WO2020054850A1 - Process monitoring assistance device, process monitoring assistance system, process monitoring assistance method, process monitoring assistance program, and terminal device - Google Patents

Process monitoring assistance device, process monitoring assistance system, process monitoring assistance method, process monitoring assistance program, and terminal device Download PDF

Info

Publication number
WO2020054850A1
WO2020054850A1 PCT/JP2019/036119 JP2019036119W WO2020054850A1 WO 2020054850 A1 WO2020054850 A1 WO 2020054850A1 JP 2019036119 W JP2019036119 W JP 2019036119W WO 2020054850 A1 WO2020054850 A1 WO 2020054850A1
Authority
WO
WIPO (PCT)
Prior art keywords
index
data
contribution
unit
indicating
Prior art date
Application number
PCT/JP2019/036119
Other languages
French (fr)
Japanese (ja)
Inventor
理 山中
諒 難波
卓巳 小原
由紀夫 平岡
Original Assignee
株式会社東芝
東芝インフラシステムズ株式会社
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 株式会社東芝, 東芝インフラシステムズ株式会社 filed Critical 株式会社東芝
Priority to MX2021002926A priority Critical patent/MX2021002926A/en
Publication of WO2020054850A1 publication Critical patent/WO2020054850A1/en
Priority to PH12021550545A priority patent/PH12021550545A1/en

Links

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
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring

Definitions

  • the embodiments of the present invention relate to a process monitoring support device, a process monitoring support system, a process monitoring support method, a process monitoring support program, and a terminal device.
  • the process monitoring support device includes a data acquisition unit, an index calculation unit, a contribution ratio calculation unit, and a display control unit.
  • the data acquisition unit acquires a plurality of types of time-series data of process variables indicating the state of the process to be monitored.
  • the index calculation unit calculates an index indicating a possibility that the monitoring target process is healthy without any abnormality based on the plurality of types of the process variables.
  • Contribution rate calculation unit for the index calculated by the index calculation unit, the contribution rate indicating the rate of contribution to increase the possibility that there is no abnormality and is healthy, for each of the plurality of types of process variables Is calculated.
  • the display control unit generates display information so as to display information on some process variables having a relatively small value of the contribution ratio on a screen of the user terminal.
  • FIG. 1 is a diagram showing a specific example of the configuration of the process monitoring support device 2 of the embodiment.
  • FIG. 1 shows a specific example in which the monitoring target of the process monitoring support device 2 is the advanced sewage treatment process 1.
  • Sewage advanced treatment process 1 is a process aimed at removing nitrogen and phosphorus from sewage.
  • the advanced sewage treatment process 1 has a sedimentation basin 101, an anaerobic tank 102, an anoxic tank 103, an aerobic tank 104 and a final sedimentation basin 105.
  • the sewage to be treated (hereinafter referred to as “treatment water”) is first sent to the sedimentation basin 101, the anaerobic tank 102, the anoxic tank 103, the aerobic tank 104, and the final sedimentation basin 105 in this order.
  • the water to be treated returned from the aerobic tank 104 at the subsequent stage is mixed with the water to be treated sent from the anaerobic tank 102, and is agitated without supplying air.
  • the nitric acid in the water to be treated is decomposed into nitrogen by the action of microorganisms and released to the atmosphere. Generally, this treatment is called denitrification.
  • the aerobic tank 104 is a tank for decomposing organic substances in the water to be treated, removing phosphorus, and nitrifying ammonia. Specifically, air is supplied to the water to be treated to activate the microorganisms, and the microorganisms decompose organic substances, and the microorganisms absorb phosphorus in the water to be treated. Microorganisms that discharge phosphorus in an anaerobic state and instead accumulate organic matter absorb the phosphorus that is exhaled by being activated, so that the phosphorus in the water to be treated is removed. In the aerobic tank 104, ammonia is decomposed into nitric acid by supplying air to the water to be treated. This treatment is generally called nitrification.
  • the final sedimentation basin 105 is a reservoir for treated water from which phosphorus has been removed and ammonia has been nitrified.
  • solids remaining in the water to be treated are separated by sedimentation, and the supernatant clear water is discharged as treated water.
  • the first settling basin excess sludge extraction pump 111 is a pump for extracting and removing sludge settled from the first settling basin 101.
  • the first settling basin excess sludge extraction pump 111 has a flow rate sensor for measuring the flow rate of the extracted sludge.
  • the blower 112 is a blower that supplies oxygen to the aerobic tank 104.
  • the blower 112 has a flow sensor that measures the flow rate of the supplied air.
  • the circulation pump 113 is a pump that returns the water to be treated from the aerobic tank 104 to the anoxic tank 103.
  • the circulation pump 113 has a flow rate sensor that measures the flow rate of the returned treated water.
  • the rainfall sensor 121 is a sensor that measures the rainfall near the sewage altitude treatment process 1.
  • the sewage inflow sensor 122 is a sensor that measures the flow rate of sewage flowing into the sewage advanced treatment process 1 (hereinafter, referred to as “inflow sewage”).
  • the inflow TN sensor 123 is a sensor that measures the total amount of nitrogen (TN) contained in the inflow sewage.
  • the inflow TP sensor 124 is a sensor that measures the total amount of phosphorus (TP) contained in the inflow sewage.
  • the inflowing organic matter sensor 125 is a UV (absorbance) sensor or a COD (chemical oxygen demand) sensor that measures the amount of organic matter contained in the inflowing sewage.
  • the ORP sensor 126 is a sensor that measures the ORP (oxidation-reduction potential) of the anaerobic tank 102.
  • Anaerobic tank pH sensor 127 is a sensor that measures the pH of anaerobic tank 102.
  • the anoxic tank ORP sensor 128 is a sensor that measures the ORP of the anoxic tank 103.
  • the anoxic tank pH sensor 129 is a sensor that measures the pH of the anoxic tank 103.
  • the phosphoric acid sensor 130 is a sensor that measures the concentration of phosphoric acid in the aerobic tank 104.
  • the DO sensor 131 is a sensor that measures the dissolved oxygen concentration (DO) in the aerobic tank 104.
  • the ammonia sensor 132 is a sensor that measures the ammonia concentration in the aerobic tank 104.
  • the MLSS sensor 133 is a sensor that measures the activated sludge concentration (MLSS) in at least one of the anaerobic tank 102, the anaerobic tank 103, and the aerobic tank 104.
  • the water temperature sensor 134 is a sensor that measures the water temperature in at least one location of the anoxic tank 103 or the aerobic tank 104.
  • the surplus sludge SS sensor 135 is a sensor that measures the solid (SS) concentration of sludge extracted from the final sedimentation basin 105.
  • the discharge SS sensor 136 is a sensor that measures the SS concentration of the water discharged from the final sedimentation basin 105.
  • the sludge interface sensor 137 is a sensor that measures the sludge interface level of the final settling basin 105.
  • the sewage discharge flow sensor 138 is a sensor that measures the flow rate of discharged water.
  • the discharge TN sensor 139 is a sensor that measures the total amount of nitrogen contained in the discharge water.
  • the discharge TP sensor 140 is a sensor that measures the total amount of phosphorus contained in the discharge water.
  • the discharged organic substance sensor 141 is a UV sensor or a COD sensor that measures the amount of organic substances contained in the discharged water.
  • Each of the above-mentioned devices such as the first settling basin excess sludge pulling pump 111, the blower 112, the circulation pump 113, the return sludge pump 114, and the final settling basin excess sludge pulling pump 115 operates under the control of a predetermined cycle.
  • the above-described sensors including the flow rate sensors included in the equipment of the initial sedimentation tank excess sludge extraction pump 111, the blower 112, the circulation pump 113, the return sludge pump 114, and the final sedimentation tank excess sludge extraction pump 115 are sensed at predetermined intervals. Measure the target.
  • the flow rate sensors included in the first settling tank excess sludge extraction pump 111, the blower 112, the circulation pump 113, the return sludge pump 114, and the final settling tank excess sludge extraction pump 115 are collectively referred to as operation amount sensors, and other sensors are referred to.
  • process sensors are collectively referred to as process sensors.
  • Each of the operation amount sensors and each of the process sensors transmits measurement data obtained by sensing in a predetermined cycle to the process monitoring support device 2 as process data.
  • the process monitoring support device 2 includes a CPU (Central Processing Unit), a memory, and an auxiliary storage device connected by a bus, and executes a monitoring support program.
  • the process monitoring support device 2 functions as a device including a data collection unit 201, a data storage unit 202, a state definition unit 21, a state calculation unit 22, and a display control unit 23 by executing a monitoring support program.
  • all or a part of each function of the process monitoring support device 2 may be realized using hardware such as an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array). Good.
  • ASIC Application Specific Integrated Circuit
  • PLD Programmable Logic Device
  • FPGA Field Programmable Gate Array
  • the data collection unit 201 may be implemented as a device having a housing different from that of the process monitoring support device 2 using a PLC (Programmable Logic Controller).
  • the monitoring support program may be recorded on a computer-readable recording medium.
  • the computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in a computer system.
  • the monitoring support program may be transmitted via a telecommunication line.
  • the data collection unit 201 acquires process data from each operation amount sensor and each process sensor.
  • the acquired process data is time-series data of each process variable indicating the state of the monitoring target process.
  • the data collection unit 201 records the obtained process data in the data storage unit 202 according to a predetermined format.
  • the data storage unit 202 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device.
  • the data storage unit 202 stores the process data obtained by the data collection unit 201.
  • the state definition unit 21 defines data relating to the health of the process to be monitored.
  • the state definition unit 21 reads past time-series data of process variables recorded in the data storage unit 202.
  • the state definition unit 21 includes an index (hereinafter, referred to as a “process health index”) used to determine the health of the process, and a contribution ratio (hereinafter, referred to as “process health index”) indicating the degree of influence of each process variable on the process health index. Process soundness contribution ratio.)).
  • the process health index is an index indicating a possibility that an abnormality has occurred in the state of the monitored process.
  • the process soundness contribution ratio is a value indicating a ratio of each process variable contributing to an increase in the possibility that an abnormality has occurred in the process soundness index.
  • the state definition unit 21 generates typical representative values and representative pattern data (hereinafter, referred to as “normal state patterns”) in the past normal state of the process.
  • the state definition unit 21 operates, for example, at a predetermined cycle T0.
  • the state calculation unit 22 reads the current time-series data at the time of monitoring the process variables recorded in the data storage unit 202.
  • the state calculation unit 22 calculates a process soundness index and a process soundness contribution ratio of each process variable using the definition in the state definition unit 21.
  • the state calculation unit 22 assigns a high importance to each process variable in, for example, an ascending order of the process soundness contribution rate (an ascending order of abnormality).
  • the state calculation unit 22 extracts a predetermined number of upper process variables in descending order of importance at the time of monitoring.
  • the state calculation unit 22 operates, for example, in a predetermined cycle T1 ( ⁇ T0).
  • the display control unit 23 generates an image indicating the time-series data of a predetermined number of higher-order process variables in descending order of importance based on the information obtained by the state calculation unit 22. Such an image may be generated, for example, as a trend graph.
  • the display control unit 23 generates data (display information) for displaying the time series data of the process health index and the time series data of each process variable with high importance, and outputs the data to the user terminal 3.
  • the state definition unit 21 functions as a past data acquisition unit 211, an index definition unit 212, a contribution ratio definition unit 213, and a normal state pattern generation unit 214.
  • the past data acquisition unit 211 reads out past data (offline data) that is time-series data of a predetermined period (hereinafter, referred to as “predetermined period”) of each process variable from the data storage unit 202.
  • predetermined period a predetermined period of each process variable
  • the user may be able to define a past predetermined period by operating the user terminal 3. It is preferable that such a definition can be input by a GUI on the monitoring screen.
  • the plant state often changes gradually due to a change in plant operation or a seasonal change. Therefore, it is often more realistic to update the past predetermined period defining the process health index.
  • the length of the past predetermined period is set in advance, and the predetermined past period may be updated according to the length of the past predetermined period in a predetermined cycle.
  • the past data acquisition unit 211 reads from the data storage unit 202 the past data for the past predetermined period defined at the time of data acquisition. For example, assuming that the length of the past predetermined period is one week, data for the past one week may be read as data of the past predetermined period every week.
  • the past time-series data acquired by the past data acquisition unit 211 in this manner is described as X.
  • the time-series data X is a matrix having process variables in the row direction and time-series data over a predetermined past period in the column direction. In the following description, the number of process variables is n and the number of time-series data is m. Therefore, the time series data X is m ⁇ n time series data.
  • the index definition unit 212 uses the past time-series data X read by the past data acquisition unit 211 to define a process health index.
  • the index definition unit 212 may define the process health index by using a multivariate analysis or a machine learning technique.
  • the method of defining the process health index may be implemented in any manner. Since the process soundness index is generated from the past time-series data X, it includes information on n process variables, and the indicators for measuring the soundness of the process are collected into one index.
  • ⁇ The following are specific examples of process integrity indicators.
  • Q-statistic used in a method called MSPC (Multivariate Statistical Process Management), which is an advanced process monitoring and diagnostic technique.
  • Data for detecting abnormalities called Hotelling's T ⁇ 2 statistic.
  • Mahalanobis distance used in the Taguchi method (a technique similar to MSPC) used in the field of quality engineering.
  • the Mahalanobis distance is essentially equivalent to the T ⁇ 2 statistic used in MSPC.
  • the T ⁇ 2 statistic is defined after performing dimension reduction (dimension of n ′ ⁇ n), but in the method using the Mahalanobis distance, the distance is defined in an n-dimensional space. You. When the dimensions are aligned, the Mahalanobis distance and the T ⁇ 2 statistic are essentially the same, and there is only a difference of a constant multiple.
  • MSPC is used as a specific example of the process soundness index.
  • PCA principal component analysis
  • T_a is an m ⁇ n matrix based on the number m of samples and the number n of principal components, and is called a score matrix.
  • P_a is an n ⁇ n matrix indicating a relationship between n process variables and n principal components, and is called a loading matrix.
  • T is a sub-matrix of T_a in which the principal components are truncated by p ( ⁇ n), and this T is generally called a score matrix.
  • P is a sub-matrix (n ⁇ p) of P_a in which the main component is truncated by p, and this P is generally called a loading matrix.
  • E is an error matrix (m ⁇ n) based on the number m of samples and the number n of process variables, and represents an error when the number of principal components is cut off by p.
  • T_a is referred to as a score matrix, and T is referred to as a main score matrix.
  • P_a is referred to as a loading matrix, and P is referred to as a main loading matrix. If these matrices are used, the Q statistic Q (x (t)) and Hotelling's T ⁇ 2 statistic T ⁇ 2 (x (t)) are defined as the following equations (2) and (3).
  • x (t) represents the t-th element of the past time-series data X.
  • I is a unit matrix of an appropriate size.
  • ⁇ is a matrix having the variance of the principal component as a diagonal element, which means normalization of the variance. At the time of online monitoring to be described later, this x (t) is calculated by replacing the process data measured online.
  • either one of the Q statistic or Hotelling's T ⁇ 2 statistic may be defined as the process health index.
  • the Q statistic tends to show (detect) minor fluctuations. Such minor fluctuations include abnormal signs that are difficult for the user to notice only by looking at one process variable. Therefore, when the purpose is to monitor such a small change, it is preferable to use the Q statistic as the process soundness index.
  • Hotelling's T ⁇ 2 statistic tends to detect relatively large variations in each process variable. Therefore, when it is desired to monitor a clear change in the process, it is preferable to use Hotelling's T ⁇ 2 statistic as the process soundness index. Also, any fluctuations, including signs of process anomalies, are detected in either the Q statistic or the Hotelling T ⁇ 2 statistic.
  • the one having a larger value of the Q statistic and the Hotelling's T ⁇ 2 statistic may be used as the process soundness index.
  • the sum of the Q statistic and the Hotelling's T ⁇ 2 statistic (that is, a comprehensive abnormality index) may be used as the process soundness index.
  • the degree of soundness may be defined in a range of 0 to 1, wherein "1" indicates that the soundness is completely sound, and "0" may be defined so as to indicate that the soundness is completely abnormal.
  • a value obtained by converting the Q statistic and the T ⁇ 2 statistic by a conversion formula such as exp ( ⁇ a ⁇ statistic) may be defined as the soundness.
  • “a” is an adjustment parameter and is a value larger than 0.
  • the contribution ratio definition unit 213 defines a process soundness contribution ratio of each process variable with respect to the index defined by the index definition unit 212.
  • the contribution to each statistic may be defined as the process soundness contribution.
  • the contribution of the Q statistic and the contribution of the Hotelling's T ⁇ 2 statistic are defined as in the following Expressions (4) and (5), respectively.
  • Equation (4) is an equation representing the projection of the Q statistic on the axis of the nth process variable, and it is possible to calculate how much each process variable contributes to the Q statistic by Equation (5).
  • Equation (5) is an equation for decomposing the T ⁇ 2 statistic into the sum of n process variable components. Using equation (5), it is possible to calculate how much each process variable contributes to the T ⁇ 2 statistic.
  • the process soundness index If the larger of the Q statistic and Hotelling's T ⁇ 2 statistic is used as the process soundness index, the corresponding contribution to the larger statistic may be used as the process soundness contribution rate. .
  • the sum of the Q statistic and the Hotelling's T ⁇ 2 statistic is used as the process soundness index, the sum of the respective values obtained by Expressions (4) and (5) is the process soundness. It may be used as a contribution rate.
  • the normal state pattern generation unit 214 generates a normal state pattern by using the past data used by the index definition unit 212 and the contribution ratio definition unit 213 to define the process soundness index and the process soundness contribution ratio.
  • the normal pattern is information indicating a typical value or a typical pattern of each process variable.
  • the normal pattern generation unit 214 generates a normal pattern of each process variable using the past time-series data X. For example, the normal pattern generation unit 214 may calculate a position parameter of each time-series data of each process variable of the past time-series data X (that is, each column vector of X) as a normal pattern.
  • the position parameter is, for example, a value such as an average value, a median value, and a pruning average value.
  • a normal pattern may be obtained by such a simple method.
  • the normal pattern generation unit 214 may further calculate, as a part of the value of the normal pattern, a scale parameter indicating the degree of variation around the typical value obtained as the normal pattern.
  • the scale parameter is, for example, a value such as a standard deviation, a median absolute deviation (MAD: Median Absolute Deviation), or a quadrant.
  • the normal state pattern generation unit 214 may generate typical pattern data of each process data instead of a mere position parameter.
  • a pattern having a periodicity on a daily basis according to a human lifestyle is generally recognized.
  • holidays such as Saturday and Sunday, the behavior often differs from that on weekdays.
  • a pattern having a weekly periodicity is recognized.
  • the data for which daily and weekly patterns are recognized include not only sewage treatment processes but also water demand patterns for water purification, energy demand patterns for energy plants, traffic patterns for automobiles and other traffic, and so on. This is particularly noticeable in infrastructure plants that are closely related to human lifestyles.
  • the normal state pattern generation unit 214 calculates the position parameter in, for example, a predetermined time unit (for example, 1 minute unit, 1 hour unit), and calculates the calculated value. May be generated as a normal pattern by connecting time series data for a predetermined period (for example, one day, one week). In such a case, the scale parameter may be calculated as a part of the normal pattern.
  • the scale parameter in a predetermined time unit may be calculated in the same manner as when the normal pattern is generated. Further, the scale parameter for each column vector of the past time-series data X may be calculated uniformly.
  • the state calculation unit 22 functions as a current data acquisition unit 221, an index calculation unit 222, a contribution ratio calculation unit 223, and an important variable extraction unit 224.
  • the current data acquisition unit 221 reads the current data (online data) of each process variable from the data storage unit 202.
  • the current data is the data of each process variable at the time of monitoring (hereinafter referred to as “monitoring time”).
  • the index operation unit 222 performs an operation using the current data read by the current data acquisition unit 221 and the definition formula of the process health index defined by the index definition unit 212.
  • the index calculation unit 222 acquires an index indicating the current degree of health of the process as the calculation result.
  • a specific example of the process of the index calculation unit 222 will be described.
  • the index calculation unit 222 performs outlier (outlier) processing on the data at the time of monitoring read by the current data acquisition unit 221 as necessary. Then, the index calculation unit 222 calculates the process health index defined by the index definition unit 212 using the data on which the outlier processing has been performed.
  • the index calculating unit 222 calculates the Q statistic by inputting the data at the monitoring time into X (t) of Expression (2). Calculate the process soundness index.
  • the index calculation unit 222 inputs the data at the monitoring time point into X (t) of the T ⁇ 2 statistic defined by Expression (3). By calculating the T ⁇ 2 statistic, the process soundness index is calculated.
  • the current process health index can be calculated by performing the calculation using the equations (2) and (3).
  • the important variable extraction unit 224 sorts the process variables in the order of low process soundness contribution ratio (high order of abnormalities).
  • the important variable extracting unit 224 extracts a predetermined number of upper process variables set in advance based on the sorting result.
  • the important variable extracting unit 224 assigns a larger value to each process variable in descending order of the process soundness contribution ratio.
  • the important variable extraction unit 224 extracts a predetermined number (M) of process variables in descending order of importance.
  • the time-series monitoring data of the M process variables is displayed on a screen (a screen of the user terminal) used for monitoring by the user as described later. Therefore, the value of M may normally be set as a value of up to about ten.
  • the display control unit 23 functions as the state display control unit 231 and the monitoring data display control unit 232.
  • the state display control unit 231 generates data for displaying time-series data of the process health index calculated by the index calculation unit 222 for a predetermined period (from a predetermined past to the present: hereinafter, referred to as a “target period”). Generate.
  • the monitoring data display control unit 232 generates data for displaying the process monitoring data (time-series data) of each process variable extracted by the important variable extracting unit 224 during the target period.
  • the state display control unit 231 generates data for displaying time-series data of the process health index calculated in real time by the index calculation unit 222, as shown in the upper part 41 of FIG.
  • the time-series data of the process soundness index may be displayed as, for example, time-series data retroactive to a predetermined period in the past from the present. In the example of FIG. 2, the time-series data for the past 24 hours is displayed, but this display period may be changed according to the operation of the user.
  • the time-series data indicated by a dashed line indicates the normal pattern generated by the normal pattern generation unit 214.
  • the normal pattern may be displayed as one horizontal bar indicating the value in the graph of the time-series data.
  • Comparing the plant monitoring method of the present embodiment with the conventional plant monitoring method has the following differences.
  • the trend graph of each process data displayed on the screen is selected by the user. Therefore, the trend graph of the process data not selected by the user is not displayed on the screen.
  • the important variable extraction unit 224 determines M pieces of process monitoring data with high importance. Then, the determined M pieces of process monitoring data are displayed on the screen of the user terminal 3.
  • the normal pattern generated by the normal pattern generating unit 214 is displayed on a trend graph. Therefore, it can be seen at a glance how much the current trend graph deviates from a typical pattern (normal pattern). Accordingly, the user can easily and visually and intuitively understand how much the situation at the time of monitoring has deviated from the typical state, without necessarily determining whether the system is abnormal or normal.
  • the normal pattern and the time-series data of the actually measured values are simultaneously displayed.
  • the fluctuation range of the normal value hereinafter, referred to as “normal fluctuation range”. May be displayed.
  • the normal fluctuation range is a range in which the value fluctuates in a state where no abnormality has occurred.
  • the normal fluctuation range may be set based on a predetermined criterion based on the scale parameter of the time-series data in the past predetermined period.
  • a representative value such as an average value and a control limit value defined by a standard deviation may be displayed.
  • a user has been monitoring a plant using a trend graph of process monitoring data, whereas a method of monitoring an abnormal degree (contribution plot) of process monitoring data using MSPC has a problem that it is difficult for a user to intuitively understand. there were. Further, when the user feels that something is abnormal, it is necessary to open and confirm the trend graph by himself. Therefore, the affinity with the conventional monitoring method is low. That is, conventionally, a trend graph is always displayed even in a normal case, but MSPC has low affinity with such a configuration.
  • FIG. 6 is a display example when a normal pattern (dashed-dotted line) and measured value data (solid line) are displayed.
  • a normal MSPC the state shown in FIG. 6 is determined to be abnormal. If there is no display corresponding to the normal state pattern, it is difficult for a user with little experience to determine why the state shown in FIG. 6 is abnormal. In such a case, by displaying the normal pattern and the data of the actually measured values together, the movement to be referred to is visualized, so that it is easier to grasp the current situation.
  • the user confirms “process variables to be viewed now” automatically determined by the system as an extension of the conventional plant monitoring of “monitoring process data by a trend graph”. be able to. Therefore, plant monitoring can be performed efficiently. Furthermore, when the normal pattern of the “process variable to be viewed now” is displayed, it is easy to determine whether the time-series data of the actually measured values up to the monitoring time point deviates from the normal pattern (whether or not it is abnormal). It is possible to make a judgment.
  • FIG. 9 is a block diagram illustrating a functional configuration of a process monitoring support device 2a according to the second embodiment.
  • the process monitoring support device 2a according to the second embodiment includes a state definition unit 21a, a state calculation unit 22a, and a display control unit 23a instead of the state definition unit 21, the state calculation unit 22, and the display control unit 23.
  • the display control unit 23a receives an instruction to change the display content from the user terminal 3, and changes the display according to the instruction.
  • a process monitoring support device 2a according to the second embodiment will be described.
  • the contribution ratio definition unit 213 defines the process soundness contribution ratio of each process variable of the process soundness index as in the following equation (6).
  • Equation (6) indicates that when ⁇ i (t) and ⁇ i (t) are defined by an average value and a standard deviation, respectively, without depending on time t, a statistic called a Z value used in SPC or the like is obtained. Corresponding. This value indicates the evaluation of the variation around the average value by the magnification relative to the standard deviation.
  • the index definition unit 212a defines a process health index by integrating the process health contribution rates defined by Expression (6).
  • the integration method may be any method.
  • the sum or average of the process soundness contribution rates calculated by equation (6) may be defined as the process soundness index.
  • the maximum value of the process soundness contribution rate calculated by the equation (6) (that is, the one closest to the abnormality) may be defined as the process soundness index.
  • the process soundness index When the process soundness index is defined by the sum of the process soundness contribution rates, it may be set to n (the number of process variables) times the reference set for the process soundness contribution rates.
  • the process soundness index When the process soundness index is defined by the maximum value of the process soundness contribution rate, the same threshold value may be used as a criterion. As described above, by applying the concept of the control limit of the SPC, the process soundness index and the abnormality judgment criterion of the process soundness contribution ratio are set.
  • an abnormality determination criterion in a normal MSPC may be applied to the abnormality determination criterion. That is, for example, the abnormality determination criterion may be set using the theoretical threshold values of the following equations (7) and (8) for the Q statistic and the T ⁇ 2 statistic. Equation (7) is the theoretical threshold of the Q statistic. Equation (8) is the theoretical threshold of the T ⁇ 2 statistic.
  • p is the number of process variables left in the model.
  • ⁇ _i is a diagonal element of ⁇ . That is, ⁇ _i is the sum of the i-th power of each component included in the error term.
  • Equation (8) p is the number of selected process variables (remaining in the model). m is the number of all process variables.
  • F (p, mp, ⁇ ) represents the F distribution when the degree of freedom is (p, mp) and the confidence limit is ⁇ . Note that 0.01 or 0.05 is often used for ⁇ .
  • An abnormality determination threshold value can be determined using Expression (7) or Expression (8) according to the statistics used for the process soundness index.
  • the statistics include the Q statistics, the T ⁇ 2 statistics, the sum of the Q statistics and the T ⁇ 2 statistics, and the maximum values of the Q statistics and the T ⁇ 2 statistics.
  • a threshold value within a predetermined range is defined as the ratio of the process soundness contribution rate of each process variable to the statistics (process soundness contribution rate / statistics). (For example, 20% to 50%), and a criterion for determining a process variable having a large process soundness contribution rate as abnormal can be provided.
  • the operations of the current data acquisition unit 221, the index calculation unit 222, the contribution ratio calculation unit 223, and the important variable extraction unit 224 in the second embodiment are almost the same as the operations of the same function unit in the first embodiment.
  • the MSPC is applied to the definition of the process soundness index and the process soundness contribution ratio
  • the process soundness index and the process soundness are deviated from typical patterns. The calculation is different because the contribution is defined.
  • the state calculation unit 22a in the second embodiment further includes an abnormal data extraction unit 225.
  • the abnormal data extraction unit 225 determines the process health index calculated by the index calculation unit 222 and the process health contribution ratio of each process monitoring data calculated by the contribution ratio calculation unit 223 in the determination criterion definition unit. It is determined based on the criteria defined by H.215 whether it is normal.
  • the abnormal data extraction unit 225 adds a flag indicating a determination result to each value. For example, a flag “0” may be added for normal, and a flag “1” for abnormal.
  • the display control unit 23a according to the second embodiment further includes a monitoring point display control unit 233 and a contribution rate display control unit 234.
  • FIGS. 10 and 11 are diagrams illustrating a specific example of a screen displayed by the operation of the display control unit 23a according to the second embodiment.
  • the display control unit 23a will be described with reference to FIGS.
  • abnormal data data to which a flag of abnormal has been added by the abnormal data extracting unit 225 in displaying the time series data of the process health index.
  • abnormal data data to which a flag of normal is added: hereinafter, referred to as “normal data”.
  • the abnormal data may be displayed, for example, in an emphasized manner. Specifically, the abnormal data may be displayed in a more prominent color (for example, red) than the normal data, or may be displayed using a prominent line type (for example, a thicker line). Further, the display may be performed in a different manner depending on the magnitude of the abnormality of the abnormality data.
  • a flicker display may be performed for a case where the degree of abnormality (degree of deviation) is larger than a predetermined reference. Further, the larger the degree of the abnormality, the larger the display area of the graph may be displayed.
  • abnormal data is indicated by a broken line
  • normal data is indicated by a solid line.
  • the threshold value defined by the criterion definition unit 215 may be displayed in the time series data of the process health index.
  • the threshold value may be displayed in a line type (for example, yellow) different from the abnormal data and the normal data.
  • abnormal data and normal data may be displayed in different modes.
  • abnormal data may be displayed in a manner emphasized compared to normal data.
  • the display control unit 23a may be configured to control the user terminal 3 so as to issue an audio alarm when the highlighted display is performed. It is also possible to simultaneously issue an alarm by voice or text. When the alarm is issued by voice, the alarm may be issued at a volume lower than the normal volume. With such a configuration, it is possible to support the user's plant monitoring without overemphasizing “abnormality” and avoiding flooding of alarms.
  • the monitoring point display control unit 233 controls not a display of time-series data of process variables such as a trend graph, but a display of devices installed in a plant as in a process flow.
  • FIG. 11 is a diagram illustrating a specific example of a screen controlled by the monitoring point display control unit 233.
  • the monitoring point display control unit 233 performs highlighting near the image of the device related to the abnormal data. For example, near the device from which the abnormal data was obtained, the name of the process variable from which the abnormal data was obtained may be displayed. At this time, the names of the process variables may be displayed in the same manner (for example, red) as the highlighting performed by the monitoring data display control unit 232.
  • the character strings “aeration air volume” and “MLSS” are displayed as specific examples of the name of the process variable from which the abnormal data was obtained.
  • the monitoring point display control unit 233 displays an image of this device or an image such as the name of the above-described process variable to prompt the user to pay attention to this image.
  • an image of an arrow in which an arrow is directed to an image of a device or an image of a name of a process variable may be displayed.
  • the target image may be displayed in the same mode (for example, red) as the highlighting performed by the monitoring data display control unit 232.
  • the image of the arrow is only a specific example of the image of interest.
  • the image of interest may be a character such as "!, An image of a star, or an image indicating that the image is generally dangerous.
  • FIG. 11 to the left of the character strings “aeration air volume” and “MLSS”, an image of a right-pointing arrow with an arrow directed to these character strings is displayed as a specific example of the image of interest.
  • an image indicating the time-series data of the process variable from which the abnormal data has been acquired may be further displayed.
  • an attention image or the like is displayed in the process flow because abnormal data is obtained for the aeration air volume and the MLSS, but the time series of the aeration air volume and the MLSS for which the abnormal data is obtained along with the process flow are further displayed.
  • An image showing the data is displayed.
  • the monitoring point display control unit 233 displays the display including the device related to the time-series data of the selected process variable (for example, the process flow may be displayed as shown in FIG.
  • the ⁇ contribution ratio display control unit 234 displays a process variable having a large value of the process soundness contribution ratio of each process variable.
  • the contribution rate display control unit 234 generates the contribution rate image 43, for example.
  • a graph for example, a bar graph
  • the contribution image 43 may display, for example, the names of a predetermined number of process variables and the values of the process health contribution in the descending order of the process soundness contribution. It is preferable that the contribution rate image 43 is displayed as a bar graph sorted in a criterion such as a descending order of the value of the process soundness contribution rate.
  • the arrangement of the process variables in the contribution ratio image 43 may be arranged vertically or horizontally.
  • the selection may be performed by an operation of placing the cursor on the graph and clicking the graph. Then, the user selects the process variable displayed on the contribution rate image 43.
  • the monitoring data display control unit 232 displays a graph of the time-series data of the process variable selected by the user in an area where the process data of “total energy intensity” is displayed. At this time, the magnitude of the process soundness contribution ratio is displayed in a bar graph. Therefore, the user can select the process data to be displayed while referring to the magnitude of the process soundness contribution ratio.
  • FIG. 12 is a flowchart illustrating an example of a processing flow of the state definition unit 21a of the process monitoring support device 2a according to the second embodiment.
  • the state defining unit 21a waits until the timing of the predetermined cycle T0 (step S301-NO).
  • the past data acquisition unit 211 acquires past data (step S302).
  • the normal pattern generation unit 214 generates a normal pattern using the acquired past data (step S303).
  • the contribution ratio definition unit 213a defines a process soundness contribution ratio for each process variable using the acquired normal pattern and past data (step S304).
  • the index definition unit 212a defines a process health index using the calculated process health contribution rate (step S305).
  • the process soundness index, the process soundness contribution rate, and the normal state pattern thus obtained are used in the subsequent processing of the state calculation unit 22a.
  • the processing of the state calculation unit 22a is basically the same as that of the first embodiment, and a description thereof will be omitted.
  • the process soundness index is defined in a stacked manner based on the deviation from the normal state pattern without using a special diagnosis algorithm.
  • Such a process has an effect that the soundness can be easily interpreted from the viewpoint that the user monitors the sign of abnormality. That is, since a special diagnosis algorithm is not used, when the soundness is poor (the degree of abnormality is high), it can be easily determined by comparing the badness with the normal pattern.
  • the data determined to be abnormal data by the abnormal data extraction unit 225 is displayed in a mode different from normal data.
  • the abnormal data may be displayed as highlighting.
  • the monitoring point display control unit 233 performs highlighting near the image of the device related to the abnormal data. Therefore, when an abnormality occurs in the plant, the user can more easily estimate the location of the abnormality and the cause of the abnormality.
  • the contribution rate image 43 is displayed. Therefore, based on the contribution rate image 43, the user can easily select a process variable of the trend graph that the user wants to display on the screen. Therefore, plant monitoring that easily reflects the user's intention can be realized.
  • FIG. 13 is a diagram illustrating a modification of the second embodiment.
  • a part of the function implemented as the process monitoring support device 2a in the second embodiment is implemented in an information processing device (process monitoring support server) installed at a remote location via the network 4. I have.
  • the process monitoring support server 9 and the process monitoring support device 2b are communicably connected via the network 4.
  • the process monitoring support server 9 and the process monitoring support device 2b are communicably connected to each other by the communication units (the communication unit 91 and the communication unit 24) included therein functioning.
  • the functions of the state definition unit 21a and the state calculation unit 22a in the second embodiment are implemented in the process monitoring support server 9.
  • the data collection unit 201 of the process monitoring support device 2b transmits the data recorded in the data storage unit 202 to the process monitoring support server 9 at a predetermined timing.
  • the timing may be a timing at which the past data acquisition unit 211 of the process monitoring support server 9 transmits request data indicating that data is requested, or a timing of a predetermined cycle.
  • the state definition unit 21a of the process monitoring support server 9 performs a process based on the received data.
  • the data collection unit 201 of the process monitoring support device 2b transmits the current data to the process monitoring support server 9 at a predetermined timing.
  • the state calculation unit 22a of the process monitoring support server 9 performs a process based on the received current data.
  • the state calculation unit 22a transmits data indicating the processing result to the process monitoring support device 2b.
  • the display control unit 23a of the process monitoring support device 2b generates data representing a screen based on the received data, and causes the user terminal 3 to display the data.
  • the function of the display control unit 23a may be implemented in the process monitoring support server 9.
  • the data collection unit 201 and the data storage unit 202 may be installed in the plant as devices different from the process monitoring support device 2b.
  • the data collection unit 201 may transmit the past data and the current data by communicating with the process monitoring support server 9.
  • the process monitoring support device 2b itself may be implemented as a user terminal.
  • the user terminal may be a mobile terminal device such as a smartphone or tablet.
  • the functions of the process monitoring support device 2 described in the first and second embodiments can be implemented in a cloud, and a support service can be provided to a plant or the like where a user is located.
  • the state calculation unit 22 and the state calculation unit 22a do not necessarily need to be designed to process only the time-series data (current data) at the time of monitoring.
  • the processing may be performed on time-series data at a past point in time specified by the user.
  • an advanced monitoring / diagnosis system can be monitored while maintaining a plant monitoring method having high affinity with plant monitoring currently widely performed. It becomes possible to incorporate into. As a result, it is possible to realize efficient plant monitoring and plant monitoring that can reduce the possibility of the user overlooking the state at an abnormal time (for example, at the time of abnormality) and can respond more quickly.

Abstract

The purpose of the present invention is to more effectively utilize an analysis result for data obtained from a process while maintaining a display to which a user is accustomed to using. A process monitoring assistance device (2) according to an embodiment has a data acquisition unit (221), an index calculation unit (222), a contribution rate calculation unit (223), and a display control unit (23). The data acquisition unit acquires time-series data for a plurality of types of process variables indicating a state of a process (1) being monitored. The index calculation unit calculates an index indicating the possibility that no abnormality is occurring in the process being monitored and the process is sound, on the basis of the plurality of types of process variables. In relation to the index calculated by the index calculation unit the contribution rate calculation unit calculates, for each of the plurality of types of process variables, a contribution rate indicating the percentage of contribution to an increased possibility that no abnormality is occurring in the process being monitored and the process is sound. The display control unit generates display information so as to display, on a screen of a user terminal (3), information pertaining to a portion of the process variables for which the contribution rate is relatively small.

Description

プロセス監視支援装置、プロセス監視支援システム、プロセス監視支援方法、プロセス監視支援プログラム及び端末装置Process monitoring support device, process monitoring support system, process monitoring support method, process monitoring support program, and terminal device
　本発明の実施形態は、プロセス監視支援装置、プロセス監視支援システム、プロセス監視支援方法、プロセス監視支援プログラム及び端末装置に関する。 The embodiments of the present invention relate to a process monitoring support device, a process monitoring support system, a process monitoring support method, a process monitoring support program, and a terminal device.
　従来、監視対象となるプロセスから取得可能なデータをＭＳＰＣ（Multivariate Statistical Process Control：多変量統計的プロセス管理）等の手法を用いて分析することで、監視対象プロセスの状態を識別するとともに、監視対象プロセスの状態に応じた支援情報をユーザに提供する技術が考案されている。しかしながら、実際の多くのプラントでは、上述した技術が適用されることなく、プロセス監視装置（ＳＣＡＤＡ：Supervisory Control And Data Acquisition）で収集したプロセスデータ（流量、温度、水質、操作量など）のトレンドグラフを表示することで監視を行うケースが圧倒的に多い。その理由の一つとして、従来から使いなれているトレンドグラフ等の表示の方がユーザにとって視認性が高いことが挙げられる。そのため、従来の技術では分析結果が有効に活用されていなかった。 Conventionally, by analyzing data that can be obtained from a process to be monitored by using a method such as MSPC (Multivariate Statistical Process Control), the state of the process to be monitored is identified, and the status of the process to be monitored is identified. Techniques have been devised for providing support information to a user according to the state of a process. However, in many actual plants, a trend graph of process data (flow rate, temperature, water quality, operation amount, etc.) collected by a process monitoring apparatus (SCADA: Supervisory Control And Data Acquisition) without applying the above technology is applied. There are overwhelmingly many cases of monitoring by displaying. One of the reasons is that the display of a trend graph or the like which has been conventionally used has higher visibility for the user. Therefore, the analysis results have not been effectively used in the conventional technology.
特開平０８－２４１１２１号公報JP 08-241121 A 特開２００４－３０３００７号公報JP 2004-303007 A 特開２００７－０６５８８３号公報JP 2007-065883 A
　本発明が解決しようとする課題は、ユーザが使い慣れた表示を維持しつつ、プロセスから得られたデータの分析結果をより有効に活用することを可能とするプロセス監視支援装置、プロセス監視支援システム、プロセス監視支援方法、プロセス監視支援プログラム及び端末装置を提供することである。 The problem to be solved by the present invention is to provide a process monitoring support device, a process monitoring support system, and a process monitoring method that enable a user to more effectively utilize an analysis result of data obtained from a process while maintaining a familiar display. An object of the present invention is to provide a process monitoring support method, a process monitoring support program, and a terminal device.
　実施形態のプロセス監視支援装置は、データ取得部と、指標演算部と、寄与率演算部と、表示制御部と、を持つ。データ取得部は、監視対象プロセスの状態を示すプロセス変数の時系列データを複数種取得する。指標演算部は、複数種の前記プロセス変数に基づいて、前記監視対象プロセスの状態に異常が生じておらず健全である可能性を示す指標を算出する。寄与率演算部は、前記指標演算部によって算出される前記指標に関して、異常が生じておらず健全である可能性が高くなることに寄与した割合を示す寄与率を、複数種の前記プロセス変数毎に算出する。表示制御部は、前記寄与率の値が相対的に小さい一部のプロセス変数に関する情報をユーザ端末の画面に表示するように表示情報を生成する。 The process monitoring support device according to the embodiment includes a data acquisition unit, an index calculation unit, a contribution ratio calculation unit, and a display control unit. The data acquisition unit acquires a plurality of types of time-series data of process variables indicating the state of the process to be monitored. The index calculation unit calculates an index indicating a possibility that the monitoring target process is healthy without any abnormality based on the plurality of types of the process variables. Contribution rate calculation unit, for the index calculated by the index calculation unit, the contribution rate indicating the rate of contribution to increase the possibility that there is no abnormality and is healthy, for each of the plurality of types of process variables Is calculated. The display control unit generates display information so as to display information on some process variables having a relatively small value of the contribution ratio on a screen of the user terminal.
第１実施形態のプロセス監視支援装置２の構成の具体例を示す図である。FIG. 2 is a diagram illustrating a specific example of a configuration of a process monitoring support device 2 according to the first embodiment. 本実施形態のプロセス監視支援装置２によってユーザに提供される画面の具体例を示す図である。FIG. 4 is a diagram illustrating a specific example of a screen provided to a user by the process monitoring support device 2 of the present embodiment. ユーザに提供される画面の具体例を示す図である。It is a figure showing the example of the screen provided to the user. ユーザに提供される画面の具体例を示す図である。It is a figure showing the example of the screen provided to the user. ＭＳＰＣの通常の監視方法で提供される画面例を示す図である。It is a figure showing the example of a screen provided by the usual monitoring method of MSPC. 常態パターン（一点鎖線）と実測値のデータ（実線）とを表示した場合の表示例である。It is a display example in the case of displaying a normal state pattern (dashed-dotted line) and measured value data (solid line). 第１実施形態のプロセス監視支援装置２の状態定義部２１の処理の流れの例を示すフローチャートである。5 is a flowchart illustrating an example of a processing flow of a state definition unit 21 of the process monitoring support device 2 according to the first embodiment. 第１実施形態のプロセス監視支援装置２の状態演算部２２及び表示制御部２３の処理の流れの例を示すフローチャートである。5 is a flowchart illustrating an example of a processing flow of a state calculation unit 22 and a display control unit 23 of the process monitoring support device 2 according to the first embodiment. 第２実施形態のプロセス監視支援装置２ａの機能構成を示すブロック図である。It is a block diagram showing the functional composition of process monitoring support device 2a of a 2nd embodiment. 第２実施形態における表示制御部２３ａの動作によって表示される画面の具体例を示す図である。It is a figure showing the example of the screen displayed by operation of display control part 23a in a 2nd embodiment. 第２実施形態における表示制御部２３ａの動作によって表示される画面の具体例を示す図である。It is a figure showing the example of the screen displayed by operation of display control part 23a in a 2nd embodiment. 第２実施形態のプロセス監視支援装置２の状態定義部２１ａの処理の流れの例を示すフローチャートである。It is a flowchart which shows the example of the flow of a process of the state definition part 21a of the process monitoring support apparatus 2 of 2nd Embodiment. 変形例の構成を示す図である。It is a figure showing composition of a modification.
　以下、実施形態のプロセス監視支援装置、プロセス監視支援システム、プロセス監視支援方法、プロセス監視支援プログラム及び端末装置を、図面を参照して説明する。なお、“ａ＿ｂ”は、“ａ”という文字の右下に小さい“ｂ”が付されていることを示す。また、“ａ＾ｂ”は、“ａ”という文字の右上に小さい“ｂ”が付されていることを示す。 Hereinafter, a process monitoring support device, a process monitoring support system, a process monitoring support method, a process monitoring support program, and a terminal device according to an embodiment will be described with reference to the drawings. Note that “a_b” indicates that a small “b” is attached to the lower right of the character “a”. Also, “a @ b” indicates that a small “b” is attached to the upper right of the character “a”.
　図１は、実施形態のプロセス監視支援装置２の構成の具体例を示す図である。図１は、プロセス監視支援装置２の監視対象が下水高度処理プロセス１である具体例を示している。下水高度処理プロセス１は、下水から窒素及びリンを除去することを目的としたプロセスである。下水高度処理プロセス１は、最初沈澱池１０１、嫌気槽１０２、無酸素槽１０３、好気槽１０４及び最終沈澱池１０５を有する。処理対象の下水（以下「被処理水」という。）は、最初沈澱池１０１、嫌気槽１０２、無酸素槽１０３、好気槽１０４、最終沈澱池１０５の順に送水され処理される。 FIG. 1 is a diagram showing a specific example of the configuration of the process monitoring support device 2 of the embodiment. FIG. 1 shows a specific example in which the monitoring target of the process monitoring support device 2 is the advanced sewage treatment process 1. Sewage advanced treatment process 1 is a process aimed at removing nitrogen and phosphorus from sewage. The advanced sewage treatment process 1 has a sedimentation basin 101, an anaerobic tank 102, an anoxic tank 103, an aerobic tank 104 and a final sedimentation basin 105. The sewage to be treated (hereinafter referred to as “treatment water”) is first sent to the sedimentation basin 101, the anaerobic tank 102, the anoxic tank 103, the aerobic tank 104, and the final sedimentation basin 105 in this order.
　最初沈澱池１０１は、下水高度処理プロセス１に送られてくる被処理水の貯水池である。最初沈澱池１０１では、沈澱により比重の重い固形物が被処理水から分離される。嫌気槽１０２は、有機物を分解する微生物を被処理水に投入するための水槽である。嫌気槽１０２において、被処理水は空気が供給されない状態で攪拌される。これにより、微生物に体内のリンを吐き出させる。一般にこの処理をリン吐出という。無酸素槽１０３は、被処理水から窒素を除去するための水槽である。具体的には、無酸素槽１０３では、後段の好気槽１０４から戻された被処理水が嫌気槽１０２から送られてきた被処理水に混ぜられ、空気を供給されない状態で攪拌される。無酸素槽１０３では、微生物の働きにより被処理水中の硝酸が窒素に分解され、大気に放出される。一般にこの処理を脱窒という。 The first settling basin 101 is a reservoir for the water to be treated sent to the advanced sewage treatment process 1. First, in the sedimentation basin 101, solid matter having a high specific gravity is separated from the water to be treated by the sedimentation. The anaerobic tank 102 is a tank for charging microorganisms that decompose organic substances into the water to be treated. In the anaerobic tank 102, the water to be treated is stirred without supplying air. This causes the microorganisms to exhale phosphorus in the body. This process is generally called phosphorus discharge. The oxygen-free tank 103 is a water tank for removing nitrogen from the water to be treated. Specifically, in the anoxic tank 103, the water to be treated returned from the aerobic tank 104 at the subsequent stage is mixed with the water to be treated sent from the anaerobic tank 102, and is agitated without supplying air. In the anoxic tank 103, the nitric acid in the water to be treated is decomposed into nitrogen by the action of microorganisms and released to the atmosphere. Generally, this treatment is called denitrification.
　好気槽１０４は、被処理水中の有機物の分解と、リンの除去及びアンモニアの硝化とを行うための水槽である。具体的には、被処理水に空気を供給して微生物を活性化させ、微生物に有機物を分解させるとともに、微生物に被処理水中のリンを吸収させる。嫌気状態でリンを吐出しその代りに有機物を蓄積した状態の微生物は活性化されることにより吐き出した以上のリンを吸収するため、被処理水中のリンが除去される。また、好気槽１０４では、被処理水に空気が供給されることによりアンモニアが硝酸に分解される。一般にこの処理を硝化という。 The aerobic tank 104 is a tank for decomposing organic substances in the water to be treated, removing phosphorus, and nitrifying ammonia. Specifically, air is supplied to the water to be treated to activate the microorganisms, and the microorganisms decompose organic substances, and the microorganisms absorb phosphorus in the water to be treated. Microorganisms that discharge phosphorus in an anaerobic state and instead accumulate organic matter absorb the phosphorus that is exhaled by being activated, so that the phosphorus in the water to be treated is removed. In the aerobic tank 104, ammonia is decomposed into nitric acid by supplying air to the water to be treated. This treatment is generally called nitrification.
　最終沈澱池１０５は、リンの除去及びアンモニアの硝化が行われた被処理水の貯水池である。最終沈澱池１０５では沈澱によって被処理水に残存する固形物が分離され、上澄みの清澄水が処理済みの水として放流される。 (4) The final sedimentation basin 105 is a reservoir for treated water from which phosphorus has been removed and ammonia has been nitrified. In the final sedimentation basin 105, solids remaining in the water to be treated are separated by sedimentation, and the supernatant clear water is discharged as treated water.
　最初沈澱池余剰汚泥引き抜きポンプ１１１は、最初沈澱池１０１から沈澱した汚泥を引き抜いて除去するポンプである。最初沈澱池余剰汚泥引き抜きポンプ１１１は、引き抜いた汚泥の流量を計測する流量センサを有する。 The first settling basin excess sludge extraction pump 111 is a pump for extracting and removing sludge settled from the first settling basin 101. The first settling basin excess sludge extraction pump 111 has a flow rate sensor for measuring the flow rate of the extracted sludge.
　ブロワ１１２は、好気槽１０４に酸素を供給する送風機である。ブロワ１１２は、供給した空気の流量を計測する流量センサを有する。 The blower 112 is a blower that supplies oxygen to the aerobic tank 104. The blower 112 has a flow sensor that measures the flow rate of the supplied air.
　循環ポンプ１１３は、被処理水を好気槽１０４から無酸素槽１０３に返送するポンプである。循環ポンプ１１３は、返送した被処理水の流量を計測する流量センサを有する。 The circulation pump 113 is a pump that returns the water to be treated from the aerobic tank 104 to the anoxic tank 103. The circulation pump 113 has a flow rate sensor that measures the flow rate of the returned treated water.
　返送汚泥ポンプ１１４は、最終沈澱池１０５から沈澱した汚泥の一部を引き抜いて嫌気槽１０２に返送するポンプである。返送汚泥ポンプ１１４は、返送した汚泥の流量を計測する流量センサを有する。 The return sludge pump 114 is a pump for extracting a part of the sludge settled from the final settling basin 105 and returning the sludge to the anaerobic tank 102. The returned sludge pump 114 has a flow rate sensor that measures the flow rate of returned sludge.
　最終沈澱池余剰汚泥引き抜きポンプ１１５は、最終沈澱池１０５から沈澱した汚泥を引き抜いて除去するポンプである。最終沈澱池余剰汚泥引き抜きポンプ１１５は、引き抜いた汚泥の流量を計測する流量センサを有する。 The final settling basin excess sludge extraction pump 115 is a pump for extracting and removing the sludge settled from the final settling basin 105. The final sedimentation tank excess sludge extraction pump 115 has a flow rate sensor for measuring the flow rate of the extracted sludge.
　雨量センサ１２１は、下水高度処理プロセス１に流入する付近の雨量を計測するセンサである。下水流入量センサ１２２は、下水高度処理プロセス１に流入する下水（以下「流入下水」という。）の流量を計測するセンサである。流入ＴＮセンサ１２３は、流入下水に含まれる全窒素量（ＴＮ）を計測するセンサである。流入ＴＰセンサ１２４は、流入下水に含まれる全リン量（ＴＰ）を計測するセンサである。流入有機物センサ１２５は、流入下水に含まれる有機物量を計測するＵＶ（吸光度）センサ又はＣＯＤ（化学的酸素要求量）センサである。 The rainfall sensor 121 is a sensor that measures the rainfall near the sewage altitude treatment process 1. The sewage inflow sensor 122 is a sensor that measures the flow rate of sewage flowing into the sewage advanced treatment process 1 (hereinafter, referred to as “inflow sewage”). The inflow TN sensor 123 is a sensor that measures the total amount of nitrogen (TN) contained in the inflow sewage. The inflow TP sensor 124 is a sensor that measures the total amount of phosphorus (TP) contained in the inflow sewage. The inflowing organic matter sensor 125 is a UV (absorbance) sensor or a COD (chemical oxygen demand) sensor that measures the amount of organic matter contained in the inflowing sewage.
　ＯＲＰセンサ１２６は、嫌気槽１０２のＯＲＰ（酸化－還元電位）を計測するセンサである。嫌気槽ｐＨセンサ１２７は、嫌気槽１０２のｐＨを計測するセンサである。無酸素槽ＯＲＰセンサ１２８は、無酸素槽１０３のＯＲＰを計測するセンサである。無酸素槽ｐＨセンサ１２９は、無酸素槽１０３のｐＨを計測するセンサである。リン酸センサ１３０は、好気槽１０４のリン酸濃度を計測するセンサである。ＤＯセンサ１３１は、好気槽１０４の溶存酸素濃度（ＤＯ）を計測するセンサである。アンモニアセンサ１３２は、好気槽１０４のアンモニア濃度を計測するセンサである。ＭＬＳＳセンサ１３３は、嫌気槽１０２、無酸素槽１０３又は好気槽１０４の少なくとも一箇所で活性汚泥濃度（ＭＬＳＳ）を計測するセンサである。 The ORP sensor 126 is a sensor that measures the ORP (oxidation-reduction potential) of the anaerobic tank 102. Anaerobic tank pH sensor 127 is a sensor that measures the pH of anaerobic tank 102. The anoxic tank ORP sensor 128 is a sensor that measures the ORP of the anoxic tank 103. The anoxic tank pH sensor 129 is a sensor that measures the pH of the anoxic tank 103. The phosphoric acid sensor 130 is a sensor that measures the concentration of phosphoric acid in the aerobic tank 104. The DO sensor 131 is a sensor that measures the dissolved oxygen concentration (DO) in the aerobic tank 104. The ammonia sensor 132 is a sensor that measures the ammonia concentration in the aerobic tank 104. The MLSS sensor 133 is a sensor that measures the activated sludge concentration (MLSS) in at least one of the anaerobic tank 102, the anaerobic tank 103, and the aerobic tank 104.
　水温センサ１３４は、無酸素槽１０３又は好気槽１０４の少なくとも一箇所で水温を計測するセンサである。余剰汚泥ＳＳセンサ１３５は、最終沈澱池１０５から引き抜かれる汚泥の固形物（ＳＳ）濃度を計測するセンサである。放流ＳＳセンサ１３６は、最終沈澱池１０５から放流される水のＳＳ濃度を計測するセンサである。汚泥界面センサ１３７は、最終沈澱池１０５の汚泥界面レベルを計測するセンサである。下水放流量センサ１３８は、放流水の流量を計測するセンサである。放流ＴＮセンサ１３９は、放流水に含まれる全窒素量を計測するセンサである。放流ＴＰセンサ１４０は、放流水に含まれる全リン量を計測するセンサである。放流有機物センサ１４１は、放流水に含まれる有機物量を計測するＵＶセンサ又はＣＯＤセンサである。 The water temperature sensor 134 is a sensor that measures the water temperature in at least one location of the anoxic tank 103 or the aerobic tank 104. The surplus sludge SS sensor 135 is a sensor that measures the solid (SS) concentration of sludge extracted from the final sedimentation basin 105. The discharge SS sensor 136 is a sensor that measures the SS concentration of the water discharged from the final sedimentation basin 105. The sludge interface sensor 137 is a sensor that measures the sludge interface level of the final settling basin 105. The sewage discharge flow sensor 138 is a sensor that measures the flow rate of discharged water. The discharge TN sensor 139 is a sensor that measures the total amount of nitrogen contained in the discharge water. The discharge TP sensor 140 is a sensor that measures the total amount of phosphorus contained in the discharge water. The discharged organic substance sensor 141 is a UV sensor or a COD sensor that measures the amount of organic substances contained in the discharged water.
　なお、上記の最初沈澱池余剰汚泥引き抜きポンプ１１１、ブロワ１１２、循環ポンプ１１３、返送汚泥ポンプ１１４及び最終沈澱池余剰汚泥引き抜きポンプ１１５など機器のそれぞれは所定周期の制御で動作する。また、最初沈澱池余剰汚泥引き抜きポンプ１１１、ブロワ１１２、循環ポンプ１１３、返送汚泥ポンプ１１４及び最終沈澱池余剰汚泥引き抜きポンプ１１５の機器それぞれが有する流量センサを含む上記の各センサは、所定周期でセンシング対象を計測する。以下、最初沈澱池余剰汚泥引き抜きポンプ１１１、ブロワ１１２、循環ポンプ１１３、返送汚泥ポンプ１１４及び最終沈澱池余剰汚泥引き抜きポンプ１１５のそれぞれが有する流量センサを総称して操作量センサと称し、その他のセンサを総称してプロセスセンサと称する。各操作量センサ及び各プロセスセンサは、所定周期のセンシングによって得られた計測データをプロセスデータとしてプロセス監視支援装置２に送信する。 機器 Each of the above-mentioned devices such as the first settling basin excess sludge pulling pump 111, the blower 112, the circulation pump 113, the return sludge pump 114, and the final settling basin excess sludge pulling pump 115 operates under the control of a predetermined cycle. In addition, the above-described sensors including the flow rate sensors included in the equipment of the initial sedimentation tank excess sludge extraction pump 111, the blower 112, the circulation pump 113, the return sludge pump 114, and the final sedimentation tank excess sludge extraction pump 115 are sensed at predetermined intervals. Measure the target. Hereinafter, the flow rate sensors included in the first settling tank excess sludge extraction pump 111, the blower 112, the circulation pump 113, the return sludge pump 114, and the final settling tank excess sludge extraction pump 115 are collectively referred to as operation amount sensors, and other sensors are referred to. Are collectively referred to as process sensors. Each of the operation amount sensors and each of the process sensors transmits measurement data obtained by sensing in a predetermined cycle to the process monitoring support device 2 as process data.
　次に、プロセス監視支援装置の２つの実施形態について説明する。まず、第１実施形態のプロセス監視支援装置２について説明する。 Next, two embodiments of the process monitoring support device will be described. First, the process monitoring support device 2 according to the first embodiment will be described.
　［第１実施形態］
　プロセス監視支援装置２は、バスで接続されたＣＰＵ（Central Processing Unit）やメモリや補助記憶装置などを備え、監視支援プログラムを実行する。プロセス監視支援装置２は、監視支援プログラムの実行によってデータ収集部２０１、データ保存部２０２、状態定義部２１、状態演算部２２及び表示制御部２３を備える装置として機能する。なお、プロセス監視支援装置２の各機能の全て又は一部は、ＡＳＩＣ（Application Specific Integrated Circuit）やＰＬＤ（Programmable Logic Device）やＦＰＧＡ（Field Programmable Gate Array）等のハードウェアを用いて実現されてもよい。また、データ収集部２０１は、ＰＬＣ（Programmable Logic Controller）を用いてプロセス監視支援装置２とは異なる筐体の装置として実装されてもよい。監視支援プログラムは、コンピュータ読み取り可能な記録媒体に記録されてもよい。コンピュータ読み取り可能な記録媒体とは、例えばフレキシブルディスク、光磁気ディスク、ＲＯＭ、ＣＤ－ＲＯＭ等の可搬媒体、コンピュータシステムに内蔵されるハードディスク等の記憶装置である。監視支援プログラムは、電気通信回線を介して送信されてもよい。 [First Embodiment]
The process monitoring support device 2 includes a CPU (Central Processing Unit), a memory, and an auxiliary storage device connected by a bus, and executes a monitoring support program. The process monitoring support device 2 functions as a device including a data collection unit 201, a data storage unit 202, a state definition unit 21, a state calculation unit 22, and a display control unit 23 by executing a monitoring support program. Note that all or a part of each function of the process monitoring support device 2 may be realized using hardware such as an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array). Good. Further, the data collection unit 201 may be implemented as a device having a housing different from that of the process monitoring support device 2 using a PLC (Programmable Logic Controller). The monitoring support program may be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in a computer system. The monitoring support program may be transmitted via a telecommunication line.
　データ収集部２０１は、各操作量センサ及び各プロセスセンサからプロセスデータを取得する。取得されるプロセスデータは、監視対象プロセスの状態を示す各プロセス変数の時系列データである。データ収集部２０１は、取得されたプロセスデータを、予め決められたフォーマットにしたがってデータ保存部２０２に記録する。 The data collection unit 201 acquires process data from each operation amount sensor and each process sensor. The acquired process data is time-series data of each process variable indicating the state of the monitoring target process. The data collection unit 201 records the obtained process data in the data storage unit 202 according to a predetermined format.
　データ保存部２０２は、磁気ハードディスク装置や半導体記憶装置等の記憶装置を用いて構成される。データ保存部２０２は、データ収集部２０１によって取得されたプロセスデータを記憶する。 The data storage unit 202 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The data storage unit 202 stores the process data obtained by the data collection unit 201.
　状態定義部２１は、監視対象となっているプロセスの健全性に関するデータを定義する。状態定義部２１は、データ保存部２０２に記録されているプロセス変数の過去の時系列データを読み出す。状態定義部２１は、プロセスの健全性を判断するために用いられる指標（以下「プロセス健全性指標」という。）と、各プロセス変数がプロセス健全性指標に与える影響度を示す寄与率（以下「プロセス健全性寄与率」という。）と、を定義する。プロセス健全性指標は、監視対象プロセスの状態に異常が生じている可能性を示す指標である。プロセス健全性寄与率は、プロセス健全性指標に関して、異常が生じている可能性が高くなることに各プロセス変数が寄与した割合を示す値である。したがって、プロセス健全性寄与率はプロセス変数毎に算出される。状態定義部２１は、プロセスの過去の通常状態における典型的な代表値や代表パターンデータ（以下「常態パターン」という。）を生成する。状態定義部２１は、例えば所定の周期Ｔ０で動作する。 (4) The state definition unit 21 defines data relating to the health of the process to be monitored. The state definition unit 21 reads past time-series data of process variables recorded in the data storage unit 202. The state definition unit 21 includes an index (hereinafter, referred to as a “process health index”) used to determine the health of the process, and a contribution ratio (hereinafter, referred to as “process health index”) indicating the degree of influence of each process variable on the process health index. Process soundness contribution ratio.)). The process health index is an index indicating a possibility that an abnormality has occurred in the state of the monitored process. The process soundness contribution ratio is a value indicating a ratio of each process variable contributing to an increase in the possibility that an abnormality has occurred in the process soundness index. Therefore, the process soundness contribution rate is calculated for each process variable. The state definition unit 21 generates typical representative values and representative pattern data (hereinafter, referred to as “normal state patterns”) in the past normal state of the process. The state definition unit 21 operates, for example, at a predetermined cycle T0.
　状態演算部２２は、データ保存部２０２に記録されているプロセス変数の監視時点の現在の時系列データを読み出す。状態演算部２２は、状態定義部２１における定義を用いて、プロセス健全性指標と各プロセス変数のプロセス健全性寄与率とを算出する。状態演算部２２は、例えばプロセス健全性寄与率の低い順（異常度の高い順）で各プロセス変数に対して高い重要度を付与する。状態演算部２２は、監視時点において重要度の高いものから順に所定数の上位のプロセス変数を抽出する。状態演算部２２は、例えば所定の周期Ｔ１（＜＜Ｔ０）で動作する。 The state calculation unit 22 reads the current time-series data at the time of monitoring the process variables recorded in the data storage unit 202. The state calculation unit 22 calculates a process soundness index and a process soundness contribution ratio of each process variable using the definition in the state definition unit 21. The state calculation unit 22 assigns a high importance to each process variable in, for example, an ascending order of the process soundness contribution rate (an ascending order of abnormality). The state calculation unit 22 extracts a predetermined number of upper process variables in descending order of importance at the time of monitoring. The state calculation unit 22 operates, for example, in a predetermined cycle T1 (<< T0).
　表示制御部２３は、状態演算部２２で得られた情報に基づいて、重要度の高いものから順に所定数の上位のプロセス変数の時系列データを示す画像を生成する。このような画像は、例えばトレンドグラフとして生成されてもよい。表示制御部２３は、プロセス健全性指標の時系列データと、重要度の高い各プロセス変数の時系列データと、を表示するためのデータ（表示情報）を生成し、ユーザ端末３に出力する。 The display control unit 23 generates an image indicating the time-series data of a predetermined number of higher-order process variables in descending order of importance based on the information obtained by the state calculation unit 22. Such an image may be generated, for example, as a trend graph. The display control unit 23 generates data (display information) for displaying the time series data of the process health index and the time series data of each process variable with high importance, and outputs the data to the user terminal 3.
　ユーザ端末３は、表示制御部２３によって生成された情報に基づいて画像を表示する。ユーザ端末３は、表示された画像に対する操作を受け付ける。ユーザ端末３は、例えばプラントの管理者やオペレータ（運転員）などのユーザによって使用される。 The user terminal 3 displays an image based on the information generated by the display control unit 23. The user terminal 3 receives an operation on the displayed image. The user terminal 3 is used by a user such as a plant manager or an operator (operator).
　次に、プロセス監視支援装置２の各機能部についてより詳細に説明する。まず、状態定義部２１について詳細に説明する。状態定義部２１は、過去データ取得部２１１、指標定義部２１２、寄与率定義部２１３及び常態パターン生成部２１４として機能する。 Next, each functional unit of the process monitoring support device 2 will be described in more detail. First, the state definition unit 21 will be described in detail. The state definition unit 21 functions as a past data acquisition unit 211, an index definition unit 212, a contribution ratio definition unit 213, and a normal state pattern generation unit 214.
　過去データ取得部２１１は、データ保存部２０２から各プロセス変数の所定期間（以下「過去所定期間」という。）の時系列データである過去データ（オフラインデータ）を読み出す。プロセス健全性指標をユニークに定義する場合は、ユーザがユーザ端末３を操作することによって過去所定期間を定義できてもよい。このような定義は、監視画面上のＧＵＩで入力可能であることが好ましい。一方、多くの実プラントでは、プラント運用の変更や季節的な変化によりプラント状態も徐々に変化していく場合が多い。そのため、プロセス健全性指標を定義する過去所定期間を更新していく方が現実的である場合も多い。このような場合には、過去所定期間の長さが予め設定されており、所定のサイクルで過去所定期間の長さに応じた過去所定期間が更新されてもよい。この場合、過去データ取得部２１１は、データ取得の時点で定義されている過去所定期間の過去データをデータ保存部２０２から読み出す。例えば、過去所定期間の長さを１週間とした場合、１週間毎に過去１週間分のデータが過去所定期間のデータとして読み出されてもよい。このようにして過去データ取得部２１１によって取得された過去の時系列データをＸと記載する。この時系列データＸは、行方向にプロセス変数、列方向に過去所定期間にわたる時系列データを持つ行列である。以下の説明では、プロセス変数の数をｎ、時系列データ数をｍとする。したがって、時系列データＸはｍ×ｎの時系列データである。 (4) The past data acquisition unit 211 reads out past data (offline data) that is time-series data of a predetermined period (hereinafter, referred to as “predetermined period”) of each process variable from the data storage unit 202. When the process soundness index is uniquely defined, the user may be able to define a past predetermined period by operating the user terminal 3. It is preferable that such a definition can be input by a GUI on the monitoring screen. On the other hand, in many actual plants, the plant state often changes gradually due to a change in plant operation or a seasonal change. Therefore, it is often more realistic to update the past predetermined period defining the process health index. In such a case, the length of the past predetermined period is set in advance, and the predetermined past period may be updated according to the length of the past predetermined period in a predetermined cycle. In this case, the past data acquisition unit 211 reads from the data storage unit 202 the past data for the past predetermined period defined at the time of data acquisition. For example, assuming that the length of the past predetermined period is one week, data for the past one week may be read as data of the past predetermined period every week. The past time-series data acquired by the past data acquisition unit 211 in this manner is described as X. The time-series data X is a matrix having process variables in the row direction and time-series data over a predetermined past period in the column direction. In the following description, the number of process variables is n and the number of time-series data is m. Therefore, the time series data X is m × n time series data.
　指標定義部２１２は、過去データ取得部２１１によって読み出された過去の時系列データＸを用いて、プロセス健全性指標を定義する。例えば、指標定義部２１２は、多変量解析や機械学習技術を使用することによって、プロセス健全性指標を定義してもよい。プロセス健全性指標の定義手法はどのように実装されてもよい。プロセス健全性指標は、過去の時系列データＸから生成されているため、ｎ個のプロセス変数の情報を含んでおり、プロセスの健全性を計る指標が１つの指標に集約されている。 The index definition unit 212 uses the past time-series data X read by the past data acquisition unit 211 to define a process health index. For example, the index definition unit 212 may define the process health index by using a multivariate analysis or a machine learning technique. The method of defining the process health index may be implemented in any manner. Since the process soundness index is generated from the past time-series data X, it includes information on n process variables, and the indicators for measuring the soundness of the process are collected into one index.
　プロセス健全性指標の具体例として、以下のようなものがある。アドバンストなプロセス監視診断技術であるＭＳＰＣ（多変量統計的プロセス管理）と呼ばれる方法で用いられるＱ統計量。HotellingのＴ＾２統計量と呼ばれる異常検出用のデータ。品質工学の分野で用いられるタグチ法（ＭＳＰＣと類似の技術）等で用いられるマハラノビス距離。なお、マハラノビス距離は、ＭＳＰＣで用いられるＴ＾２統計量と本質的に同等のものである。ただし、ＭＳＰＣでは、次元削減（ｎ’＜＜ｎの次元）を行った上でＴ＾２統計量が定義されるが、マハラノビス距離を用いた方法では、ｎ次元の空間上で距離が定義される。次元をそろえた場合、本質的にマハラノビス距離とＴ＾２統計量とはほぼ同一のものであり、定数倍の差があるだけである。 具体 The following are specific examples of process integrity indicators. Q-statistic used in a method called MSPC (Multivariate Statistical Process Management), which is an advanced process monitoring and diagnostic technique. Data for detecting abnormalities called Hotelling's T ＾ 2 statistic. Mahalanobis distance used in the Taguchi method (a technique similar to MSPC) used in the field of quality engineering. The Mahalanobis distance is essentially equivalent to the T と 2 statistic used in MSPC. However, in the MSPC, the T ＾ 2 statistic is defined after performing dimension reduction (dimension of n ′ << n), but in the method using the Mahalanobis distance, the distance is defined in an n-dimensional space. You. When the dimensions are aligned, the Mahalanobis distance and the T ＾ 2 statistic are essentially the same, and there is only a difference of a constant multiple.
　本実施形態では、プロセス健全性指標の具体例としてＭＳＰＣを用いる。ＭＳＰＣで用いられる主成分分析（ＰＣＡ）を用いると、時系列データＸは以下の式（１）のように分解できる。 In this embodiment, MSPC is used as a specific example of the process soundness index. When the principal component analysis (PCA) used in MSPC is used, the time-series data X can be decomposed as in the following equation (1).
Figure JPOXMLDOC01-appb-M000001
Figure JPOXMLDOC01-appb-M000001
　Ｔ＿ａはサンプル数ｍと主成分数ｎとによるｍ×ｎ行列であり、スコア行列と呼ばれる。Ｐ＿ａはｎ個のプロセス変数とｎ個の主成分との関係を示すｎ×ｎ行列であり、ローディング行列と呼ばれる。Ｔは主成分をｐ（＜＜ｎ）個で打ち切ったＴ＿ａの部分行列であり、通常はこのＴをスコア行列と呼ぶのが一般的である。同様にＰは主成分をｐ個で打ち切ったＰ＿ａの部分行列（ｎ×ｐ）であり、通常はこのＰをローディング行列と呼ぶのが一般的である。また、Ｅはサンプル数ｍとプロセス変数の数ｎとによる誤差行列（ｍ×ｎ）であり、主成分をｐ個で打ち切った場合の誤差を表す。 T_a is an m × n matrix based on the number m of samples and the number n of principal components, and is called a score matrix. P_a is an n × n matrix indicating a relationship between n process variables and n principal components, and is called a loading matrix. T is a sub-matrix of T_a in which the principal components are truncated by p (<< n), and this T is generally called a score matrix. Similarly, P is a sub-matrix (n × p) of P_a in which the main component is truncated by p, and this P is generally called a loading matrix. E is an error matrix (m × n) based on the number m of samples and the number n of process variables, and represents an error when the number of principal components is cut off by p.
　以下では、Ｔ＿ａとＴとを明確に区別するため、Ｔ＿ａをスコア行列と称し、Ｔを主要スコア行列と称する。同様に、以下では、Ｐ＿ａとＰとを明確に区別するため、Ｐ＿ａをローディング行列と称し、Ｐを主要ローディング行列と称する。これら各行列を用いればＱ統計量Ｑ（ｘ（ｔ））及びHotellingのＴ＾２統計量Ｔ＾２（ｘ（ｔ））は次の式（２）及び式（３）と定義される。 In the following, in order to clearly distinguish T_a from T, T_a is referred to as a score matrix, and T is referred to as a main score matrix. Similarly, hereinafter, in order to clearly distinguish P_a from P, P_a is referred to as a loading matrix, and P is referred to as a main loading matrix. If these matrices are used, the Q statistic Q (x (t)) and Hotelling's T ＾ 2 statistic T ＾ 2 (x (t)) are defined as the following equations (2) and (3).
Figure JPOXMLDOC01-appb-M000002
Figure JPOXMLDOC01-appb-M000002
Figure JPOXMLDOC01-appb-M000003
Figure JPOXMLDOC01-appb-M000003
　式（２）及び式（３）において、ｘ（ｔ）は過去の時系列データＸのｔ番目の要素を表す。Ｉは適当なサイズの単位行列である。Λは主成分の分散を対角要素として持つ行列であり、分散の正規化を意味する。後述するオンラインの監視の際には、このｘ（ｔ）が、オンラインで計測されてくるプロセスデータに置き換わって計算される。 に お い て In Equations (2) and (3), x (t) represents the t-th element of the past time-series data X. I is a unit matrix of an appropriate size. Λ is a matrix having the variance of the principal component as a diagonal element, which means normalization of the variance. At the time of online monitoring to be described later, this x (t) is calculated by replacing the process data measured online.
　式（２）のＱ統計量は、過去の時系列データＸに含まれる各プロセス変数の関係性（相関）がどの程度維持されているか（又はどの程度くずれているか）を示す指標である。式（３）のHotellingのＴ＾２統計量は、過去の時系列データＸに含まれる各プロセス変数がどの程度変動しているかを正規化して表現した指標である。 Ｑ The Q statistic in the equation (2) is an index indicating how much the relationship (correlation) of each process variable included in the past time-series data X is maintained (or how much it is distorted). The Hotelling's T ＾ 2 statistic in Expression (3) is an index that represents, by normalizing, how much each process variable included in the past time-series data X fluctuates.
　通常のＭＳＰＣでは、これら二つの統計量が異常検出用の指標として用いられる。しかし、寄与率定義部２１３で定義されるプロセス健全性寄与率の大小を各プロセス変数に対して一意に定義しておくことが好ましい。そのため、プロセス健全性指標も一つだけ定義されてもよい。 で は In a normal MSPC, these two statistics are used as indices for abnormality detection. However, it is preferable that the magnitude of the process soundness contribution rate defined by the contribution rate definition unit 213 is uniquely defined for each process variable. Therefore, only one process health index may be defined.
　例えば、Ｑ統計量又はHotellingのＴ＾２統計量のいずれか一つがプロセス健全性指標として定義されてもよい。一般的に、Ｑ統計量の方が軽微な変動を示す（検出する）傾向がある。このような軽微な変動には、ユーザが一つのプロセス変数だけを見ているだけでは気づきにくい異常兆候も含まれる。そのため、このような軽微な変動を監視することを目的とする場合は、Ｑ統計量をプロセス健全性指標とすることが好ましい。一方、HotellingのＴ＾２統計量は、各プロセス変数の比較的大きな変動を検出する傾向がある。そのため、プロセスの明確な変動を監視したい場合にはHotellingのＴ＾２統計量をプロセス健全性指標とすることが好ましい。また、プロセスの異常兆候を含むなんらかの変動は、Ｑ統計量又はHotellingのＴ＾２統計量のいずれかで検出される。そのため、Ｑ統計量とHotellingのＴ＾２統計量との大きな値を持つ方（すなわち異常度が高い方）がプロセス健全性指標として用いられてもよい。また、Ｑ統計量とHotellingのＴ＾２統計量との和（すなわち総合的な異常指標）がプロセス健全性指標として用いられてもよい。
　また、健全度を０～１の範囲で定義し、“１”は完全に健全であることを示し、“０”は完全に異常であることを示すように定義されてもよい。この場合、Ｑ統計量及びＴ＾２統計量が、ｅｘｐ（－ａ×統計量）のような変換式で変換された値が健全度として定義されてもよい。この場合、“ａ”は調整パラメータであり、０より大きい値である。 For example, either one of the Q statistic or Hotelling's T ＾ 2 statistic may be defined as the process health index. Generally, the Q statistic tends to show (detect) minor fluctuations. Such minor fluctuations include abnormal signs that are difficult for the user to notice only by looking at one process variable. Therefore, when the purpose is to monitor such a small change, it is preferable to use the Q statistic as the process soundness index. On the other hand, Hotelling's T ＾ 2 statistic tends to detect relatively large variations in each process variable. Therefore, when it is desired to monitor a clear change in the process, it is preferable to use Hotelling's T ＾ 2 statistic as the process soundness index. Also, any fluctuations, including signs of process anomalies, are detected in either the Q statistic or the Hotelling T ＾ 2 statistic. Therefore, the one having a larger value of the Q statistic and the Hotelling's T ＾ 2 statistic (that is, the one having the higher abnormality level) may be used as the process soundness index. Further, the sum of the Q statistic and the Hotelling's T ＾ 2 statistic (that is, a comprehensive abnormality index) may be used as the process soundness index.
Further, the degree of soundness may be defined in a range of 0 to 1, wherein "1" indicates that the soundness is completely sound, and "0" may be defined so as to indicate that the soundness is completely abnormal. In this case, a value obtained by converting the Q statistic and the T ＾ 2 statistic by a conversion formula such as exp (−a × statistic) may be defined as the soundness. In this case, “a” is an adjustment parameter and is a value larger than 0.
　寄与率定義部２１３は、指標定義部２１２によって定義された指標に対する各プロセス変数のプロセス健全性寄与率を定義する。プロセス健全性指標が、Ｑ統計量やHotellingのＴ＾２統計量で定義されている場合は、各統計量に対する寄与率をプロセス健全性寄与率の定義とすればよい。Ｑ統計量の寄与量と、HotellingのＴ＾２統計量の寄与量とは、それぞれ以下の式（４）及び式（５）のように定義される。 The contribution ratio definition unit 213 defines a process soundness contribution ratio of each process variable with respect to the index defined by the index definition unit 212. When the process soundness index is defined by the Q statistic or Hotelling's T ＾ 2 statistic, the contribution to each statistic may be defined as the process soundness contribution. The contribution of the Q statistic and the contribution of the Hotelling's T 統計 2 statistic are defined as in the following Expressions (4) and (5), respectively.
Figure JPOXMLDOC01-appb-M000004
Figure JPOXMLDOC01-appb-M000004
Figure JPOXMLDOC01-appb-M000005
Figure JPOXMLDOC01-appb-M000005
　式（４）及び式（５）において、ｘ（ｔ，ｎ）はある時刻ｔにおける過去データｘ（ｔ）のうちｎ番目のプロセス変数を表す。また、式（４）において、Ｆ（：，ｎ）の記載は、行列Ｆからｎ列を抽出することを表す。同様に、式（５）において、Ｐ（：，ｎ）の記載は、行列Ｐからｎ列を抽出することを表す。式（４）は、ｎ番目のプロセス変数の軸に対するＱ統計量の射影を表す式であり、式（５）によって各プロセス変数がＱ統計量にどの程度寄与しているかを計算することができる。また、式（５）は、Ｔ＾２統計量をｎ個の各プロセス変数成分の和に分解する式である。式（５）によって各プロセス変数がＴ＾２統計量にどの程度寄与しているかを計算することができる。 に お い て In Equations (4) and (5), x (t, n) represents the n-th process variable in the past data x (t) at a certain time t. In the equation (4), the description of F (:, n) indicates that n columns are extracted from the matrix F. Similarly, in Expression (5), the description of P (:, n) indicates that n columns are extracted from the matrix P. Equation (4) is an equation representing the projection of the Q statistic on the axis of the nth process variable, and it is possible to calculate how much each process variable contributes to the Q statistic by Equation (5). . Equation (5) is an equation for decomposing the T ＾ 2 statistic into the sum of n process variable components. Using equation (5), it is possible to calculate how much each process variable contributes to the T ＾ 2 statistic.
　Ｑ統計量とHotellingのＴ＾２統計量とのうち大きい方がプロセス健全性指標として用いられる場合には、対応する大きい方の統計量に対する寄与量がプロセス健全性寄与率として用いられてもよい。一方、Ｑ統計量とHotellingのＴ＾２統計量との和がプロセス健全性指標として用いられる場合には、式（４）及び式（５）で得られたそれぞれの値の和がプロセス健全性寄与率として用いられてもよい。 If the larger of the Q statistic and Hotelling's T ＾ 2 statistic is used as the process soundness index, the corresponding contribution to the larger statistic may be used as the process soundness contribution rate. . On the other hand, when the sum of the Q statistic and the Hotelling's T ＾ 2 statistic is used as the process soundness index, the sum of the respective values obtained by Expressions (4) and (5) is the process soundness. It may be used as a contribution rate.
　このように構成されることによって、ある時刻において各プロセス変数の監視データが入力されると、各プロセス変数のその時刻におけるプロセス健全性寄与率を順位づけることができる。 With this configuration, when monitoring data of each process variable is input at a certain time, the process health contribution ratio of each process variable at that time can be ranked.
　なお、以上の処理は、通常のＭＳＰＣで行われている処理とほぼ同様である。プロセス健全性指標やプロセス健全性寄与率を一意に決定するために、Ｑ統計量及びＴ＾２統計量に若干の処理を加える場合がある点で異なる。 Note that the above processing is almost the same as the processing performed by normal MSPC. The difference is that some processing may be added to the Q statistic and the T ＾ 2 statistic in order to uniquely determine the process soundness index and the process soundness contribution rate.
　常態パターン生成部２１４は、指標定義部２１２及び寄与率定義部２１３によってプロセス健全性指標及びプロセス健全性寄与率の定義に用いられた過去データを用いて、常態パターンを生成する。常態パターンとは、各プロセス変数の典型的な値や典型的なパターンを示す情報である。以下、常態パターン生成部２１４の詳細について説明する。 The normal state pattern generation unit 214 generates a normal state pattern by using the past data used by the index definition unit 212 and the contribution ratio definition unit 213 to define the process soundness index and the process soundness contribution ratio. The normal pattern is information indicating a typical value or a typical pattern of each process variable. Hereinafter, details of the normal pattern generation unit 214 will be described.
　常態パターン生成部２１４は、過去の時系列データＸを用いて各プロセス変数の常態パターンを生成する。例えば、常態パターン生成部２１４は、過去の時系列データＸの各プロセス変数の時系列データ（すなわちＸのそれぞれの列ベクトル）毎の位置母数を常態パターンとして算出してもよい。位置母数は、例えば平均値、中央値、刈込平均値などの値である。過去の時系列データＸに特別なパターンが認められない場合は、このようなシンプルな方法で常態パターンが得られてもよい。常態パターン生成部２１４は、常態パターンとして得られる典型値の周りにどの程度のばらつきがあるかを示す尺度母数をさらに常態パターンの値の一部として算出してもよい。尺度母数は、例えば標準偏差、中央値絶対偏差（ＭＡＤ：Median Absolute Deviation）、四分範囲等の値である。 The normal pattern generation unit 214 generates a normal pattern of each process variable using the past time-series data X. For example, the normal pattern generation unit 214 may calculate a position parameter of each time-series data of each process variable of the past time-series data X (that is, each column vector of X) as a normal pattern. The position parameter is, for example, a value such as an average value, a median value, and a pruning average value. When a special pattern is not recognized in the past time-series data X, a normal pattern may be obtained by such a simple method. The normal pattern generation unit 214 may further calculate, as a part of the value of the normal pattern, a scale parameter indicating the degree of variation around the typical value obtained as the normal pattern. The scale parameter is, for example, a value such as a standard deviation, a median absolute deviation (MAD: Median Absolute Deviation), or a quadrant.
　常態パターン生成部２１４は、過去の時系列データＸの中に何等かパターンが認められる場合には、単なる位置母数ではなく、各プロセスデータの典型的なパターンデータを生成してもよい。例えば下水処理プロセスにおいては、通常は人間の生活様式に応じて日単位の周期性を持つパターンが認められる。また、土日などの休日には平日と異なる挙動を示す事も多い。このような場合は、週単位の周期性を持つパターンが認められる。このような日単位や週単位のパターンが認められるデータは、下水処理プロセスだけでなく、浄水処理の水需要パターン、エネルギープラントのエネルギー需要パターン、自動車などの交通量の交通量パターン、等のように人間の生活様式と密接に関連するインフラ系のプラントでは特に顕著に認められる。このような特徴を過去の時系列データＸが持つ場合には、常態パターン生成部２１４は、例えば所定時間単位（例えば１分単位、１時間単位）で位置母数を算出し、算出された値を所定期間分（例えば１日分、１週間分）繋げた時系列データを常態パターンとして生成してもよい。このような場合にも、常態パターンの一部として尺度母数が算出されてもよい。尺度母数の算出に際しては、常態パターンが生成された場合と同じ様に所定時間単位の尺度母数が算出されてもよい。また、一様に過去の時系列データＸの列ベクトル毎の尺度母数が算出されてもよい。 (4) When any pattern is found in the past time-series data X, the normal state pattern generation unit 214 may generate typical pattern data of each process data instead of a mere position parameter. For example, in a sewage treatment process, a pattern having a periodicity on a daily basis according to a human lifestyle is generally recognized. In addition, on holidays such as Saturday and Sunday, the behavior often differs from that on weekdays. In such a case, a pattern having a weekly periodicity is recognized. The data for which daily and weekly patterns are recognized include not only sewage treatment processes but also water demand patterns for water purification, energy demand patterns for energy plants, traffic patterns for automobiles and other traffic, and so on. This is particularly noticeable in infrastructure plants that are closely related to human lifestyles. When the past time-series data X has such a characteristic, the normal state pattern generation unit 214 calculates the position parameter in, for example, a predetermined time unit (for example, 1 minute unit, 1 hour unit), and calculates the calculated value. May be generated as a normal pattern by connecting time series data for a predetermined period (for example, one day, one week). In such a case, the scale parameter may be calculated as a part of the normal pattern. When calculating the scale parameter, the scale parameter in a predetermined time unit may be calculated in the same manner as when the normal pattern is generated. Further, the scale parameter for each column vector of the past time-series data X may be calculated uniformly.
　次に、状態演算部２２について詳細に説明する。状態演算部２２は、現在データ取得部２２１、指標演算部２２２、寄与率演算部２２３及び重要変数抽出部２２４として機能する。 Next, the state calculation unit 22 will be described in detail. The state calculation unit 22 functions as a current data acquisition unit 221, an index calculation unit 222, a contribution ratio calculation unit 223, and an important variable extraction unit 224.
　現在データ取得部２２１は、データ保存部２０２から各プロセス変数の現在データ（オンラインデータ）を読み出す。現在データとは、監視を行う時点（以下「監視時点」という。）における各プロセス変数のデータである。 (4) The current data acquisition unit 221 reads the current data (online data) of each process variable from the data storage unit 202. The current data is the data of each process variable at the time of monitoring (hereinafter referred to as “monitoring time”).
　指標演算部２２２は、現在データ取得部２２１によって読み出された現在データと、指標定義部２１２によって定義されたプロセス健全性指標の定義式と、を用いた演算を行う。指標演算部２２２は、演算結果として、現在のプロセスの健全度を示す指標を取得する。以下、指標演算部２２２の処理の具体例について説明する。 The index operation unit 222 performs an operation using the current data read by the current data acquisition unit 221 and the definition formula of the process health index defined by the index definition unit 212. The index calculation unit 222 acquires an index indicating the current degree of health of the process as the calculation result. Hereinafter, a specific example of the process of the index calculation unit 222 will be described.
　指標演算部２２２は、現在データ取得部２２１によって読み出された監視時点のデータに対し、必要に応じてアウトライア（外れ値）処理を行う。そして、指標演算部２２２は、アウトライア処理が行われたデータを用いて、指標定義部２１２によって定義されたプロセス健全性指標を算出する。 The index calculation unit 222 performs outlier (outlier) processing on the data at the time of monitoring read by the current data acquisition unit 221 as necessary. Then, the index calculation unit 222 calculates the process health index defined by the index definition unit 212 using the data on which the outlier processing has been performed.
　例えば、プロセス健全性指標がＱ統計量で定義された場合には、指標演算部２２２は、式（２）のＸ（ｔ）に監視時点のデータを入力してＱ統計量を算出することによってプロセス健全性指標を算出する。Ｔ＾２統計量でプロセス健全性指標が定義された場合には、指標演算部２２２は、式（３）で定義されたＴ＾２統計量のＸ（ｔ）に監視時点のデータを入力してＴ＾２統計量を算出することによってプロセス健全性指標を算出する。Ｑ統計量とＴ＾２統計量との大きい方（悪い方）でプロセス健全性指標が定義されている場合や、Ｑ統計量及びＴ＾２統計量の和でプロセス健全性指標が定義されている場合にも、式（２）及び式（３）を用いて演算が行われることによって、現時点でのプロセス健全性指標を算出することができる。 For example, when the process health index is defined by the Q statistic, the index calculating unit 222 calculates the Q statistic by inputting the data at the monitoring time into X (t) of Expression (2). Calculate the process soundness index. When the process soundness index is defined by the T ＾ 2 statistic, the index calculation unit 222 inputs the data at the monitoring time point into X (t) of the T ＾ 2 statistic defined by Expression (3). By calculating the T ＾ 2 statistic, the process soundness index is calculated. When the process health index is defined by the larger (bad) of the Q statistic and the T ＾ 2 statistic, or when the process health index is defined by the sum of the Q statistic and the T ＾ 2 statistic Even in the case where the calculation is performed, the current process health index can be calculated by performing the calculation using the equations (2) and (3).
　寄与率演算部２２３は、現在データ取得部２２１によって読み出された現在データと、寄与率定義部２１３によって定義されたプロセス健全性指標に対する各プロセス変数のプロセス健全性寄与率の定義式と、を用いた演算を行う。寄与率演算部２２３は、演算結果として、現在の各プロセスのプロセス健全度寄与率を取得する。具体的には、寄与率演算部２２３は、式（４）及び式（５）を用いて演算を行う。Ｑ統計量が採用されている場合には、寄与率演算部２２３は式（４）を用いてプロセス健全性寄与率を算出する。Ｔ＾２統計量が採用されている場合には、寄与率演算部２２３は式（５）を用いてプロセス健全性寄与率を算出する。Ｑ統計量及びＴ＾２統計量の大きい方（悪い方）でプロセス健全性指標が定義されている場合や、Ｑ統計量及びＴ＾２統計量の和でプロセス健全性指標が定義されている場合には、式（４）及び式（５）を用いてプロセス健全性寄与率が算出される。 The contribution rate calculation unit 223 calculates the current data read by the current data acquisition unit 221 and the definition formula of the process health contribution rate of each process variable with respect to the process health index defined by the contribution rate definition unit 213. Perform the calculations used. The contribution rate calculation unit 223 acquires the current process health degree contribution rate of each process as the calculation result. Specifically, the contribution ratio calculation unit 223 performs the calculation using the equations (4) and (5). When the Q statistic is employed, the contribution rate calculation unit 223 calculates the process soundness contribution rate using Expression (4). When the T ＾ 2 statistic is employed, the contribution ratio calculation unit 223 calculates the process soundness contribution ratio using Expression (5). The process health index is defined by the larger (bad) Q statistic and T ＾ 2 statistic, or the process health index is defined by the sum of the Q statistic and T ＾ 2 statistic. In this case, the process soundness contribution rate is calculated using the equations (4) and (5).
　重要変数抽出部２２４は、各プロセス変数のプロセス健全性寄与率が低い順（異常度の高い順）にソートする。重要変数抽出部２２４は、ソート結果に基づいて、予め設定された所定の数の上位のプロセス変数を抽出する。以下、重要変数抽出部２２４の処理の具体例について説明する。重要変数抽出部２２４は、各プロセス変数について、そのプロセス健全性寄与率が大きい順に大きな値の重要度を付与する。重要変数抽出部２２４は、重要度の大きいものから順に所定数（Ｍ個）のプロセス変数を抽出する。Ｍ個のプロセス変数の時系列監視データは、後述するようにユーザによって監視に用いられる画面（ユーザ端末の画面）に表示される。そのため、Ｍの値は、通常は高々１０個程度までの値として設定されてもよい。 (4) The important variable extraction unit 224 sorts the process variables in the order of low process soundness contribution ratio (high order of abnormalities). The important variable extracting unit 224 extracts a predetermined number of upper process variables set in advance based on the sorting result. Hereinafter, a specific example of the process of the important variable extraction unit 224 will be described. The important variable extracting unit 224 assigns a larger value to each process variable in descending order of the process soundness contribution ratio. The important variable extraction unit 224 extracts a predetermined number (M) of process variables in descending order of importance. The time-series monitoring data of the M process variables is displayed on a screen (a screen of the user terminal) used for monitoring by the user as described later. Therefore, the value of M may normally be set as a value of up to about ten.
　表示制御部２３は、状態表示制御部２３１及び監視データ表示制御部２３２として機能する。 The display control unit 23 functions as the state display control unit 231 and the monitoring data display control unit 232.
　状態表示制御部２３１は、指標演算部２２２で演算されたプロセス健全性指標の所定の期間（所定の過去から現在まで：以下「対象期間」という。）の時系列データを表示するためのデータを生成する。 The state display control unit 231 generates data for displaying time-series data of the process health index calculated by the index calculation unit 222 for a predetermined period (from a predetermined past to the present: hereinafter, referred to as a “target period”). Generate.
　監視データ表示制御部２３２は、重要変数抽出部２２４によって抽出された各プロセス変数の対象期間のプロセス監視データ（時系列データ）を表示するためのデータを生成する。 The monitoring data display control unit 232 generates data for displaying the process monitoring data (time-series data) of each process variable extracted by the important variable extracting unit 224 during the target period.
　次に、プロセス監視支援装置２の表示制御部２３によってユーザに提示されるユーザインタフェースについて詳細に説明する。ユーザとシステムとが情報交換を行うインターフェイス部分に着目したことが本実施形態の特徴の一つである。ユーザインタフェースを、従来のプラント監視からの自然な延長・拡張な形で実現することに本実施形態の主眼の一つがある。そのため、従来の監視方法、ＭＳＰＣによるアドバンストな監視方法、本実施形態の監視方法、の３つの監視方法を比較しながら本実施形態の作用について説明する。 Next, the user interface presented to the user by the display control unit 23 of the process monitoring support device 2 will be described in detail. One of the features of the present embodiment is that the user and the system pay attention to an interface for exchanging information. One of the main objectives of the present embodiment is to realize the user interface in a natural extension / expansion form from the conventional plant monitoring. Therefore, the operation of the present embodiment will be described while comparing three monitoring methods, the conventional monitoring method, the advanced monitoring method using the MSPC, and the monitoring method of the present embodiment.
　図２は、本実施形態のプロセス監視支援装置２によってユーザに提供される画面の具体例を示す図である。図３は、ユーザに提供される画面の具体例を示す図である。 FIG. 2 is a diagram showing a specific example of a screen provided to the user by the process monitoring support device 2 of the present embodiment. FIG. 3 is a diagram illustrating a specific example of a screen provided to the user.
　状態表示制御部２３１は、図２の上段４１のように、指標演算部２２２によってリアルタイムで算出されたプロセス健全性指標の時系列データを表示するためのデータを生成する。プロセス健全性指標の時系列データは、例えば現在から所定期間の過去に遡った時系列データとして表示されてもよい。図２の例では、過去２４時間分の時系列データが表示されているが、この表示期間はユーザの操作に応じて変更されてもよい。 The state display control unit 231 generates data for displaying time-series data of the process health index calculated in real time by the index calculation unit 222, as shown in the upper part 41 of FIG. The time-series data of the process soundness index may be displayed as, for example, time-series data retroactive to a predetermined period in the past from the present. In the example of FIG. 2, the time-series data for the past 24 hours is displayed, but this display period may be changed according to the operation of the user.
　従来のプラント監視では、このようなプロセス健全性指標の表示は行われていなかった。一方で、従来のプラント監視では、ＭＳＰＣによるアドバンスト監視方法によってＱ統計量やＴ＾２統計量を用いた監視が行われていた。ＭＳＰＣではＱ統計量及びＴ＾２統計量という２種類の異常指標を区別して表示がなされていたのに対し、本実施形態では、一つのプロセス健全性指標として表示される。 (4) In the conventional plant monitoring, such a process integrity index was not displayed. On the other hand, in the conventional plant monitoring, monitoring using the Q statistic and the T ＾ 2 statistic has been performed by the advanced monitoring method using the MSPC. In the MSPC, two types of abnormal indicators, ie, the Q statistic and the T ＾ 2 statistic, are displayed separately, whereas in the present embodiment, they are displayed as one process health index.
　監視データ表示制御部２３２は、図２の下段４２のように、重要変数抽出部２２４によって抽出されたＭ個（図２の例ではＭ＝８）のプロセス変数の時系列データを表示するためのデータを生成する。監視データ表示制御部２３２は、上段４１に表示されるプロセス健全性指標の時系列データと同じ期間の時系列データ（いわゆるトレンドグラフ）が表示されることが望ましい。この際、監視データ表示制御部２３２は、常態パターン生成部２１４によって生成された常態パターンのデータも併せて表示してもよい。図２の下段４２の各時系列データのグラフのうち、実線で示された時系列データは、監視時点までに計測された実測値の時系列データを示す。図２の下段４２の各時系列データのグラフのうち、一点鎖線で示された時系列データは、常態パターン生成部２１４によって生成された常態パターンを示す。常態パターンが一つの値として算出される場合には、時系列データのグラフにおいて、その値を示す一本の横棒として常態パターンが表示されてもよい。 The monitoring data display control unit 232 is for displaying the time series data of M process variables (M = 8 in the example of FIG. 2) extracted by the important variable extraction unit 224 as shown in the lower part 42 of FIG. Generate data. It is desirable that the monitoring data display control unit 232 displays time series data (so-called trend graph) of the same period as the time series data of the process health index displayed in the upper stage 41. At this time, the monitoring data display control unit 232 may also display the normal pattern data generated by the normal pattern generation unit 214. In the time series data graph in the lower part 42 of FIG. 2, the time series data indicated by the solid line indicates the time series data of the actually measured values measured up to the monitoring time. In the graph of each time-series data in the lower part 42 of FIG. 2, the time-series data indicated by a dashed line indicates the normal pattern generated by the normal pattern generation unit 214. When the normal pattern is calculated as one value, the normal pattern may be displayed as one horizontal bar indicating the value in the graph of the time-series data.
　以上の一連の処理は、状態演算部２２の処理の周期であるＴ１以上の長い周期である周期Ｔ２で繰り返し実行されて表示が行われてもよい。例えば、Ｔ１＝１分の場合、Ｔ２＝１時間であってもよい。 The series of processes described above may be repeatedly executed and displayed in a cycle T2 which is a long cycle longer than T1 which is a cycle of the process of the state calculation unit 22. For example, when T1 = 1 minute, T2 = 1 hour.
　このような本実施形態のプラント監視方法と従来のプラント監視方法とを比較すると以下の様な違いがある。従来のプラント監視方法では、画面に表示される各プロセスデータのトレンドグラフは、ユーザによって選択されていた。そのため、ユーザによって選択されなかったプロセスデータのトレンドグラフは画面に表示されなかった。これに対し、本実施形態では、重要変数抽出部２２４によって、重要度の高いＭ個のプロセス監視データが判断される。そして、判断されたＭ個のプロセス監視データがユーザ端末３の画面に表示される。 比較 Comparing the plant monitoring method of the present embodiment with the conventional plant monitoring method has the following differences. In the conventional plant monitoring method, the trend graph of each process data displayed on the screen is selected by the user. Therefore, the trend graph of the process data not selected by the user is not displayed on the screen. On the other hand, in the present embodiment, the important variable extraction unit 224 determines M pieces of process monitoring data with high importance. Then, the determined M pieces of process monitoring data are displayed on the screen of the user terminal 3.
　また、本実施形態では、常態パターン生成部２１４によって生成された常態パターンがトレンドグラフに表示される。そのため、現在のトレンドグラフが典型的なパターン（常態パターン）からどの程度ずれているかという事が一見してわかる。これにより、必ずしもシステムによって異常や正常の判断を行わなくても、ユーザは監視時点の状況が典型的な状態からどの程度ずれているか視覚的、直観的に容易に把握することができる。例えば図２の具体例では、常態パターンと実測値の時系列データとを同時に表示しているが、図３に示すように、通常時の値の変動範囲（以下「通常変動範囲」という。）が表示されてもよい。通常変動範囲は、異常が生じていない状態で値が変動する範囲である。例えば過去所定期間における時系列データの尺度母数等に基づいて所定の基準で通常変動範囲が設定されてもよい。 Also, in the present embodiment, the normal pattern generated by the normal pattern generating unit 214 is displayed on a trend graph. Therefore, it can be seen at a glance how much the current trend graph deviates from a typical pattern (normal pattern). Accordingly, the user can easily and visually and intuitively understand how much the situation at the time of monitoring has deviated from the typical state, without necessarily determining whether the system is abnormal or normal. For example, in the specific example of FIG. 2, the normal pattern and the time-series data of the actually measured values are simultaneously displayed. However, as shown in FIG. 3, the fluctuation range of the normal value (hereinafter, referred to as “normal fluctuation range”). May be displayed. The normal fluctuation range is a range in which the value fluctuates in a state where no abnormality has occurred. For example, the normal fluctuation range may be set based on a predetermined criterion based on the scale parameter of the time-series data in the past predetermined period.
　また、プロセス監視データが日変動や週変動などのパターンを持たない場合には、図４に示すように、平均値などの代表値と標準偏差などで定義した管理限界値が表示されてもよい。 Further, when the process monitoring data does not have a pattern such as a daily variation or a weekly variation, as shown in FIG. 4, a representative value such as an average value and a control limit value defined by a standard deviation may be displayed. .
　アドバンスト監視として知られる通常のＭＳＰＣと本実施形態との相違は以下のとおりである。図５はＭＳＰＣの通常の監視方法で提供される画面例を示す。ＭＳＰＣでは、図５の下段に示すように、各プロセス変数のプロセス健全性寄与率が、寄与率プロットと呼ばれるバーグラフなどの形態で表示される。そして、どのプロセス変数に異常の兆候が認められるかについての判断は、ユーザに委ねられる。ユーザは、自身が確認する必要があると判断したプロセス変数について、そのトレンドグラフを表示するような操作を行い、画面に表示されたトレンドグラフを視認することによってプラント監視が行われる。ユーザは従来からプロセス監視データのトレンドグラフでプラント監視を行っているのに対し、ＭＳＰＣでプロセス監視データの異常度（寄与率プロット）を監視する方法では、ユーザにとって直観的にわかりづらいという問題があった。さらに、ユーザが自ら異常と感じた際に自らそのトレンドグラフを開いて確認する必要があった。そのため、従来の監視方法との親和性が低い。すなわち、従来は正常な場合でも常時トレンドグラフが表示されていたが、ＭＳＰＣはそのような構成との親和性が低い。 The differences between the normal MSPC known as advanced monitoring and this embodiment are as follows. FIG. 5 shows an example of a screen provided by a normal monitoring method of the MSPC. In the MSPC, as shown in the lower part of FIG. 5, the process soundness contribution ratio of each process variable is displayed in a form such as a bar graph called a contribution ratio plot. Then, it is left to the user to determine which process variable has a sign of abnormality. The user performs an operation of displaying a trend graph for a process variable determined to be required to be checked by himself, and performs plant monitoring by visually recognizing the trend graph displayed on the screen. Conventionally, a user has been monitoring a plant using a trend graph of process monitoring data, whereas a method of monitoring an abnormal degree (contribution plot) of process monitoring data using MSPC has a problem that it is difficult for a user to intuitively understand. there were. Further, when the user feels that something is abnormal, it is necessary to open and confirm the trend graph by himself. Therefore, the affinity with the conventional monitoring method is low. That is, conventionally, a trend graph is always displayed even in a normal case, but MSPC has low affinity with such a configuration.
　また、各プロセス変数の常態パターンに相当する表示も行われていなかった。そのため、異常が生じた時の要因推定が困難であった。図６は常態パターン（一点鎖線）と実測値のデータ（実線）とを表示した場合の表示例である。通常のＭＳＰＣでは、図６のような状態は異常と判断される。もし常態パターンに相当する表示が無い場合には、なぜ図６のような状態で異常状態なのかを判断することは、特に経験の浅いユーザには難しい。このような時、常態パターンと実測値のデータとを併せて表示することによって、参照すべき動きが可視化されるため現状の把握が一層容易になる。 表示 Also, the display corresponding to the normal pattern of each process variable was not performed. For this reason, it was difficult to estimate a factor when an abnormality occurred. FIG. 6 is a display example when a normal pattern (dashed-dotted line) and measured value data (solid line) are displayed. In a normal MSPC, the state shown in FIG. 6 is determined to be abnormal. If there is no display corresponding to the normal state pattern, it is difficult for a user with little experience to determine why the state shown in FIG. 6 is abnormal. In such a case, by displaying the normal pattern and the data of the actually measured values together, the movement to be referred to is visualized, so that it is easier to grasp the current situation.
　図７は、第１実施形態のプロセス監視支援装置２の状態定義部２１の処理の流れの例を示すフローチャートである。状態定義部２１は、所定の周期Ｔ０のタイミングまでは待機する（ステップＳ１０１－ＮＯ）。所定の周期Ｔ０のタイミングが到来すると（ステップＳ１０１－ＹＥＳ）、過去データ取得部２１１は、過去データを取得する（ステップＳ１０２）。次に、指標定義部２１２は、取得された過去データを用いてプロセス健全性指標を定義する（ステップＳ１０３）。次に、寄与率定義部２１３は、取得された過去データを用いて、各プロセス変数についてプロセス健全性寄与率を定義する（ステップＳ１０４）。そして、常態パターン生成部２１４は、取得された過去データを用いて常態パターンを生成する（ステップＳ１０５）。このようにして得られたプロセス健全性指標、プロセス健全性寄与率及び常態パターンは、その後の状態演算部２２の処理で使用される。 FIG. 7 is a flowchart illustrating an example of a processing flow of the state definition unit 21 of the process monitoring support device 2 according to the first embodiment. The state definition unit 21 waits until the timing of the predetermined cycle T0 (step S101-NO). When the timing of the predetermined cycle T0 comes (step S101-YES), the past data acquisition unit 211 acquires past data (step S102). Next, the index definition unit 212 defines a process soundness index using the acquired past data (step S103). Next, the contribution ratio definition unit 213 defines a process soundness contribution ratio for each process variable using the acquired past data (step S104). Then, the normal pattern generation unit 214 generates a normal pattern using the acquired past data (step S105). The process soundness index, the process soundness contribution rate, and the normal state pattern thus obtained are used in the subsequent processing of the state calculation unit 22.
　図８は、第１実施形態のプロセス監視支援装置２の状態演算部２２及び表示制御部２３の処理の流れの例を示すフローチャートである。状態演算部２２は、所定の周期Ｔ１のタイミングまでは待機する（ステップＳ２０１－ＮＯ）。所定の周期Ｔ１のタイミングが到来すると（ステップＳ２０１－ＹＥＳ）、現在データ取得部２２１は、現在データを取得する（ステップＳ２０２）。次に、指標演算部２２２は、取得された現在データを用いて、指標定義部２１２における最新の定義にしたがってプロセス健全性指標を算出する（ステップＳ２０３）。次に、寄与率演算部２２３は、取得された現在データを用いて、各プロセス変数について定義されたプロセス健全性寄与率の定義にしたがってプロセス健全性寄与率を算出する（ステップＳ２０４）。重要変数抽出部２２４は、各プロセス変数をプロセス健全性寄与率が低い順（異常度の高い順）にソートし、プロセス健全性寄与率が低いものから順に所定の数のプロセス変数を抽出する（ステップＳ２０５）。そして、表示制御部２３は、ステップＳ２０２～Ｓ２０５の処理結果に基づいて表示情報を生成し、ユーザ端末３に出力する（ステップＳ２０６）。 FIG. 8 is a flowchart illustrating an example of a processing flow of the state calculation unit 22 and the display control unit 23 of the process monitoring support device 2 according to the first embodiment. The state calculation unit 22 waits until the timing of the predetermined cycle T1 (Step S201-NO). When the timing of the predetermined cycle T1 comes (step S201-YES), the current data acquisition unit 221 acquires the current data (step S202). Next, the index calculation unit 222 calculates a process health index according to the latest definition in the index definition unit 212 using the acquired current data (step S203). Next, the contribution ratio calculation unit 223 calculates the process soundness contribution ratio according to the definition of the process soundness contribution ratio defined for each process variable using the acquired current data (step S204). The important variable extraction unit 224 sorts each process variable in the order of low process soundness contribution ratio (high order of abnormalities), and extracts a predetermined number of process variables in ascending order of process soundness contribution ratio ( Step S205). Then, the display control unit 23 generates display information based on the processing results of steps S202 to S205, and outputs the display information to the user terminal 3 (step S206).
　このように構成された本実施形態では、ユーザは、「プロセスデータのトレンドグラフによる監視」という従来のプラント監視の延長として、システムによって自動的に判断された「今見るべきプロセス変数」を確認することができる。そのため、プラント監視を効率よく行うことが可能となる。さらに、「今見るべきプロセス変数」の常態パターンが表示される場合には、監視時点までの実測値の時系列データが常態パターンから離れているか否か（異常であるか否か）を容易に判断することが可能となる。 In the present embodiment configured as described above, the user confirms “process variables to be viewed now” automatically determined by the system as an extension of the conventional plant monitoring of “monitoring process data by a trend graph”. be able to. Therefore, plant monitoring can be performed efficiently. Furthermore, when the normal pattern of the “process variable to be viewed now” is displayed, it is easy to determine whether the time-series data of the actually measured values up to the monitoring time point deviates from the normal pattern (whether or not it is abnormal). It is possible to make a judgment.
　［第２実施形態］
　次に、第２実施形態のプロセス監視支援装置２ａについて説明する。図９は、第２実施形態のプロセス監視支援装置２ａの機能構成を示すブロック図である。第２実施形態のプロセス監視支援装置２ａは、状態定義部２１、状態演算部２２及び表示制御部２３に代えて、状態定義部２１ａ、状態演算部２２ａ及び表示制御部２３ａを備える。第２実施形態では、表示制御部２３ａは、ユーザ端末３から表示内容を変更する指示を受け付け、指示に応じて表示を変更する。以下、第２実施形態のプロセス監視支援装置２ａについて説明する。 [Second embodiment]
Next, a process monitoring support device 2a according to a second embodiment will be described. FIG. 9 is a block diagram illustrating a functional configuration of a process monitoring support device 2a according to the second embodiment. The process monitoring support device 2a according to the second embodiment includes a state definition unit 21a, a state calculation unit 22a, and a display control unit 23a instead of the state definition unit 21, the state calculation unit 22, and the display control unit 23. In the second embodiment, the display control unit 23a receives an instruction to change the display content from the user terminal 3, and changes the display according to the instruction. Hereinafter, a process monitoring support device 2a according to the second embodiment will be described.
　第２実施形態における状態定義部２１ａでは、まず常態パターン生成部２１４が各プロセス変数の常態パターンを定義する。次に、寄与率演算部２２３が、常態パターンを用いてプロセス健全性寄与率を定義する。そして、指標定義部２１２がプロセス健全性指標を定義する。以下、各機能について説明する。 In the state definition unit 21a according to the second embodiment, the normal state pattern generation unit 214 first defines the normal state pattern of each process variable. Next, the contribution ratio calculation unit 223 defines the process soundness contribution ratio using the normal state pattern. Then, the index definition unit 212 defines a process soundness index. Hereinafter, each function will be described.
　まず、過去データ取得部２１１が過去所定期間の時系列データＸを取得する。
　次に、常態パターン生成部２１４は、取得された時系列データＸに基づいて常態パターンを生成する。以下では、説明のために、パターン（日変動パターンや週変動パターン）データや位置母数をμｉ（ｔ）、ｉ＝１，２，・・・，ｎと記載する。また、パターンや位置母数の変動を表す尺度母数を、σｉ（ｔ）、ｉ＝１，２，…，ｎと記載する。なお、μ及びσは、必ずしもそれぞれ平均及び標準偏差のみを表しているわけではない。μは、メジアンや刈り込み平均などその他のロバスト性を持った位置母数であってもよい。σは、標準偏差だけでなく、四分位点や、ＭＡＤであってもよい。 First, the past data acquisition unit 211 acquires the time-series data X for the past predetermined period.
Next, the normal pattern generation unit 214 generates a normal pattern based on the acquired time-series data X. In the following, for the sake of explanation, pattern (daily variation pattern or weekly variation pattern) data and position parameters are described as μi (t), i = 1, 2,..., N. Further, a scale parameter representing a variation of a pattern or a position parameter is described as σi (t), i = 1, 2,..., N. Note that μ and σ do not always represent only the average and the standard deviation, respectively. μ may be a position parameter having other robustness such as a median or a pruned average. σ may be not only the standard deviation, but also a quartile or MAD.
　次に，寄与率定義部２１３は、プロセス健全性指標の各プロセス変数のプロセス健全性寄与率を以下の式（６）のように定義する。 Next, the contribution ratio definition unit 213 defines the process soundness contribution ratio of each process variable of the process soundness index as in the following equation (6).
Figure JPOXMLDOC01-appb-M000006
Figure JPOXMLDOC01-appb-M000006
　ここで，ｘｉ（ｔ）は、過去の時系列データＸのｉ列ｔ行目の要素を示す。Ｚｉ（ｘｉ（ｔ））は、ｉ番目のプロセス変数の時刻ｔにおけるプロセス健全性寄与率を示す。監視の際には、このｘｉ（ｔ）が監視時点で計測されたプロセスデータに置き換わって計算される。 Here, xi (t) indicates the element in the i-th column and the t-th row of the past time-series data X. Zi (xi (t)) indicates the process soundness contribution ratio at the time t of the i-th process variable. At the time of monitoring, xi (t) is calculated by replacing the process data measured at the time of monitoring.
　式（６）は、μｉ（ｔ）及びσｉ（ｔ）が時間ｔに依存せずにそれぞれ平均値及び標準偏差で定義される場合には、ＳＰＣ等で用いられるいわゆるＺ値と呼ばれる統計量に対応する。この値は、平均値まわりのばらつきを標準偏差に対する倍率での評価を示す。 Equation (6) indicates that when μi (t) and σi (t) are defined by an average value and a standard deviation, respectively, without depending on time t, a statistic called a Z value used in SPC or the like is obtained. Corresponding. This value indicates the evaluation of the variation around the average value by the magnification relative to the standard deviation.
　次に、指標定義部２１２ａは、式（６）で定義されたプロセス健全性寄与率を統合することによってプロセス健全性指標を定義する。統合する方法はどのような方法であってもよい。例えば、式（６）で計算されたプロセス健全性寄与率の総和又は平均がプロセス健全性指標と定義されてもよい。式（６）で計算されたプロセス健全性寄与率の最大値（すなわち最も異常に近いもの）がプロセス健全性指標として定義されてもよい。 Next, the index definition unit 212a defines a process health index by integrating the process health contribution rates defined by Expression (6). The integration method may be any method. For example, the sum or average of the process soundness contribution rates calculated by equation (6) may be defined as the process soundness index. The maximum value of the process soundness contribution rate calculated by the equation (6) (that is, the one closest to the abnormality) may be defined as the process soundness index.
　判定基準定義部２１５は、プロセスの異常と正常とを判定するための基準を設定する。なお、本実施形態は、異常を検出しアラーム発報することを主眼としたものでは無く、異常及び正常にかかわらずプラント監視の支援を目的としたものである。ただし、支援情報の一つとしてプラントの異常の有無を監視画面上に提示することは好ましいと考えられる。そのため、第２実施形態では、判定基準定義部２１５が異常の有無を示す情報を取得するための基準（以下「判定基準」という。）を設定する。 (4) The criterion definition unit 215 sets a criterion for determining whether a process is abnormal or not. Note that the present embodiment does not focus on detecting an abnormality and issuing an alarm, but aims at supporting plant monitoring regardless of whether the abnormality is normal or not. However, it is considered preferable to present the presence or absence of a plant abnormality on the monitoring screen as one of the support information. For this reason, in the second embodiment, the criterion definition unit 215 sets a criterion (hereinafter, referred to as a “criterion”) for acquiring information indicating the presence or absence of an abnormality.
　判定基準は、プロセス健全性指標とプロセス健全性寄与率に関して定義される。式（６）でプロセス健全性寄与率が定義される場合には、ＳＰＣで用いられる異常判定しきい値の考え方が適用されてもよい。具体的には以下の通りである。通常のＳＰＣでは２σ～３σ程度で管理限界が設定される。そのため、式（６）のしきい値を２～３程度の値で設定し、この設定値以下か設定値以上で異常と判断されてもよい。一方、プロセス健全性指標に対する異常は以下のように定義されてもよい。プロセス健全性指標がプロセス健全性寄与率の平均で定義された場合には、同じしきい値が判定基準として用いられてもよい。プロセス健全性指標がプロセス健全性寄与率の総和で定義された場合には、プロセス健全性寄与率に設定された基準のｎ（プロセス変数の数）倍で設定されてもよい。また、プロセス健全性寄与率の最大値でプロセス健全性指標が定義された場合は、同じしきい値が判定基準として用いられてもよい。このように、ＳＰＣの管理限界の概念が適用されることによって、プロセス健全性指標及びプロセス健全性寄与率の異常判定基準が設定される。 Criteria are defined with respect to the process soundness index and the process soundness contribution rate. When the process soundness contribution rate is defined by Expression (6), the concept of the abnormality determination threshold value used in the SPC may be applied. Specifically, it is as follows. In a normal SPC, a management limit is set at about 2σ to 3σ. Therefore, the threshold value of the equation (6) may be set to a value of about 2 to 3, and an abnormality may be determined to be less than or equal to this set value. On the other hand, the abnormality for the process health index may be defined as follows. If the process health index is defined as the average of the process health contribution rates, the same threshold may be used as a criterion. When the process soundness index is defined by the sum of the process soundness contribution rates, it may be set to n (the number of process variables) times the reference set for the process soundness contribution rates. When the process soundness index is defined by the maximum value of the process soundness contribution rate, the same threshold value may be used as a criterion. As described above, by applying the concept of the control limit of the SPC, the process soundness index and the abnormality judgment criterion of the process soundness contribution ratio are set.
　なお第１実施形態で示されたＭＳＰＣが状態定義部２１ａに適用された場合、異常判定基準には通常のＭＳＰＣにおける異常判定基準が適用されてもよい。すなわち、例えばＱ統計量とＴ＾２統計量に対する以下の式（７）及び式（８）の理論しきい値を用いて異常判定基準が設定されてもよい。式（７）は、Ｑ統計量の理論しきい値である。式（８）は、Ｔ＾２統計量の理論しきい値である。 When the MSPC described in the first embodiment is applied to the state definition unit 21a, an abnormality determination criterion in a normal MSPC may be applied to the abnormality determination criterion. That is, for example, the abnormality determination criterion may be set using the theoretical threshold values of the following equations (7) and (8) for the Q statistic and the T ＾ 2 statistic. Equation (7) is the theoretical threshold of the Q statistic. Equation (8) is the theoretical threshold of the T ＾ 2 statistic.
Figure JPOXMLDOC01-appb-M000007
Figure JPOXMLDOC01-appb-M000007
Figure JPOXMLDOC01-appb-M000008
Figure JPOXMLDOC01-appb-M000008
　式（７）において、ｐはモデルの中に残されたプロセス変数の数である。ｃ＿αは、信頼区間の限界が１－αである場合の標準正規分布の標準偏差のずれである。例えば、α＝０．０１である場合、ｃ＿α＝２．５３である。例えば、α＝０．０５である場合、ｃ＿α＝１．９６である。また、λ＿ｉはΛの対角要素である。つまり、Θ＿ｉは、誤差項に含まれる各成分のｉ乗和である。 Ｐ In equation (7), p is the number of process variables left in the model. c_α is the deviation of the standard deviation of the standard normal distribution when the limit of the confidence interval is 1−α. For example, when α = 0.01, c_α = 2.53. For example, when α = 0.05, c_α = 1.96. Λ_i is a diagonal element of Λ. That is, Θ_i is the sum of the i-th power of each component included in the error term.
　式（８）において、ｐは選択された（モデルの中に残された）プロセス変数の数である。ｍは、全プロセス変数の数である。Ｆ（ｐ，ｍ－ｐ，α）は、自由度が（ｐ，ｍ－ｐ）であり、信頼限界をαとした場合のＦ分布を表す。なお、αには０．０１又は０．０５が用いられることが多い。 Ｐ In equation (8), p is the number of selected process variables (remaining in the model). m is the number of all process variables. F (p, mp, α) represents the F distribution when the degree of freedom is (p, mp) and the confidence limit is α. Note that 0.01 or 0.05 is often used for α.
　プロセス健全性指標に用いられた統計量に応じて式（７）や式（８）を用いて異常判定しきい値を定めることができる。なお、統計量は、Ｑ統計量、Ｔ＾２統計量、Ｑ統計量及びＴ＾２統計量の総和、Ｑ統計量及びＴ＾２統計量の最大値などである。 異常 An abnormality determination threshold value can be determined using Expression (7) or Expression (8) according to the statistics used for the process soundness index. The statistics include the Q statistics, the T ＾ 2 statistics, the sum of the Q statistics and the T ＾ 2 statistics, and the maximum values of the Q statistics and the T ＾ 2 statistics.
　各プロセス変数のプロセス健全性寄与率については、各プロセス変数のプロセス健全性寄与率が統計量に占める割合（プロセス健全性寄与率／統計量）に対して、所定の範囲でのしきい値（例えば、20％～50％）を定め、プロセス健全性寄与率が大きいプロセス変数を異常と判定する基準を設けておくことができる。 Regarding the process soundness contribution rate of each process variable, a threshold value within a predetermined range (process soundness contribution rate / statistics) is defined as the ratio of the process soundness contribution rate of each process variable to the statistics (process soundness contribution rate / statistics). (For example, 20% to 50%), and a criterion for determining a process variable having a large process soundness contribution rate as abnormal can be provided.
　第２実施形態における現在データ取得部２２１、指標演算部２２２、寄与率演算部２２３及び重要変数抽出部２２４の動作は、第１実施形態の同機能部の動作とほぼ同じである。ただし、第１実施形態ではプロセス健全性指標やプロセス健全性寄与率の定義にＭＳＰＣが適用されたのに対し、第２実施形態では典型的なパターンからの乖離でプロセス健全性指標やプロセス健全性寄与率を定義されているため、その計算は異なる。 The operations of the current data acquisition unit 221, the index calculation unit 222, the contribution ratio calculation unit 223, and the important variable extraction unit 224 in the second embodiment are almost the same as the operations of the same function unit in the first embodiment. However, in the first embodiment, the MSPC is applied to the definition of the process soundness index and the process soundness contribution ratio, whereas in the second embodiment, the process soundness index and the process soundness are deviated from typical patterns. The calculation is different because the contribution is defined.
　第２実施形態における状態演算部２２ａは、異常データ抽出部２２５をさらに備える。異常データ抽出部２２５は、指標演算部２２２によって算出されたプロセス健全性指標と、寄与率演算部２２３によって算出された各プロセス監視データのプロセス健全性寄与率と、に対して、判定基準定義部２１５によって定義された基準に基づいて正常か否かを判定する。異常データ抽出部２２５は、判定結果を示すフラグを各値に付加する。例えば、正常の場合は“０”、異常の場合は“１”というフラグが付加されてもよい。 状態 The state calculation unit 22a in the second embodiment further includes an abnormal data extraction unit 225. The abnormal data extraction unit 225 determines the process health index calculated by the index calculation unit 222 and the process health contribution ratio of each process monitoring data calculated by the contribution ratio calculation unit 223 in the determination criterion definition unit. It is determined based on the criteria defined by H.215 whether it is normal. The abnormal data extraction unit 225 adds a flag indicating a determination result to each value. For example, a flag “0” may be added for normal, and a flag “1” for abnormal.
　第２実施形態における表示制御部２３ａは、監視ポイント表示制御部２３３及び寄与率表示制御部２３４をさらに備える。図１０及び図１１は、第２実施形態における表示制御部２３ａの動作によって表示される画面の具体例を示す図である。以下、図１０及び図１１を用いて表示制御部２３ａについて説明する。 The display control unit 23a according to the second embodiment further includes a monitoring point display control unit 233 and a contribution rate display control unit 234. FIGS. 10 and 11 are diagrams illustrating a specific example of a screen displayed by the operation of the display control unit 23a according to the second embodiment. Hereinafter, the display control unit 23a will be described with reference to FIGS.
　図１０に示されるように、画面上段４１には、プロセス健全性指標の時系列データの表示において、異常データ抽出部２２５によって異常というフラグが付加されたデータ（以下「異常データ」という。）についてはその他のデータ（例えば正常というフラグが付加されたデータ：以下「正常データ」という。）とは異なる態様で表示される。異常データは、例えば強調された態様で表示されてもよい。具体的には、異常データは正常データよりも目立つ色（例えば赤色）で表示されても良いし、目立つ線種（例えばより太い線）を用いて表示されてもよい。また、異常データの異常の程度の大きさに応じて異なる態様で表示が行われてもよい。例えば、異常の程度（乖離の程度）が所定の基準よりも大きいものはフリッカ表示されてもよい。また、異常の程度が大きいほどグラフの表示面積が大きく表示されてもよい。本実施形態の図面では、異常データは破線で示されており、正常データは実線で示されている。 As shown in FIG. 10, in the upper part of the screen 41, data (hereinafter, referred to as “abnormal data”) to which a flag of abnormal has been added by the abnormal data extracting unit 225 in displaying the time series data of the process health index. Are displayed in a manner different from other data (for example, data to which a flag of normal is added: hereinafter, referred to as “normal data”). The abnormal data may be displayed, for example, in an emphasized manner. Specifically, the abnormal data may be displayed in a more prominent color (for example, red) than the normal data, or may be displayed using a prominent line type (for example, a thicker line). Further, the display may be performed in a different manner depending on the magnitude of the abnormality of the abnormality data. For example, a flicker display may be performed for a case where the degree of abnormality (degree of deviation) is larger than a predetermined reference. Further, the larger the degree of the abnormality, the larger the display area of the graph may be displayed. In the drawings of the present embodiment, abnormal data is indicated by a broken line, and normal data is indicated by a solid line.
　また、判定基準定義部２１５によって定義されたしきい値がプロセス健全性指標の時系列データにおいて表示されてもよい。例えば、しきい値は異常データ及び正常データとは異なる線種（例えば黄色）で表示されてもよい。 (4) The threshold value defined by the criterion definition unit 215 may be displayed in the time series data of the process health index. For example, the threshold value may be displayed in a line type (for example, yellow) different from the abnormal data and the normal data.
　図１０に示されるように、画面下段４２に表示される各プロセス変数の時系列データの表示においても、上段４１の表示と同様に、異常データと正常データとが異なる態様で表示されてもよい。例えば、異常データが正常データに比べて強調された態様で表示されてもよい。このように強調した態様で表示されることによって、ユーザは異常の発生についてより容易に気がつくことが可能となるとともに、どのプロセス変数のどの日時のデータが異常状態であったかについて視覚的に判断することが可能となる。なお、強調表示がなされる際に音声によるアラームを発報するようにユーザ端末３を制御するように表示制御部２３ａが構成されてもよい。ことも可能であり，音声あるいは文字で同時にアラーム発報を行ってもよい。音声によるアラーム発報が行われる場合には、通常の音量よりも低い音量でアラーム発報が行われてもよい。このように構成されることによって、「異常」を強調しすぎず、アラームの洪水を避けて、ユーザのプラント監視を支援することが可能となる。 As shown in FIG. 10, in the display of the time-series data of each process variable displayed in the lower part 42 of the screen, similarly to the display of the upper part 41, abnormal data and normal data may be displayed in different modes. . For example, abnormal data may be displayed in a manner emphasized compared to normal data. By being displayed in such an emphasized manner, the user can more easily notice the occurrence of the abnormality, and visually determine which data of which process variable and which date was abnormal. Becomes possible. Note that the display control unit 23a may be configured to control the user terminal 3 so as to issue an audio alarm when the highlighted display is performed. It is also possible to simultaneously issue an alarm by voice or text. When the alarm is issued by voice, the alarm may be issued at a volume lower than the normal volume. With such a configuration, it is possible to support the user's plant monitoring without overemphasizing “abnormality” and avoiding flooding of alarms.
　監視ポイント表示制御部２３３は、トレンドグラフのようなプロセス変数の時系列データの表示では無く、プロセスフローのようにプラントに設置された機器を示す表示を制御する。図１１は、監視ポイント表示制御部２３３によって制御される画面の具体例を示す図である。監視ポイント表示制御部２３３は、監視データ表示制御部２３２によって異常データが表示された場合に、異常データに関連する機器の画像の近傍で強調表示を行う。例えば、異常データが取得された機器の近傍において、異常データが取得されたプロセス変数の名称を表示してもよい。このとき、プロセス変数の名称は、監視データ表示制御部２３２によって行われる強調表示と同じ態様（例えば赤色）で表示されてもよい。図１１において、“曝気風量”及び“ＭＬＳＳ”という文字列が、異常データが取得されたプロセス変数の名称の具体例として表示されている。 The monitoring point display control unit 233 controls not a display of time-series data of process variables such as a trend graph, but a display of devices installed in a plant as in a process flow. FIG. 11 is a diagram illustrating a specific example of a screen controlled by the monitoring point display control unit 233. When the monitoring data display control unit 232 displays the abnormal data, the monitoring point display control unit 233 performs highlighting near the image of the device related to the abnormal data. For example, near the device from which the abnormal data was obtained, the name of the process variable from which the abnormal data was obtained may be displayed. At this time, the names of the process variables may be displayed in the same manner (for example, red) as the highlighting performed by the monitoring data display control unit 232. In FIG. 11, the character strings “aeration air volume” and “MLSS” are displayed as specific examples of the name of the process variable from which the abnormal data was obtained.
　また、監視ポイント表示制御部２３３は、異常データが取得された機器の近傍において、この機器の画像又は上述したプロセス変数の名称等の画像に対して、この画像にユーザが注目することを促す画像（以下「注目画像」という。）を表示してもよい。例えば、注目画像の具体例として、機器の画像又はプロセス変数の名称の画像に対して矢が向けられた矢印の画像が表示されてもよい。このとき、注目画像は、監視データ表示制御部２３２によって行われる強調表示と同じ態様（例えば赤色）で表示されてもよい。矢印の画像は注目画像の具体例に過ぎない。例えば、注目画像は“！”等の文字であってもよいし、星印の画像や、一般的に危険であることを示す画像であってもよい。図１１において、“曝気風量”及び“ＭＬＳＳ”という文字列の左に、これらの文字列に矢が向けられた右向きの矢印の画像が注目画像の具体例として表示されている。 In addition, in the vicinity of the device from which the abnormal data was acquired, the monitoring point display control unit 233 displays an image of this device or an image such as the name of the above-described process variable to prompt the user to pay attention to this image. (Hereinafter, referred to as “attention image”). For example, as a specific example of the target image, an image of an arrow in which an arrow is directed to an image of a device or an image of a name of a process variable may be displayed. At this time, the target image may be displayed in the same mode (for example, red) as the highlighting performed by the monitoring data display control unit 232. The image of the arrow is only a specific example of the image of interest. For example, the image of interest may be a character such as "!", An image of a star, or an image indicating that the image is generally dangerous. In FIG. 11, to the left of the character strings “aeration air volume” and “MLSS”, an image of a right-pointing arrow with an arrow directed to these character strings is displayed as a specific example of the image of interest.
　監視ポイント表示制御部２３３によってこのような表示が行われる場合には、異常データが取得されたプロセス変数の時系列データを示す画像がさらに表示されてもよい。図１１の例では、曝気風量及びＭＬＳＳについて異常データが取得されたためにプロセスフローにおいて注目画像などが表示されているが、さらにプロセスフローと供に異常データが取得された曝気風量及びＭＬＳＳの時系列データを示す画像が表示されている。 In the case where such a display is performed by the monitoring point display control unit 233, an image indicating the time-series data of the process variable from which the abnormal data has been acquired may be further displayed. In the example of FIG. 11, an attention image or the like is displayed in the process flow because abnormal data is obtained for the aeration air volume and the MLSS, but the time series of the aeration air volume and the MLSS for which the abnormal data is obtained along with the process flow are further displayed. An image showing the data is displayed.
　また、図１０に示される表示において、異常データが取得された時系列データをユーザが選択した場合に、監視ポイント表示制御部２３３は、選択されたプロセス変数の時系列データに関する機器を含む表示（例えばプロセスフロー）を図１１に示されるように表示してもよい。 Further, in the display shown in FIG. 10, when the user selects the time-series data from which the abnormal data is acquired, the monitoring point display control unit 233 displays the display including the device related to the time-series data of the selected process variable ( For example, the process flow may be displayed as shown in FIG.
　このように時系列データとプロセスフロー上の位置との対応関係をとってプラント監視画面に表示することによって、ユーザは異常が発生した際にその要因を推測することが容易となる。 (4) By displaying the correspondence between the time-series data and the position on the process flow on the plant monitoring screen, the user can easily estimate the cause of the occurrence of the abnormality when it occurs.
　寄与率表示制御部２３４は、各プロセス変数のプロセス健全性寄与率の値が大きいプロセス変数に関する表示を行う。寄与率表示制御部２３４は、例えば寄与率画像４３を生成する。寄与率画像４３には、プロセス健全性寄与率の値を示すグラフ（例えば棒グラフ）が表示される。寄与率画像４３には、例えばプロセス健全性寄与率が高いものから順に所定数のプロセス変数の名称とプロセス健全性寄与率の値とが表示されてもよい。寄与率画像４３には、プロセス健全性寄与率の値が高い順などの基準でソーティングされて棒グラフが表示されることが好ましい。寄与率画像４３の各プロセス変数の並び方は、縦に並んでもよいし横に並んでもよい。 The 率 contribution ratio display control unit 234 displays a process variable having a large value of the process soundness contribution ratio of each process variable. The contribution rate display control unit 234 generates the contribution rate image 43, for example. In the contribution ratio image 43, a graph (for example, a bar graph) showing the value of the process soundness contribution ratio is displayed. The contribution image 43 may display, for example, the names of a predetermined number of process variables and the values of the process health contribution in the descending order of the process soundness contribution. It is preferable that the contribution rate image 43 is displayed as a bar graph sorted in a criterion such as a descending order of the value of the process soundness contribution rate. The arrangement of the process variables in the contribution ratio image 43 may be arranged vertically or horizontally.
　以下、図１０の表示に対するユーザの操作の具体例について説明する。ユーザは、まず図１０のような画像が表示された画面を見て監視を行う。ユーザが、下段４２に表示されている重要プロセス変数のいずれかのプロセスデータの表示データを、他のプロセスデータの表示に変更したい場合、ユーザ端末３の入力装置を操作することによって変更をプロセス監視支援装置２に指示する。例えば、ユーザが８つの監視対象の中の“全体エネルギー原単位”のプロセスデータを他のプロセスデータに変更して表示したい場合、監視画面上に表示されている“全体エネルギー原単位”のプロセスデータの時系列データのグラフを選択する。選択という作業は、マウスやキーボードやタッチパネル等の入力装置に対する操作として行われる。例えば、グラフにカーソルを合わせてクリックするという操作によって選択が行われてもよい。そして、ユーザは寄与率画像４３に表示されたプロセス変数を選択する。監視データ表示制御部２３２は、ユーザによって選択されたプロセス変数の時系列データのグラフを、“全体エネルギー原単位”のプロセスデータが表示されていた領域に表示する。この際、プロセス健全性寄与率の大小が棒グラフで表示されている。そのため、ユーザはプロセス健全性寄与率の大きさを参考にしながら、表示させるプロセスデータを選択することができる。 Hereinafter, a specific example of the user's operation on the display of FIG. 10 will be described. First, the user performs monitoring by viewing a screen on which an image as shown in FIG. 10 is displayed. When the user wants to change the display data of one of the important process variables displayed in the lower part 42 to the display of other process data, the change is monitored by operating the input device of the user terminal 3. Instruct the support device 2. For example, if the user wants to change the process data of “total energy intensity” among the eight monitoring targets to another process data and display the same, the process data of “total energy intensity” displayed on the monitoring screen Select a graph of time series data. The operation of selection is performed as an operation on an input device such as a mouse, a keyboard, and a touch panel. For example, the selection may be performed by an operation of placing the cursor on the graph and clicking the graph. Then, the user selects the process variable displayed on the contribution rate image 43. The monitoring data display control unit 232 displays a graph of the time-series data of the process variable selected by the user in an area where the process data of “total energy intensity” is displayed. At this time, the magnitude of the process soundness contribution ratio is displayed in a bar graph. Therefore, the user can select the process data to be displayed while referring to the magnitude of the process soundness contribution ratio.
　図１２は、第２実施形態のプロセス監視支援装置２ａの状態定義部２１ａの処理の流れの例を示すフローチャートである。状態定義部２１ａは、所定の周期Ｔ０のタイミングまでは待機する（ステップＳ３０１－ＮＯ）。所定の周期Ｔ０のタイミングが到来すると（ステップＳ３０１－ＹＥＳ）、過去データ取得部２１１は、過去データを取得する（ステップＳ３０２）。次に、常態パターン生成部２１４は、取得された過去データを用いて常態パターンを生成する（ステップＳ３０３）。次に、寄与率定義部２１３ａは、取得された常態パターンと過去データとを用いて、各プロセス変数についてプロセス健全性寄与率を定義する（ステップＳ３０４）。そして、指標定義部２１２ａは、算出されたプロセス健全性寄与率を用いてプロセス健全性指標を定義する（ステップＳ３０５）。このようにして得られたプロセス健全性指標、プロセス健全性寄与率及び常態パターンは、その後の状態演算部２２ａの処理で使用される。なお、状態演算部２２ａの処理は、第１実施形態と原則として変わらないため説明を省略する。 FIG. 12 is a flowchart illustrating an example of a processing flow of the state definition unit 21a of the process monitoring support device 2a according to the second embodiment. The state defining unit 21a waits until the timing of the predetermined cycle T0 (step S301-NO). When the timing of the predetermined cycle T0 comes (step S301-YES), the past data acquisition unit 211 acquires past data (step S302). Next, the normal pattern generation unit 214 generates a normal pattern using the acquired past data (step S303). Next, the contribution ratio definition unit 213a defines a process soundness contribution ratio for each process variable using the acquired normal pattern and past data (step S304). Then, the index definition unit 212a defines a process health index using the calculated process health contribution rate (step S305). The process soundness index, the process soundness contribution rate, and the normal state pattern thus obtained are used in the subsequent processing of the state calculation unit 22a. Note that the processing of the state calculation unit 22a is basically the same as that of the first embodiment, and a description thereof will be omitted.
　以上のように構成された第２実施形態の状態定義部２１ａでは、特別な診断アルゴリズムを用いる事なく、単純に常態パターンとの乖離を基準に積み上げ式でプロセス健全性指標が定義される。このような処理は、異常兆候をユーザが監視するという観点においては、健全性の解釈が容易であるという効果がある。すなわち、特別な診断アルゴリズムが用いられていないため、健全度が悪い（異常度が高い）場合には、どこが悪いのか常態パターンと比較することで容易に判断できる。 In the state definition unit 21a of the second embodiment configured as described above, the process soundness index is defined in a stacked manner based on the deviation from the normal state pattern without using a special diagnosis algorithm. Such a process has an effect that the soundness can be easily interpreted from the viewpoint that the user monitors the sign of abnormality. That is, since a special diagnosis algorithm is not used, when the soundness is poor (the degree of abnormality is high), it can be easily determined by comparing the badness with the normal pattern.
　また、第２実施形態の表示制御部２３ａでは、異常データ抽出部２２５によって異常データであると判定されたデータは、通常データとは異なる態様で表示される。異常データは強調表示として表示されてもよい。このように構成されることによって、プラントに異常が発生した場合に、ユーザがその事象に気づきやすいプラント監視を実現することが可能になる In the display control unit 23a of the second embodiment, the data determined to be abnormal data by the abnormal data extraction unit 225 is displayed in a mode different from normal data. The abnormal data may be displayed as highlighting. With this configuration, when an abnormality occurs in the plant, it is possible to realize plant monitoring in which the user can easily notice the event.
　また、第２実施形態の表示制御部２３ａでは、監視ポイント表示制御部２３３によって、異常データに関連する機器の画像の近傍で強調表示が行われる。そのため、プラントに異常が発生した場合に、ユーザはより容易に異常の発生個所や異常の要因を推定することが可能となる。 Also, in the display control unit 23a of the second embodiment, the monitoring point display control unit 233 performs highlighting near the image of the device related to the abnormal data. Therefore, when an abnormality occurs in the plant, the user can more easily estimate the location of the abnormality and the cause of the abnormality.
　また、第２実施形態では、寄与率画像４３が表示される。そのため、寄与率画像４３に基づいて、ユーザが画面に表示させたいトレンドグラフのプロセス変数を選択することが容易に可能となる。そのため、ユーザの意志を反映しやすいプラント監視の実現が可能になる。 で は In the second embodiment, the contribution rate image 43 is displayed. Therefore, based on the contribution rate image 43, the user can easily select a process variable of the trend graph that the user wants to display on the screen. Therefore, plant monitoring that easily reflects the user's intention can be realized.
　［変形例］
　図１３は、第２実施形態の変形例を示す図である。変形例では、第２実施形態においてプロセス監視支援装置２ａとして実装されていた機能の一部が、ネットワーク４を介して離れた位置に設置された情報処理装置（プロセス監視支援サーバ）に実装されている。図１３の例では、プロセス監視支援サーバ９と、プロセス監視支援装置２ｂとがネットワーク４を介して通信可能に接続されている。プロセス監視支援サーバ９及びプロセス監視支援装置２ｂとは、それぞれが備える通信部（通信部９１、通信部２４）が機能することによって通信可能に接続される。第２実施形態における状態定義部２１ａ及び状態演算部２２ａの機能が、プロセス監視支援サーバ９に実装されている。プロセス監視支援装置２ｂのデータ収集部２０１は、所定のタイミングで、データ保存部２０２に記録されているデータをプロセス監視支援サーバ９に送信する。例えば、プロセス監視支援サーバ９の過去データ取得部２１１がデータを要求することを示す要求データを送信してきたタイミングであってもよいし、予め定められた周期のタイミングであってもよい。プロセス監視支援サーバ９の状態定義部２１ａは、過去データを受信すると、受信されたデータに基づいて処理を行う。また、プロセス監視支援装置２ｂのデータ収集部２０１は、所定のタイミングで現在データをプロセス監視支援サーバ９に送信する。プロセス監視支援サーバ９の状態演算部２２ａは、受信された現在データに基づいて処理を行う。状態演算部２２ａは、処理結果を示すデータをプロセス監視支援装置２ｂに送信する。プロセス監視支援装置２ｂの表示制御部２３ａは、受信されたデータに基づいて画面を表すデータを生成し、ユーザ端末３に表示させる。 [Modification]
FIG. 13 is a diagram illustrating a modification of the second embodiment. In the modification, a part of the function implemented as the process monitoring support device 2a in the second embodiment is implemented in an information processing device (process monitoring support server) installed at a remote location via the network 4. I have. In the example of FIG. 13, the process monitoring support server 9 and the process monitoring support device 2b are communicably connected via the network 4. The process monitoring support server 9 and the process monitoring support device 2b are communicably connected to each other by the communication units (the communication unit 91 and the communication unit 24) included therein functioning. The functions of the state definition unit 21a and the state calculation unit 22a in the second embodiment are implemented in the process monitoring support server 9. The data collection unit 201 of the process monitoring support device 2b transmits the data recorded in the data storage unit 202 to the process monitoring support server 9 at a predetermined timing. For example, the timing may be a timing at which the past data acquisition unit 211 of the process monitoring support server 9 transmits request data indicating that data is requested, or a timing of a predetermined cycle. When receiving the past data, the state definition unit 21a of the process monitoring support server 9 performs a process based on the received data. The data collection unit 201 of the process monitoring support device 2b transmits the current data to the process monitoring support server 9 at a predetermined timing. The state calculation unit 22a of the process monitoring support server 9 performs a process based on the received current data. The state calculation unit 22a transmits data indicating the processing result to the process monitoring support device 2b. The display control unit 23a of the process monitoring support device 2b generates data representing a screen based on the received data, and causes the user terminal 3 to display the data.
　変形例の構成は図１３の構成に限定されない。たとえば、プロセス監視支援サーバ９において表示制御部２３ａの機能も実装されてもよい。この場合、データ収集部２０１及びデータ保存部２０２がプロセス監視支援装置２ｂとは異なる装置としてプラントに設置されてもよい。この場合、データ収集部２０１がプロセス監視支援サーバ９と通信することによって、過去データ及び現在データを送信してもよい。この場合、プロセス監視支援装置２ｂそのものがユーザ端末として実装されてもよい。このような実装では、ユーザ端末はスマートフォンやタブレット等の携帯端末装置であってもよい。 構成 The configuration of the modification is not limited to the configuration of FIG. For example, the function of the display control unit 23a may be implemented in the process monitoring support server 9. In this case, the data collection unit 201 and the data storage unit 202 may be installed in the plant as devices different from the process monitoring support device 2b. In this case, the data collection unit 201 may transmit the past data and the current data by communicating with the process monitoring support server 9. In this case, the process monitoring support device 2b itself may be implemented as a user terminal. In such an implementation, the user terminal may be a mobile terminal device such as a smartphone or tablet.
　このように構成されることによって、第１実施形態や第２実施形態において示したプロセス監視支援装置２の機能をクラウドで実装し、ユーザが配置されたプラント等に支援サービスを提供することが可能となる。 With this configuration, the functions of the process monitoring support device 2 described in the first and second embodiments can be implemented in a cloud, and a support service can be provided to a plant or the like where a user is located. Becomes
　状態演算部２２及び状態演算部２２ａは、必ずしも監視時点の時系列データ（現在データ）のみについて処理を行うように設計される必要は無い。例えば、ユーザによって指定された過去の時点の時系列データについて処理を行うように構成されてもよい。このように構成されることによって、過去のプラントの状態を振り返りたい場合に、任意の過去のデータに基づいてプロセス健全性指標やプロセス健全性寄与率の値に基づいた画面をユーザに提供することが可能となる。 The state calculation unit 22 and the state calculation unit 22a do not necessarily need to be designed to process only the time-series data (current data) at the time of monitoring. For example, the processing may be performed on time-series data at a past point in time specified by the user. With this configuration, it is possible to provide a user with a screen based on a process health index or a value of a process health contribution ratio based on arbitrary past data when it is desired to look back on a past state of a plant. Becomes possible.
　以上説明した少なくともひとつの実施形態によれば、表示制御部を持つことにより、現在広く行われているプラント監視と親和性の高いプラント監視方法を維持しながら、アドバンストな監視・診断システムをプラント監視に組み込むことが可能となる。その結果、効率的なプラント監視と、非定常時（異常時等）にその状態をユーザが見落とす可能性を低減させ、より迅速に対応することが可能なプラント監視を実現できる。 According to at least one embodiment described above, by having a display control unit, an advanced monitoring / diagnosis system can be monitored while maintaining a plant monitoring method having high affinity with plant monitoring currently widely performed. It becomes possible to incorporate into. As a result, it is possible to realize efficient plant monitoring and plant monitoring that can reduce the possibility of the user overlooking the state at an abnormal time (for example, at the time of abnormality) and can respond more quickly.
　本発明のいくつかの実施形態を説明したが、これらの実施形態は、例として提示したものであり、発明の範囲を限定することは意図していない。これら実施形態は、その他の様々な形態で実施されることが可能であり、発明の要旨を逸脱しない範囲で、種々の省略、置き換え、変更を行うことができる。これら実施形態やその変形は、発明の範囲や要旨に含まれると同様に、特許請求の範囲に記載された発明とその均等の範囲に含まれるものである。 Although some embodiments of the present invention have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. These embodiments can be implemented in other various forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and equivalents thereof.
１…下水高度処理プロセス、１０１…最初沈澱池、１０２…嫌気槽、１０３…及び無酸素槽、１０４…好気槽、１０５…最終沈澱池、１１１…最初沈澱池余剰汚泥引き抜きポンプ、１１２…ブロワ、１１３…循環ポンプ、１１４…返送汚泥ポンプ、１１５…最終沈澱池余剰汚泥引き抜きポンプ、１２１…雨量センサ、１２２…下水流入量センサ、１２３…流入ＴＮセンサ、１２４…流入ＴＰセンサ、１２５…流入有機物センサ、１２６…ＯＲＰセンサ、１２７…嫌気槽ｐＨセンサ、１２８…無酸素槽ＯＲＰセンサ、１２９…無酸素槽ｐＨセンサ、１３０…リン酸センサ、１３１…ＤＯセンサ、１３２…アンモニアセンサ、１３３…ＭＬＳＳセンサ、１３４…水温センサ、１３５…余剰汚泥ＳＳセンサ、１３６…放流ＳＳセンサ、１３７…汚泥界面センサ、１３８…下水放流量センサ、１３９…放流ＴＮセンサ、１４０…放流ＴＰセンサ、１４１…放流有機物センサ、２…プロセス監視支援装置、２０１…データ収集部、２０２…データ保存部、２１…状態定義部、２２…状態演算部、２３…表示制御部、２１１…過去データ取得部、２１２…指標定義部、２１３…寄与率定義部、２１４…常態パターン生成部、２１５…判定基準定義部、２２１…現在データ取得部、２２２…指標演算部、２２３…寄与率演算部、２２４…重要変数抽出部、２２５…異常データ抽出部、２３１…状態表示制御部、２３２…監視データ表示制御部、２３３…監視ポイント表示制御部、２３４…寄与率表示制御部 DESCRIPTION OF SYMBOLS 1 ... Advanced sewage treatment process, 101 ... First sedimentation tank, 102 ... Anaerobic tank, 103 ... and anoxic tank, 104 ... Aerobic tank, 105 ... Final sedimentation tank, 111 ... First sedimentation tank excess sludge removal pump, 112 ... Blower , 113: circulation pump, 114: return sludge pump, 115: final settling basin excess sludge removal pump, 121: rainfall sensor, 122: sewage inflow sensor, 123 ... inflow TN sensor, 124 ... inflow TP sensor, 125 ... inflow organic matter Sensors 126 ORP sensor 127 Anaerobic tank pH sensor 128 Anoxic tank ORP sensor 129 Anoxic tank pH sensor 130 Phosphoric acid sensor 131 DO sensor 132 Ammonia sensor 133 MLSS sensor , 134: water temperature sensor, 135: surplus sludge SS sensor, 136: discharge SS sensor, 137: sludge Surface sensor, 138 sewage discharge flow sensor, 139 discharge TN sensor, 140 discharge TP sensor, 141 discharge organic substance sensor, 2 process monitoring support device, 201 data collection unit, 202 data storage unit, 21 state Definition unit, 22: State calculation unit, 23: Display control unit, 211: Past data acquisition unit, 212: Index definition unit, 213: Contribution ratio definition unit, 214: Normal pattern generation unit, 215: Determination criterion definition unit, 221 ... Current data acquisition unit, 222 ... Index calculation unit, 223 ... Contribution ratio calculation unit, 224 ... Important variable extraction unit, 225 ... Abnormal data extraction unit, 231 ... Status display control unit, 232 ... Monitoring data display control unit, 233 ... Monitoring point display control unit, 234... Contribution rate display control unit

Claims (12)

  1. 　監視対象プロセスの状態を示すプロセス変数の時系列データを複数種取得するデータ取得部と、
    　複数種の前記プロセス変数に基づいて、前記監視対象プロセスの状態に異常が生じておらず健全である可能性を示す指標を算出する指標演算部と、
    　前記指標演算部によって算出される前記指標に関して、異常が生じておらず健全である可能性が高くなることに寄与した割合を示す寄与率を、複数種の前記プロセス変数毎に算出する寄与率演算部と、
    　前記寄与率の値が相対的に小さい一部のプロセス変数に関する情報をユーザ端末の画面に表示するように表示情報を生成する表示制御部と、
    を備えるプロセス監視支援装置。     A data acquisition unit that acquires a plurality of types of time-series data of a process variable indicating a state of a process to be monitored;
    Based on a plurality of types of the process variables, an index calculating unit that calculates an index indicating a possibility that the state of the monitored process has no abnormality and is healthy,
    For the index calculated by the index calculation unit, a contribution ratio calculation for each of a plurality of types of the process variables, which calculates a contribution ratio indicating a ratio of contributing to a higher possibility of being healthy without abnormality. Department and
    A display control unit that generates display information so as to display information about a part of the process variable in which the value of the contribution ratio is relatively small on the screen of the user terminal,
    A process monitoring support device comprising:
  2. 　前記表示制御部は、前記指標の時系列データをさらに表示するように表示情報を生成する、請求項１に記載のプロセス監視支援装置。 The process monitoring support device according to claim 1, wherein the display control unit generates display information so as to further display the time-series data of the index.
  3. 　前記表示制御部は、前記監視対象プロセスが正常に動作している場合に前記プロセス変数が取り得る値のパターンを示す常態パターンを、各プロセス変数に関する情報とともに表示するように表示情報を生成する、請求項１又は２に記載のプロセス監視支援装置。 The display control unit generates display information to display a normal pattern indicating a pattern of possible values of the process variable when the process to be monitored is operating normally, together with information on each process variable. The process monitoring support device according to claim 1.
  4. 　前記指標演算部は、Ｑ統計量及びＴ＾２統計量のいずれか一方又は双方を用いて前記指標を算出し、
    　前記寄与率演算部は、前記指標の算出に用いられた統計量に対する各プロセス変数の寄与量に基づいて前記寄与率を算出する、請求項１から３のいずれか一項に記載のプロセス監視支援装置。     The index calculation unit calculates the index using one or both of the Q statistic and the T ＾ 2 statistic,
    4. The process monitoring support according to claim 1, wherein the contribution rate calculation unit calculates the contribution rate based on a contribution amount of each process variable to a statistic used for calculating the index. 5. apparatus.
  5. 　前記寄与率演算部は、前記監視対象プロセスが正常に動作している場合に前記プロセス変数が取り得る値のパターンを示す常態パターンと、各プロセス変数の実測値と、の乖離の程度に基づいて前記寄与率を算出し、
    　前記指標演算部は、前記寄与率演算部によって算出された各プロセス変数の寄与率の統計値に基づいて前記指標を算出する、
    請求項１又は２に記載のプロセス監視支援装置。     The contribution rate calculation unit is based on a normal pattern indicating a pattern of values that can be taken by the process variable when the process to be monitored is operating normally, and a measured value of each process variable. Calculating the contribution rate,
    The index calculation unit calculates the index based on the statistical value of the contribution ratio of each process variable calculated by the contribution ratio calculation unit,
    The process monitoring support device according to claim 1.
  6. 　前記表示制御部は、前記常態パターンからの乖離が所定の条件を満たしたプロセス変数のデータを他のデータと異なる態様で表示するように表示情報を生成する、請求項３に記載のプロセス監視支援装置。 4. The process monitoring support according to claim 3, wherein the display control unit generates display information such that data of a process variable whose deviation from the normal pattern satisfies a predetermined condition is displayed in a different form from other data. 5. apparatus.
  7. 　前記表示制御部は、前記監視対象プロセスの機器を含む画像であるフロー図をさらに表示し、前記フロー図において、前記寄与率の値が相対的に大きい一部のプロセス変数に関する前記機器の画像又は前記機器の画像の近傍において、他の機器の画像とは異なる態様の表示を行う、請求項１に記載のプロセス監視支援装置。 The display control unit further displays a flow diagram which is an image including the device of the monitoring target process, and in the flow diagram, the image of the device with respect to some process variables in which the value of the contribution ratio is relatively large or The process monitoring support device according to claim 1, wherein a display different from an image of another device is displayed near an image of the device.
  8. 　前記表示制御部は、前記寄与率の値が相対的に小さい一部のプロセス変数を示す寄与率画像をさらに表示し、前記寄与率画像に含まれる前記プロセス変数に対する操作に応じて、画面に表示されるプロセス変数の種別を変更する、請求項１に記載のプロセス監視支援装置。 The display control unit further displays a contribution image indicating a part of the process variable in which the value of the contribution is relatively small, and displays the contribution image on a screen according to an operation on the process variable included in the contribution image. The process monitoring support device according to claim 1, wherein the type of the process variable is changed.
  9. 　監視対象プロセスの状態を示すプロセス変数の時系列データを複数種取得するデータ取得部と、
    　複数種の前記プロセス変数に基づいて、前記監視対象プロセスの状態に異常が生じておらず健全である可能性を示す指標を算出する指標演算部と、
    　前記指標演算部によって算出される前記指標に関して、異常が生じておらず健全である可能性が高くなることに寄与した割合を示す寄与率を、複数種の前記プロセス変数毎に算出する寄与率演算部と、
    　前記寄与率の値が相対的に小さい一部のプロセス変数に関する情報をユーザ端末の画面に表示するように表示情報を生成する表示制御部と、
    を備えるプロセス監視支援システム。     A data acquisition unit that acquires a plurality of types of time-series data of a process variable indicating a state of a process to be monitored;
    Based on a plurality of types of the process variables, an index calculating unit that calculates an index indicating a possibility that the state of the monitored process has no abnormality and is healthy,
    For the index calculated by the index calculation unit, a contribution ratio calculation for each of a plurality of types of the process variables, which calculates a contribution ratio indicating a ratio of contributing to a higher possibility of being healthy without abnormality. Department and
    A display control unit that generates display information so as to display information about a part of the process variable in which the value of the contribution ratio is relatively small on the screen of the user terminal,
    A process monitoring support system comprising:
  10. 　監視対象プロセスの状態を示すプロセス変数の時系列データを複数種取得し、
    　複数種の前記プロセス変数に基づいて、前記監視対象プロセスの状態に異常が生じておらず健全である可能性を示す指標を算出し、
    　算出される前記指標に関して、異常が生じておらず健全である可能性が高くなることに寄与した割合を示す寄与率を、複数種の前記プロセス変数毎に算出し、
    　前記寄与率の値が相対的に小さい一部のプロセス変数に関する情報をユーザ端末の画面に表示するように表示情報を生成する、
    プロセス監視支援方法。     Acquire multiple types of time-series data of process variables indicating the status of the monitored process,
    Based on the plurality of types of process variables, calculate an index indicating a possibility that the state of the monitored process has no abnormality and is healthy,
    With respect to the calculated index, a contribution ratio indicating a ratio of contributing to a higher possibility of being healthy without an abnormality being calculated is calculated for each of a plurality of types of the process variables,
    Generating display information to display information on some process variables in which the value of the contribution ratio is relatively small on the screen of the user terminal,
    Process monitoring support method.
  11. 　請求項１から８のいずれか一項に記載のプロセス監視支援装置としてコンピュータを機能させるためのプロセス監視支援プログラム。 A process monitoring support program for causing a computer to function as the process monitoring support device according to any one of claims 1 to 8.
  12. 　監視対象プロセスの状態を示すプロセス変数の時系列データを複数種取得するデータ取得部と、複数種の前記プロセス変数に基づいて、前記監視対象プロセスの状態に異常が生じておらず健全である可能性を示す指標を算出する指標演算部と、前記指標演算部によって算出される前記指標に関して、異常が生じておらず健全である可能性が高くなることに寄与した割合を示す寄与率を、複数種の前記プロセス変数毎に算出する寄与率演算部と、前記前記寄与率の値が相対的に小さい一部のプロセス変数に関する情報をユーザ端末の画面に表示するように表示情報を生成する表示制御部と、を備えるプロセス監視支援装置から、前記表示情報を受信する通信部と、
    　前記通信部によって受信された前記表示情報に基づいて画面を表示する表示部と、
    　を備える端末装置。     A data acquisition unit for acquiring a plurality of types of time-series data of process variables indicating the state of the monitored process; and based on the plurality of types of the process variables, the state of the monitored process may be healthy without any abnormality. An index calculating unit for calculating an index indicating the property, and regarding the index calculated by the index calculating unit, a contribution rate indicating a rate of contributing to a higher possibility of being healthy without occurrence of an abnormality, A contribution ratio calculation unit that calculates for each type of the process variable, and a display control that generates display information so as to display information about a part of the process variable in which the value of the contribution ratio is relatively small on a screen of a user terminal. And a communication unit that receives the display information from a process monitoring support apparatus that includes:
    A display unit that displays a screen based on the display information received by the communication unit,
    A terminal device comprising:
PCT/JP2019/036119 2018-09-14 2019-09-13 Process monitoring assistance device, process monitoring assistance system, process monitoring assistance method, process monitoring assistance program, and terminal device WO2020054850A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
MX2021002926A MX2021002926A (en) 2018-09-14 2019-09-13 Process monitoring assistance device, process monitoring assistance system, process monitoring assistance method, process monitoring assistance program, and terminal device.
PH12021550545A PH12021550545A1 (en) 2018-09-14 2021-03-11 Process monitoring support device, process monitoring support system, process monitoring support method, process monitoring support program, and terminal device

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2018173057A JP7163117B2 (en) 2018-09-14 2018-09-14 Process monitoring support device, process monitoring support system, process monitoring support method, and process monitoring support program
JP2018-173057 2018-09-14

Publications (1)

Publication Number Publication Date
WO2020054850A1 true WO2020054850A1 (en) 2020-03-19

Family

ID=69777656

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2019/036119 WO2020054850A1 (en) 2018-09-14 2019-09-13 Process monitoring assistance device, process monitoring assistance system, process monitoring assistance method, process monitoring assistance program, and terminal device

Country Status (4)

Country Link
JP (1) JP7163117B2 (en)
MX (1) MX2021002926A (en)
PH (1) PH12021550545A1 (en)
WO (1) WO2020054850A1 (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008217617A (en) * 2007-03-06 2008-09-18 Toshiba Corp Plant state index management device and computer program for attaining the same
US20090312851A1 (en) * 2005-07-25 2009-12-17 Biogen Idec Ma Inc. System and Method for Bioprocess Control
WO2012090937A1 (en) * 2010-12-28 2012-07-05 株式会社 東芝 Process state monitoring device
JP2013008111A (en) * 2011-06-22 2013-01-10 Hitachi Engineering & Services Co Ltd Abnormality sign diagnostic device and abnormality sign diagnostic method
JP2013093027A (en) * 2011-10-24 2013-05-16 Fisher Rosemount Systems Inc Predicted defect analysis
WO2014073261A1 (en) * 2012-11-09 2014-05-15 株式会社 東芝 Process monitoring/diagnosis device and process monitoring/diagnosis program
JP2017126258A (en) * 2016-01-15 2017-07-20 横河電機株式会社 Monitoring control system and work support method
JP2018120343A (en) * 2017-01-24 2018-08-02 株式会社東芝 Process diagnostic device, process diagnostic method, and process diagnostic system

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090312851A1 (en) * 2005-07-25 2009-12-17 Biogen Idec Ma Inc. System and Method for Bioprocess Control
JP2008217617A (en) * 2007-03-06 2008-09-18 Toshiba Corp Plant state index management device and computer program for attaining the same
WO2012090937A1 (en) * 2010-12-28 2012-07-05 株式会社 東芝 Process state monitoring device
JP2013008111A (en) * 2011-06-22 2013-01-10 Hitachi Engineering & Services Co Ltd Abnormality sign diagnostic device and abnormality sign diagnostic method
JP2013093027A (en) * 2011-10-24 2013-05-16 Fisher Rosemount Systems Inc Predicted defect analysis
WO2014073261A1 (en) * 2012-11-09 2014-05-15 株式会社 東芝 Process monitoring/diagnosis device and process monitoring/diagnosis program
JP2017126258A (en) * 2016-01-15 2017-07-20 横河電機株式会社 Monitoring control system and work support method
JP2018120343A (en) * 2017-01-24 2018-08-02 株式会社東芝 Process diagnostic device, process diagnostic method, and process diagnostic system

Also Published As

Publication number Publication date
MX2021002926A (en) 2021-06-15
PH12021550545A1 (en) 2022-02-14
JP7163117B2 (en) 2022-10-31
JP2020046783A (en) 2020-03-26

Similar Documents

Publication Publication Date Title
JP4762088B2 (en) Process abnormality diagnosis device
JP5996384B2 (en) Process monitoring diagnostic device, process monitoring diagnostic program
JP4468269B2 (en) Process monitoring apparatus and method
JP6214889B2 (en) Process monitoring and diagnosis device
JP5813317B2 (en) Process status monitoring device
JP5259797B2 (en) Learning type process abnormality diagnosis device and operator judgment estimation result collection device
JP5284433B2 (en) Process monitoring / diagnosis / support equipment
JP6862190B2 (en) Process diagnostic device, process diagnostic method and process diagnostic system
CA2762415C (en) System and method for automatic quality control of clinical diagnostic processes
JP2009070071A (en) Learning process abnormality diagnostic device and operator&#39;s judgement estimation result collecting device
JP4604987B2 (en) Operation management method and apparatus for water treatment plant
JP6606003B2 (en) Process diagnostic apparatus, process diagnostic method, and computer program
JP2009199545A (en) Remote monitoring system
CN111768038A (en) Pollutant monitoring method and device, terminal equipment and storage medium
JP2024001315A (en) Process monitoring system, process monitoring method, and program
CN112948761A (en) River nitrogen pollutant quantitative source analysis system
JP4802665B2 (en) Operation management device for water treatment plant
KR100834187B1 (en) Diagnosis system for biological wastewater treatment process using bayesian networks
JP2009178713A (en) Control system for water treatment facility
WO2020054850A1 (en) Process monitoring assistance device, process monitoring assistance system, process monitoring assistance method, process monitoring assistance program, and terminal device
JP2008112428A (en) Method and apparatus for statistically predicting quality of inflow water in water disposal facility
JP2022025364A (en) Process monitoring support device, process monitoring support method, process monitoring support program and terminal device
CN114295162A (en) Environmental monitoring system based on data acquisition
Ivan Fault Detection in Wastewater Treatment: Process Supervision to Improve Wastewater Reuse
Aghaulor et al. An economic design of a modified synthetic double sampling control chart for process monitoring

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19859317

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19859317

Country of ref document: EP

Kind code of ref document: A1