CN116249945A - Process abnormality detection device and method using analysis of control part temperature signal - Google Patents
Process abnormality detection device and method using analysis of control part temperature signal Download PDFInfo
- Publication number
- CN116249945A CN116249945A CN202180057871.5A CN202180057871A CN116249945A CN 116249945 A CN116249945 A CN 116249945A CN 202180057871 A CN202180057871 A CN 202180057871A CN 116249945 A CN116249945 A CN 116249945A
- Authority
- CN
- China
- Prior art keywords
- signal
- units
- temperature signal
- pieces
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 86
- 238000001514 detection method Methods 0.000 title claims abstract description 38
- 230000005856 abnormality Effects 0.000 title claims abstract description 29
- 238000007781 pre-processing Methods 0.000 claims abstract description 24
- 238000010801 machine learning Methods 0.000 claims abstract description 16
- 238000000605 extraction Methods 0.000 claims abstract description 13
- 238000005070 sampling Methods 0.000 claims description 55
- 238000004519 manufacturing process Methods 0.000 claims description 36
- 238000011084 recovery Methods 0.000 claims description 21
- 238000006073 displacement reaction Methods 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000007637 random forest analysis Methods 0.000 claims description 10
- 230000000630 rising effect Effects 0.000 claims description 8
- 238000010586 diagram Methods 0.000 description 14
- 239000000284 extract Substances 0.000 description 5
- 238000003066 decision tree Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000003466 welding Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
- G05B19/05—Programmable logic controllers, e.g. simulating logic interconnections of signals according to ladder diagrams or function charts
- G05B19/058—Safety, monitoring
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0221—Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
- G05B23/0275—Fault isolation and identification, e.g. classify fault; estimate cause or root of failure
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Manufacturing & Machinery (AREA)
- Quality & Reliability (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
A process abnormality detection device using analysis of a control portion temperature signal according to the present invention includes: a preprocessing unit for preprocessing data of an analog temperature signal corresponding to the PLC control process; a feature information extraction unit for extracting a plurality of pieces of feature information from the preprocessed analog temperature signal; and a process abnormality detection unit that applies the extracted pieces of characteristic information to the machine learning model to detect abnormality of the PLC control process corresponding to the simulated temperature signal.
Description
Technical Field
The present invention relates to control of a Programmable Logic Controller (PLC), and more particularly, to a technique for accurately detecting defects and anomalies through analysis of a control part.
Background
A Programmable Logic Controller (PLC) is a computer programmed in a low-level language and used to control an automation system. As internal logic of the PLC, the PLC program controls the automation system through boolean operations. A PLC program designed under a general manufacturing process can be verified, and if accuracy of the program is achieved, the program can be used to actually control an automation system.
In recent automated manufacturing industries, as the complexity of manufacturing lines increases, control logic becomes redundant in design and complex, and thus, the logic of PLC programs also becomes complex. For these reasons, it has become increasingly difficult to diagnose and monitor PLC programs, and accordingly, the time for diagnosing and correcting errors has also gradually increased. According to the so-called "holy-cup" of automated control, i.e. operational diagnostics (Operational Diagnostics), the operational delay due to such misdiagnosis and identification accounts for more than 80% of the total equipment downtime. In particular, in the vehicle body assembly line, the average period is about 1 minute, and thus if the vehicle body assembly line is stopped due to an equipment failure, a large amount of loss may be caused in a short time.
A typical automated manufacturing system may consist of several robots and automated transport devices. The robot and the conveyor perform various tasks, such as welding and transportation, according to the logic of the PLC program. Recent automated manufacturing systems employ large-scale automated manufacturing lines of high complexity, and thus, involve various task failure causes such as errors in their own systems due to complexity and errors due to external causes such as the movement range of robots being interfered with. Delays in the process caused by task failure may result in significant economic losses due to false detection and increased start-up time.
In order to diagnose the errors, a diagnostic code may be added to the PLC program controlling the process and the distribution flow, thereby monitoring the automated manufacturing system. In general, in an automated manufacturing system typified by the automotive industry, when an error occurs in an automotive manufacturing line, a PLC program is monitored by an error display module that shows the predicted error. That is, the conventional monitoring method predicts a high probability region where an error occurs and builds and adds a code for an error diagnosis target object, so that all signals cannot be monitored. Therefore, the conventional method monitors only a limited range of processes, and thus, has a limitation as a monitoring method for detecting progressive errors. In other words, it is difficult to deal with task failure due to progressive degradation of the device or accessory in advance.
In particular, if the system abnormality does not significantly affect the operation of the PLC control system, the abnormality cannot be determined by the control process model, and therefore, there is a problem that all types of errors related to the analog input/output signal cannot be detected. The prior art related to the present invention is disclosed in korean patent No. 10-0414437 (registered in 12/24/2003).
Disclosure of Invention
Technical problem
The present invention relates to a process anomaly detection method and system using analysis of a control section temperature signal, which extracts a plurality of pieces of characteristic information of a Programmable Logic Controller (PLC) analog signal, and applies the extracted plurality of pieces of characteristic information to a machine learning model to detect a process anomaly associated with the PLC analog signal.
Technical proposal
In order to solve the above problems, the present invention provides a process abnormality detection apparatus using temperature signal analysis of a control section, comprising: a preprocessing unit for preprocessing data of an analog temperature signal corresponding to the PLC control process; a feature information extraction unit for extracting a plurality of pieces of feature information from the preprocessed analog temperature signal; and a process abnormality detection unit that applies the extracted pieces of characteristic information to the machine learning model to detect abnormality of the PLC control process corresponding to the simulated temperature signal.
The preprocessing unit may include: a signal dividing module configured to divide the analog temperature signal into segmented units; and a loss recovery module configured to recover a lost portion of the analog temperature signal divided into segmented units.
The signal dividing module may divide the analog temperature signal into units of the production cycle as segmented units.
The signal dividing module may divide the analog temperature signal into units of respective specific processes within the production cycle of the product as segmented units.
The loss recovery module may recover sampling temperature data of a lost portion of the first segment signal or the second segment signal divided in units of segments using a linear relationship between end point temperature data, which is data among a plurality of pieces of sampling temperature data of the first segment signal among segment signals obtained by dividing the analog temperature signal into units of segments, and start point temperature data, which is data among a plurality of pieces of sampling temperature data of the second segment signal adjacent to the first segment signal.
The loss recovery module may divide each of the segmented signals divided in units of segments into predetermined sampling portions and recover the sampling temperature data divided for each of the predetermined sampling portions using a linear relationship between the plurality of pieces of original sampling temperature data.
The characteristic information extraction unit may extract at least one of a total integrated area, an integrated area of an arc segment, a Y-axis translational integrated area, an integrated area of a temperature rising portion, an integrated area of a temperature falling portion, a slope between a start point and an end point, a displacement difference between a start point and a peak, a displacement difference between a peak and an end point, an average value, a standard deviation, a root mean square, or a shape factor in one product production cycle from the analog temperature signal as the characteristic information.
The feature information extraction unit may select some feature information from the extracted pieces of feature information using a machine learning random forest algorithm.
In order to solve the above problems, the present invention provides a process abnormality detection method using analysis of a control portion temperature signal, comprising the steps of: performing data preprocessing on the analog temperature signal corresponding to the PLC control procedure; extracting a plurality of pieces of characteristic information from the preprocessed analog temperature signal; and applying the extracted pieces of characteristic information to a machine learning model to detect an abnormality of the PLC control process corresponding to the simulated temperature signal.
The step of performing data preprocessing may include: dividing the analog temperature signal into segmented units; and recovering a lost portion of the analog temperature signal divided into segmented units.
The step of dividing into segmented units may comprise dividing the analog temperature signal into units of the production cycle of the product as segmented units.
The step of dividing into segmented units may include dividing the analog temperature signal into units of respective specific procedures within the production cycle of the product as segmented units.
The recovering of the lost portion may include recovering sampling temperature data of the lost portion of the first segment signal or the second segment signal divided in the unit of segments using a linear relationship between end point temperature data, which is data among a plurality of pieces of sampling temperature data of the first segment signal belonging to the segment signal obtained by dividing the analog temperature signal into the unit of segments, and start point temperature data, which is data among a plurality of pieces of sampling temperature data of the second segment signal adjacent to the first segment signal.
The recovering of the lost portion may include dividing each of the segmented signals divided in units of segments into predetermined sampling portions, and recovering the sampling temperature data divided for each of the predetermined sampling portions using a linear relationship between the plurality of pieces of original sampling temperature data.
The extracting of the plurality of pieces of characteristic information may include extracting at least one of a total integrated area, an integrated area of an arc segment, a Y-axis translational integrated area, an integrated area of a temperature rising portion, an integrated area of a temperature falling portion, a slope between a start point and an end point, a displacement difference between a start point and a peak, a displacement difference between a peak and an end point, an average value, a standard deviation, a root mean square, or a shape factor in one product production cycle from the analog temperature signal as the characteristic information.
The step of extracting the plurality of pieces of feature information may include selecting some feature information from the plurality of pieces of extracted feature information using a machine learning random forest algorithm.
Advantageous effects
According to the present invention, data preprocessing is performed using an analog temperature signal corresponding to a Programmable Logic Controller (PLC) control process, and a plurality of pieces of characteristic information are extracted from the preprocessed analog temperature signal to detect an abnormality of the PLC control process, thereby easily detecting a process abnormality of a specific control part in the PLC control process.
Drawings
Fig. 1 is a block diagram illustrating an embodiment of a process abnormality detection apparatus for explaining analysis using a control portion temperature signal according to the present invention.
Fig. 2 is a block diagram illustrating an embodiment of a preprocessing unit shown in fig. 1 according to the present invention.
Fig. 3 is a reference diagram illustrating an example of dividing an analog temperature signal into segmented units by a signal dividing module.
Fig. 4 is a reference diagram illustrating another example of dividing an analog temperature signal into segmented units by a signal dividing module.
Fig. 5 is a reference diagram illustrating an example of restoration of a lost portion of a segmented signal by a loss restoration module.
Fig. 6 is a reference diagram illustrating another example of restoration of a lost portion of a segmented signal by a loss restoration module.
Fig. 7 is a reference diagram illustrating pieces of characteristic information that can be extracted from the analog temperature signal by the characteristic information extraction unit.
Fig. 8 is a flowchart illustrating an embodiment of a process anomaly detection method for explaining an analysis using a control portion temperature signal according to the present invention.
Fig. 9 is a flowchart illustrating an embodiment for explaining the step of performing data preprocessing shown in fig. 8.
Detailed Description
Hereinafter, the present invention will be described in detail with reference to the accompanying drawings in order to easily embody and embody the present disclosure by those skilled in the art. However, the present invention is not limited to the examples disclosed below, but may be embodied in various forms. In addition, portions of the drawings that are not related to the specification are omitted to ensure clarity of the present disclosure. Like reference symbols in the drawings indicate like elements.
In general, detection and isolation of defects and anomalies in Programmable Logic Controller (PLC) control systems is difficult. The present invention proposes an automated tool for diagnosing manufacturing process faults, called a manufacturing process fault diagnosis tool, capable of accurately detecting and isolating defects and anomalies of a PLC control system.
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.
Fig. 1 is a block diagram illustrating an embodiment of a process anomaly detection apparatus 100 for explaining analysis using a control segment temperature signal according to the present invention.
Referring to fig. 1, the process anomaly detection apparatus 100 may include a preprocessing unit 110, a feature information extraction unit 120, and a process anomaly detection unit 130.
The preprocessing unit 110 performs data preprocessing on the acquired analog temperature signal corresponding to the PLC control process.
Fig. 2 is a block diagram illustrating an embodiment of the preprocessing unit 110 shown in fig. 1 according to the present invention.
Referring to fig. 2, the preprocessing unit 110 includes a signal dividing module 110-1 and a loss recovery module 110-2.
The signal dividing module 110-1 may divide the analog temperature signal into segmented units.
The signal dividing module 110-1 may divide the analog temperature signal into units of the production cycle of the product as segmented units.
Fig. 3 is a reference diagram illustrating an example of dividing an analog temperature signal into segmented units by the signal dividing module 110-1.
Referring to ((a) of fig. 3, conventionally, a method of dividing a complete analog temperature signal into predetermined time intervals (e.g., 1 minute intervals) without considering a control portion of the analog temperature signal and analyzing the divided signal is generally used, according to which an accurate data pattern change of the analog temperature signal cannot be reflected because of a different pattern of each analysis point.
In addition, it is possible to easily extract a data portion that does not require analysis, that is, a portion where a process abnormality occurs.
In addition, the signal dividing module 110-1 may divide the analog temperature signal into units of respective specific processes within the production cycle of the product as segmented units.
Fig. 4 is a reference diagram illustrating another example of dividing the analog temperature signal into segmented units by the signal dividing module 110-1.
Referring to fig. 4, the signal division module 110-1 may divide the analog temperature signal into units of respective specific processes within the product production cycle based on the PLC I/O log information to divide the analog temperature signal into units of respective specific processes within the product production cycle as a segmented unit. Since the individual specific processes in the product production cycle are divided as the segment units, it is possible to determine which process portion in the product production cycle has an abnormality. For example, the operation of the process a in the production cycle takes 5 seconds on average, but when an abnormality occurs and the operation time takes 7 seconds, it can be determined that the simulated temperature signal of the process a has an abnormality.
The loss recovery module 110-2 recovers the lost portion of the analog temperature signal divided into segmented units.
The loss recovery module 110-2 recovers sampling temperature data of a lost portion of the first segment signal or the second segment signal divided in units of segments using a linear relationship between end point temperature data, which is data among a plurality of pieces of sampling temperature data of the first segment signal belonging to the segment signal obtained by dividing the analog temperature signal into units of segments, and start point temperature data, which is data among a plurality of pieces of sampling temperature data of the second segment signal adjacent to the first segment signal.
Fig. 5 is a reference diagram illustrating an example of recovery of a lost portion of a segmented signal by the loss recovery module 110-2.
Referring to fig. 5, when the analog temperature signal is divided into segmented units, the start time and the end time of the segmented signal do not completely coincide with the start time and the end time of the sampling temperature data of the analog temperature signal. Therefore, since the degree of loss is different for each segment, the data preprocessing can be performed under the same condition, and the values of the start time and the end time of the sampling temperature data are unreliable. The smaller the size of the segmentation unit, the greater the impact of the loss.
To solve this problem, as shown in fig. 5, the loss recovery module 110-2 may select end point data or start point data based on a start time or end time of the segment signal and generate new sampling temperature data at a point where a straight line passing through the end point data and the start point data intersects a reference point of the unit divided into segments.
For example, the sampling temperature data RP1 corresponding to the division point of the first segment signal or the second segment signal divided into the segmented units may be calculated using a linear relational expression of a straight line passing through the end point temperature data EP1 and the start point temperature data SP1, wherein the end point temperature data EP1 is data among the plurality of pieces of sampling temperature data of the first segment signal and the start point temperature data SP1 is data among the plurality of pieces of sampling temperature data of the second segment signal adjacent to the first segment signal. Thus, the loss recovery module 110-2 calculates the sampling temperature data RP1 corresponding to the division point, thereby recovering the analog temperature signal of the lost portion ((e.g., 0.5 second portion) in the second segment signal.
In addition, the sampling temperature data RP2 of the division point corresponding to the second or third division signal (not shown) divided into the divided units may be calculated using a linear relational expression of a straight line passing through the end point temperature data EP2, which is data among the plurality of pieces of sampling temperature data of the second division signal, and the start point temperature data SP2, which is data among the plurality of pieces of sampling temperature data of the third division signal adjacent to the second division signal. Thus, the loss recovery module 110-2 calculates the sampled temperature data RP2 corresponding to the division point, thereby recovering the analog temperature signal of the lost portion (e.g., 1.5 second portion) in the second segment signal.
Further, the loss recovery module 110-2 divides each of the segmented signals divided in units of segments into predetermined sampling portions and recovers the sampling temperature data divided for each of the predetermined sampling portions using a linear relationship between the plurality of pieces of original sampling temperature data.
Fig. 6 is a reference diagram illustrating another example of recovery of a lost portion of a segmented signal by the loss recovery module 110-2.
Referring to fig. 6, when the analog temperature signal is divided based on the segment signals of the same size, the number of points of the sampling temperature data may be different for each segment, and even if the number of points of the sampling temperature data is constant, the acquisition time of the points may not be constant.
To solve the above problem, the loss recovery module 110-2 collects data points at fixed time intervals so that sampling temperature data can be extracted under the same conditions for all segmented signals.
For example, the loss recovery module 110-2 segments the sampled portion into predetermined time units for each segmented signal and calculates a data point value corresponding to each time unit. The loss recovery module 110-2 calculates the sampling temperature data RP10 divided for each predetermined sampling portion by using a linear relational expression of a straight line passing through each of the original sampling temperature data EP10 and EP 11. Thus, the loss recovery module 110-2 may recover an analog temperature signal in which the number of data points and acquisition intervals in all segmented signals are constant.
The characteristic information extraction unit 120 extracts a plurality of pieces of characteristic information from the preprocessed analog temperature signal. The characteristic information extraction unit 120 may extract at least one of a total integrated area, an integrated area of an arc segment, a Y-axis translational integrated area, an integrated area of a temperature rising portion, an integrated area of a temperature falling portion, a slope between a start point and an end point, a displacement difference between a start point and a peak, a displacement difference between a peak and an end point, an average value, a standard deviation, a root mean square, or a shape factor in one product production period from the analog temperature signal as the characteristic information.
Fig. 7 is a reference diagram illustrating pieces of characteristic information that can be extracted from the analog temperature signal by the characteristic information extraction unit 120. As shown in fig. 7, the total integrated area, the integrated area of the arc segment, the Y-axis translational integrated area, the integrated area of the temperature rising portion, the integrated area of the temperature falling portion, the slope between the start point and the end point, the displacement difference between the start point and the peak, the displacement difference between the peak and the end point, the average value, the standard deviation, the root mean square, and the shape factor in one product production cycle can be calculated by the following formulas, respectively.
The total integration area can be calculated using the following equation 1.
[ formula 1]
The integrated area of the arc segment can be calculated using the following equation 2.
[ formula 2]
The Y-axis translational integration area can be calculated using the following equation 3.
[ formula 3]
The integral area of the temperature rising portion can be calculated using the following equation 4.
[ equation 4]
The integral area of the temperature drop portion can be calculated using the following equation 5.
[ equation 5]
The slope between the start point and the end point can be calculated using the following equation 6.
[ formula 6]
The displacement difference between the start point and the peak can be calculated using the following equation 7.
[ formula 7]
Displacement difference between start point and peak=max (y) -y 0
The displacement difference between the peak and the end point can be calculated using the following equation 8.
[ formula 8]
Displacement difference between peak and end point = Max (y) -y n
The average value can be calculated using the following equation 9.
[ formula 9]
The standard deviation can be calculated using the following equation 10.
[ formula 10]
The root mean square can be calculated using the following equation 11.
[ formula 11]
The form factor can be calculated using the following equation 12.
[ formula 12]
The feature information extraction unit 120 may select some feature information from among the calculated pieces of feature information using a machine learning random forest algorithm.
Random forest algorithms refer to a method of generating a plurality of decision trees by randomly selecting elements for creating the decision trees and combining the plurality of decision trees to generate a model. By selecting some of all the feature information using a random forest algorithm, more accurate classification of the feature information can be achieved than regression analysis. The feature information extractor 120 uses a random forest algorithm and lists the ranks of the pieces of feature information according to the importance of the features, and selects the feature information according to the ranks to accurately classify the abnormality of the process. Here, the importance of a feature is a measure for evaluating which variable is most important in the process of creating a decision tree.
The process anomaly detection unit 130 applies the selected pieces of characteristic information to the machine learning model to detect anomalies in the PLC control process corresponding to the simulated temperature signal. The machine learning model is a model trained in advance using training data, and includes an input layer, a hidden layer, and an output layer, which can be used as a classifier to determine anomalies in the PLC control process. The machine learning model utilized in the process anomaly detection unit 130 is a generic model, and a detailed description thereof will not be provided here. However, the process abnormality detection unit 130 has the following structure: the selected pieces of characteristic information are re-trimmed and applied to the input layer, pass through the hidden layer learned through training, and the resultant value can be output through the output layer in which abnormality is determined.
Table 1 below is a table for explaining the determination of process abnormality by applying the selected feature information to the process abnormality detection unit 130.
TABLE 1
Referring to table 1, the following examples are shown: 3 pieces of characteristic information (i.e., (b) an integrated area of an arc segment, (f) a slope between a start point and an end point, and ((k) root mean square) are applied to the process anomaly detection unit 130, which are selected from 12 pieces of characteristic information using a random forest algorithm, to detect anomalies (no anomaly: OK, anomaly: NG) of a corresponding process for each product production cycle.
Fig. 8 is a flowchart illustrating an embodiment of a process anomaly detection method for explaining an analysis using a control portion temperature signal according to the present invention.
The analog temperature signal corresponding to the PLC control process is subjected to data preprocessing (step S1000).
Fig. 9 is a flowchart illustrating an embodiment for explaining the step of performing data preprocessing shown in fig. 8 according to the present invention.
First, the analog temperature signal is divided into segmented units (step S1010).
In the step of dividing into the segmented units, the analog temperature signal may be divided into units of the production cycle of the product as the segmented units. In addition, in the step of dividing into the segmented units, the analog temperature signal may be divided into units of respective specific processes within the production cycle of the product as the segmented units.
After step S1010, the lost portion of the analog temperature signal divided into the segmented units is restored ((step S1012).
In the recovering of the lost portion, a linear relationship between end point temperature data, which is data among a plurality of pieces of sample temperature data of a first segment signal belonging to a segment signal obtained by dividing an analog temperature signal into segmented units, and start point temperature data, which is data among a plurality of pieces of sample temperature data of a second segment signal adjacent to the first segment signal, may be used to recover the sample temperature data of the lost portion of the first segment signal or the second segment signal divided in segments.
In addition, in the step of recovering the lost portion, it is possible to divide the respective segment signals divided in units of segments into predetermined sampling portions and recover the sampling temperature data divided for each predetermined sampling portion using a linear relationship between a plurality of pieces of original sampling temperature data.
After step S1000, pieces of characteristic information are extracted from the preprocessed data ((step S1002).
In the extracting of the plurality of pieces of characteristic information, at least one of a total integrated area, an integrated area of an arc segment, a Y-axis translational integrated area, an integrated area of a temperature rising portion, an integrated area of a temperature falling portion, a slope between a start point and an end point, a displacement difference between a start point and a peak, a displacement difference between a peak and an end point, an average value, a standard deviation, a root mean square, or a shape factor in one production cycle of the product may be extracted from the analog temperature signal as the characteristic information. In the step of extracting the feature information, some feature information may be selected from the extracted pieces of feature information using a machine learning random forest algorithm.
After step S1002, the selected pieces of characteristic information may be applied to a machine learning model to detect an abnormality of the PLC control process corresponding to the simulated temperature signal (step S1004). The selected pieces of feature information may be re-trimmed and applied to the input layer, pass through the hidden layer learned through training, and output a result value regarding the anomaly through the output layer.
The present invention can be implemented in the form of a software program and recorded in a predetermined computer-readable recording medium. For example, the recording medium may be an embedded type hard disk, flash memory, RAM, ROM, or the like of each reproduction apparatus, or may be an external type optical disk such as CD-R and CD-RW, a compact flash card, a smart media, a memory stick, and a multimedia card of each reproduction apparatus.
Although the embodiments of the present invention have been described above, the embodiments disclosed herein are not intended to limit the present invention. The scope of the invention should be construed by the appended claims, and all techniques within the scope of equivalents thereof should be construed as being included in the scope of the invention.
Claims (16)
1. A process abnormality detection device for analyzing a temperature signal of a control section includes:
a preprocessing unit for performing data preprocessing on an analog temperature signal corresponding to a Programmable Logic Controller (PLC) control process;
a feature information extraction unit for extracting a plurality of pieces of feature information from the analog temperature signal subjected to the preprocessing; and
and a process abnormality detection unit that applies the extracted pieces of characteristic information to a machine learning model to detect an abnormality of the PLC control process corresponding to the simulated temperature signal.
2. The process anomaly detection apparatus according to claim 1, wherein the preprocessing unit includes:
a signal dividing module configured to divide the analog temperature signal into segmented units; and
a loss recovery module configured to recover a lost portion of the analog temperature signal divided into segmented units.
3. The process anomaly detection apparatus of claim 2, wherein the signal dividing module is configured to divide the analog temperature signal into units of a product production cycle as the segmented units.
4. The process anomaly detection apparatus of claim 2, wherein the signal dividing module is configured to divide the analog temperature signal into units of respective specific processes within a production cycle of a product as the segmented units.
5. The process anomaly detection apparatus according to claim 2, wherein the loss recovery module is configured to recover sampling temperature data of a lost portion of a first segment signal or a second segment signal divided in units of segments using a linear relationship between end point temperature data, which is data among pieces of sampling temperature data of the first segment signal among segment signals obtained by dividing the analog temperature signal into units of segments, and start point temperature data, which is data among pieces of sampling temperature data of the second segment signal adjacent to the first segment signal.
6. The process anomaly detection apparatus according to claim 5, wherein the loss recovery module is configured to divide each of the segment signals divided in segments into predetermined sampling portions, and recover the sampling temperature data divided for each of the predetermined sampling portions using a linear relationship between a plurality of pieces of raw sampling temperature data.
7. The process anomaly detection apparatus according to claim 1, wherein the characteristic information extraction unit is configured to extract at least one of a total integrated area, an integrated area of an arc segment, a Y-axis translational integrated area, an integrated area of a temperature rising portion, an integrated area of a temperature falling portion, a slope between a start point and an end point, a displacement difference between a start point and a peak, a displacement difference between a peak and an end point, an average value, a standard deviation, a root mean square, or a form factor in one production cycle of the product from the analog temperature signal as the characteristic information.
8. The process anomaly detection apparatus according to claim 7, wherein the feature information extraction unit is configured to select some feature information from the extracted pieces of feature information using a machine learning random forest algorithm.
9. A process anomaly detection method using analysis of a control portion temperature signal, comprising the steps of:
performing data preprocessing on the analog temperature signal corresponding to the PLC control procedure;
extracting a plurality of pieces of characteristic information from the preprocessed analog temperature signal; and
and applying the extracted pieces of characteristic information to a machine learning model to detect an abnormality of the PLC control process corresponding to the simulated temperature signal.
10. The process anomaly detection method according to claim 9, wherein the step of performing data preprocessing includes:
dividing the analog temperature signal into segmented units; and
recovering a lost portion of the analog temperature signal divided into segmented units.
11. The process anomaly detection method of claim 10, wherein the step of dividing into segmented units comprises:
the analog temperature signal is divided into units of a product production cycle as the segmented units.
12. The process anomaly detection method of claim 10, wherein the step of dividing into segmented units comprises:
the analog temperature signal is divided into units of each specific process within the production cycle of the product as the segmented units.
13. The process anomaly detection method according to claim 10, wherein the step of recovering the lost portion comprises:
the method includes recovering sampling temperature data of a lost portion of a first segment signal or a second segment signal divided in units of segments using a linear relationship between end point temperature data, which is data among a plurality of pieces of sampling temperature data of a first segment signal belonging to a segment signal obtained by dividing the analog temperature signal into units of segments, and start point temperature data, which is data among a plurality of pieces of sampling temperature data of the second segment signal adjacent to the first segment signal.
14. The process anomaly detection method according to claim 13, wherein the step of recovering the lost portion comprises:
dividing each of the segment signals divided in units of segments into predetermined sampling portions, and restoring sampling temperature data divided for each of the predetermined sampling portions using a linear relationship between a plurality of pieces of original sampling temperature data.
15. The process anomaly detection method according to claim 9, wherein the step of extracting a plurality of pieces of characteristic information includes:
and extracting at least one of a total integral area, an integral area of an arc section, a Y-axis translation integral area, an integral area of a temperature rising part, an integral area of a temperature falling part, a slope between a starting point and an ending point, a displacement difference between the starting point and the peak, a displacement difference between the peak and the ending point, an average value, a standard deviation, a root mean square or a shape factor in one product production period from the analog temperature signal as the characteristic information.
16. The process anomaly detection method according to claim 15, wherein the step of extracting a plurality of pieces of characteristic information includes:
and selecting some characteristic information from the extracted plurality of characteristic information by using a machine learning random forest algorithm.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR1020200096717A KR102220138B1 (en) | 2020-08-03 | 2020-08-03 | Apparatus and method for detecting process abnormality according to analysis of control section temperature signal |
| KR10-2020-0096717 | 2020-08-03 | ||
| PCT/KR2021/009532 WO2022030831A1 (en) | 2020-08-03 | 2021-07-23 | Process abnormality detection device and method using analysis of control section temperature signal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116249945A true CN116249945A (en) | 2023-06-09 |
Family
ID=74687792
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202180057871.5A Pending CN116249945A (en) | 2020-08-03 | 2021-07-23 | Process abnormality detection device and method using analysis of control part temperature signal |
Country Status (3)
| Country | Link |
|---|---|
| KR (1) | KR102220138B1 (en) |
| CN (1) | CN116249945A (en) |
| WO (1) | WO2022030831A1 (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102220138B1 (en) * | 2020-08-03 | 2021-02-26 | 주식회사 유디엠텍 | Apparatus and method for detecting process abnormality according to analysis of control section temperature signal |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100414437B1 (en) | 2001-04-06 | 2004-01-07 | 전영민 | Factory automation control method |
| WO2015014390A1 (en) * | 2013-07-30 | 2015-02-05 | Dmg Mori Seiki Co., Ltd. | Control system for controlling operation of a numerically controlled machine tool, and back-end and front-end control devices for use in such system |
| EP3335083B1 (en) * | 2015-08-11 | 2024-02-14 | Siemens Aktiengesellschaft | Rich contextualization of automation data |
| KR101865666B1 (en) * | 2016-07-22 | 2018-07-13 | 두산중공업 주식회사 | Apparatus and System for Simulating Core Protection system software |
| US10877465B2 (en) * | 2016-10-24 | 2020-12-29 | Fisher-Rosemount Systems, Inc. | Process device condition and performance monitoring |
| KR102221473B1 (en) * | 2018-12-17 | 2021-03-02 | 주식회사 유디엠텍 | System for error detection according to plc control |
| KR102220138B1 (en) * | 2020-08-03 | 2021-02-26 | 주식회사 유디엠텍 | Apparatus and method for detecting process abnormality according to analysis of control section temperature signal |
-
2020
- 2020-08-03 KR KR1020200096717A patent/KR102220138B1/en active IP Right Grant
-
2021
- 2021-07-23 WO PCT/KR2021/009532 patent/WO2022030831A1/en active Application Filing
- 2021-07-23 CN CN202180057871.5A patent/CN116249945A/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| WO2022030831A1 (en) | 2022-02-10 |
| KR102220138B1 (en) | 2021-02-26 |
| KR102220138B9 (en) | 2023-05-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102221473B1 (en) | System for error detection according to plc control | |
| EP2015186B1 (en) | Diagnostic systems and methods for predictive condition monitoring | |
| EP3105644B1 (en) | Method of identifying anomalies | |
| US6970804B2 (en) | Automated self-learning diagnostic system | |
| US20200198128A1 (en) | Learning data confirmation support device, machine learning device, and failure predicting device | |
| US11287796B2 (en) | Diagnostic device and method for monitoring a technical plan | |
| JP2004150440A (en) | Method for performing diagnosis on system | |
| US11816935B2 (en) | Predicting a repair and maintenance activity for an aircraft system | |
| CN109141625B (en) | Online state monitoring method for ball screw pair | |
| JP7012888B2 (en) | Abnormal factor estimation device, abnormal factor estimation method, and program | |
| CN116249945A (en) | Process abnormality detection device and method using analysis of control part temperature signal | |
| US6138109A (en) | Neural network diagnostic classification of complex binary systems | |
| US20090248363A1 (en) | Method of multi-level fault isolation design | |
| Eger et al. | Knowledge capturing platform in multi-stage production systems for zero-defect manufacturing | |
| CN113210824B (en) | Servo welding gun driving abnormity detection method and equipment | |
| US7814034B2 (en) | Method and system for automatically developing a fault classification system by segregation of kernels in time series data | |
| Madden et al. | Monitoring and diagnosis of multiple incipient faults using fault tree induction | |
| CN110580700A (en) | Information processing method, information processing apparatus, and program | |
| US11676055B2 (en) | System for detecting data drift in machine-learning process monitoring | |
| JP2020166407A (en) | Model generation device, abnormality occurrence prediction device, abnormality occurrence prediction model generation method and abnormality occurrence prediction method | |
| KR102265461B1 (en) | Abnormal Data Detection System in Manufacturing Process | |
| Kampker et al. | Inline Testing Method Based On Paschen’s Law With Integrated Machine Learning For Hairpin Stator Production | |
| US20050075835A1 (en) | System and method of real-time statistical bin control | |
| KR102315228B1 (en) | Apparatus and method System managing data for controlling of Programmable Logic Controller | |
| KR102414537B1 (en) | Device for tracking defect location based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |