using X data and time-corrected X data, it is possible to use collected process data for analysis in a continuous production process for producing long, continuous processed products such as wire products or submarine cables, and to correct for data time delays caused by process characteristics using a visibility graph algorithm; in particular, by using collected process data for analysis in a continuous production process for producing long, continuous processed products such as wire products or submarine cables, and to correct for data time delays caused by process characteristics using a visibility graph algorithm,By using the Visibility Graph Algorithm (VGA) instead of the existing cross-correlation algorithm to correct data temporal distortion, more accurate temporal correction of data is performed, enabling more precise analytical management during the analysis, monitoring, and prediction of process data in continuous processes.

[0118] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will understand that other specific forms can be easily modified without altering the technical spirit or essential features of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0120] The scope of the present invention is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and equivalent concepts thereof should be interpreted as being included within the scope of the present invention. Explanation of the symbols

[0121] 10: Continuous Production System 11: Parameter setting section 12: Inspection and measurement unit 13: VGA Generation Module 14: Time delay correction unit 15: Alarm section 16: Analysis Department S110: A step of designating an analysis target location for process data analysis, preparing Y data which is quality data measured in the final quality inspection corresponding to the designated analysis target location, and loading parameters required for process data analysis. S120: A step of loading X data, which constitutes the analysis data and is measured at a designated analysis target location, to a defined length. S130: Step to generate the visibility graph of the Y data S140: A step of generating a visibility graph by moving each X data by a defined interval, and checking the time delay of each X data by calculating the adjacency matrix and Frobenius norm with the Y data. S150: Step of correcting X data using the identified time delay S160: A step of checking the confirmed distortion values ​​of X data and generating a user alarm to warn if they differ from predefined values. S170: A step that statistically compares the time delay correction value of X data with the previous value and generates a warning by generating an alert if an anomaly occurs. S180: Step of conducting analysis using X data and time-corrected X data

Claims

Claim 1 A method for correcting process data time using a VGA (Visibility Graph Algorithm) for process control in a continuous production process, which is performed in a continuous production product manufacturing system (10) including a cable process for analyzing collected process data, verifying the presence or absence of equipment abnormalities, and quality monitoring, comprising: (1) a step of designating an analysis target location for process data analysis, preparing Y data which is quality data measured in the final quality inspection corresponding to the designated analysis target location, and loading parameters required for process data analysis; (2) a step of loading X data, which constitutes the analysis data measured at the analysis target location designated in step (1), by a defined length; (3) a step of generating a visibility graph of Y data to check the time lag using the data prepared in step (2); (4) generating a visibility graph while moving the X data prepared in step (2) by a defined interval, and calculating an adjacency matrix and a Frobenius norm with respect to the Y data to check the time lag of each X data. A method for correcting process data using VGA for process control in a continuous production process, characterized by comprising: (5) a step of correcting X data using the time delay identified in step (4); (6) a step of identifying the distortion value of X data through step (5) and, if different from a predefined value, generating a user alarm to warn; (7) a step of statistically comparing the correction value of the time delay of X data through step (5) with the previous value and, if an anomaly occurs, generating an alarm to warn; and (8) a step of conducting an analysis using X data and time-corrected X data. Claim 2 A method for correcting process data time using VGA for process control in a continuous production process, characterized in that, in step (1) above, a location for analysis target for process data analysis is designated, and Y data, which is quality data measured in the final quality inspection corresponding to the designated location for analysis target, is prepared, and parameters required for process data analysis are loaded, wherein the parameters required for process data analysis are defined as data lengths for detecting the occurrence of lag to verify time lag. Claim 3 A method for correcting process data time using VGA for process control in a continuous production process, wherein in step (2), X data constituting analysis data is loaded for a defined length and X data is collected by measuring at a designated analysis target location in step (1), and the X data consists of data collected by measuring at a designated analysis target location from measuring equipment each placed in a continuous production process. Claim 4 A method for correcting process data time using VGA for process control in a continuous production process, characterized in that, in step (3), a visibility graph of Y data is generated to check the time lag with the data prepared in step (2), and the process comprises loading Y data, generating a visibility graph of the loaded Y data, and generating an adjacency matrix with the generated visibility graph. Claim 5 A method for correcting process data time using VGA for process control in a continuous production process, characterized in that, in step (4), the X data prepared in step (2) are each moved by a defined interval to generate a visibility graph, and the adjacency matrix and Frobenius Norm with respect to the Y data are calculated to determine the time delay of each X data, wherein for all X data to be analyzed, the process of generating a visibility graph, generating an adjacency matrix, calculating the difference between the two matrices using the visibility graph and the adjacency matrix, and then calculating the Frobenius Norm; the process of determining the location having the smallest Frobenius Norm value for each X data; and the process of taking the data at the location having the determined minimum Frobenius Norm value for each X data and analyzing the Y data and the analysis data. Claim 6 A method for correcting process data time using VGA for process control in a continuous production process, characterized in that, in step (6), the confirmed distortion value of X data is confirmed through step (5), and if it differs from a predefined value, a user alarm is generated to warn, and if the confirmed time lag is greater or smaller than the parameter value defined by the user using the time lag confirmed in the collected process data, it is determined to be an abnormal situation of the process and an alarm is generated. Claim 7 A method for correcting process data time using VGA for process control in a continuous production process, characterized in that, in step (7), the correction value of the time delay of X data is statistically compared with the previous value through step (5) to generate an alarm and issue a warning when an anomaly occurs, and the alarm is generated when the confirmed time delay is greater than 2 sigma (σ) compared to the previously confirmed time delay.

Images