JP2009277118A - Plant monitoring device and plant monitoring method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a plant monitoring device and plant monitoring method for facilitating extraction of variation pattern. <P>SOLUTION: The plant monitoring device comprises: an input section for inputting a data group composed of a plurality of data showing the state of a plant; a selecting section for selecting the data group where at least some of the plurality of data constituting this data group varies from the data group; and a derivation section for deriving periodic and associated variation patterns from the selected data group. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a plant monitoring apparatus and a plant monitoring method for monitoring a plant.

  In a large-scale plant such as a nuclear power plant, a thermal power plant, and a chemical plant, a plant monitoring device that monitors the plant is used. The plant monitoring device monitors the state of the plant based on a signal (plant data) from the measuring instrument and determines whether the state of the plant is normal or abnormal.

Here, there is a possibility that the plant monitoring device may erroneously determine the state of the plant, and it is important to reduce the probability of this erroneous determination. For this reason, the technology of the plant monitoring device has been disclosed that excludes data within the period during which the influence of the test operation of the equipment has occurred from the abnormality detection so that the state change due to the test operation of the equipment is not judged as abnormal. (See Patent Document 1). That is, the amount of change and rate of change of the signal are determined based on the determination condition, and it is determined whether or not it is due to the driving operation.
Japanese Patent Laid-Open No. 11-95833, paragraph 0116

  In the above-described technology, it is necessary to define a change pattern corresponding to a test operation or the like. For this purpose, the actual operation history is compared with the signal change at that time, and a signal change pattern corresponding to the operation is extracted. For example, a change pattern B corresponding to the driving operation A (for example, the signal X increases and the signal Y decreases) is extracted from the history when the driving operation A is performed. By using a logical expression or the like corresponding to the change pattern B and excluding the change pattern B from changes in the plant data, erroneous determinations in the plant monitoring device can be reduced.

  However, since the plant data includes a large number of data, it is not easy to extract a change pattern corresponding to a test operation or the like.

  In view of the above, an object of the present invention is to provide a plant monitoring apparatus and a plant monitoring method that facilitate the extraction of change patterns.

  A plant monitoring apparatus according to an aspect of the present invention includes a plurality of data representing a plant state, an input unit that continuously inputs a set of data, and a set of continuously input data, A selection unit for selecting a data set in which at least one of a plurality of data constituting the data set has changed, and a derivation for deriving a cyclic change pattern corresponding to each other from the selected data set And a determination unit that excludes the derived change pattern from the selected data set and determines whether or not there is an abnormality in the plant.

  A plant monitoring method according to an aspect of the present invention includes a step of continuously inputting a data set composed of a plurality of data representing the state of the plant, and a set of the continuously input data. Selecting a data set in which at least one of a plurality of data constituting the data set has changed, deriving a cyclic change pattern corresponding to each other from the selected data set, and And excluding the derived change pattern from the selected data set to determine whether there is an abnormality in the plant.

  ADVANTAGE OF THE INVENTION According to this invention, the plant monitoring apparatus and plant monitoring method which make easy extraction of a change pattern can be provided.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a block diagram showing a plant monitoring system 100 according to an embodiment of the present invention. The plant monitoring system 100 includes a plant 110, a measuring instrument 120, and a plant monitoring device 130.

  The plant 110 is a plant such as a nuclear power plant, a thermal power plant, or a chemical plant, and is a monitoring target of the plant monitoring device 130.

  The measuring device 120 is for monitoring the state of the plant 110 and outputs a signal representing the state of the plant 110 (for example, temperature, pressure, flow rate). Various measuring instruments 120 are installed at a plurality of locations in the plant 110.

  The plant monitoring device 130 is a device that monitors the plant 110, and includes an input unit 131, a data storage unit 132, a data processing unit 133, a history data storage unit 134, a determination formula creation unit 135, a monitoring processing unit 136, and a history data providing unit. 137, and a determination formula editing unit 138.

  The input unit 131 inputs various signals (plant data) from the measuring instrument 120 at a constant cycle (a sampling cycle dt described later). Plant data (signal from the measuring device 120) is a set of data necessary for plant monitoring, and is composed of a plurality of data representing the state of the plant.

  The data storage unit 132 stores the plant data input by the input unit 131 in chronological order. The stored plant data is used by the data processing unit 133 and the monitoring processing unit 136.

  The data processing unit 133 obtains the change and generation time of each data in the plant data (data set), and selects a data set in which at least one of a plurality of data constituting the data set has changed. To do. The data processing unit 133 sequentially reads plant data stored in the data storage unit 132, detects the occurrence of a change (increase / decrease) in data (signal), and stores history data as history data along with the time. The data is stored in the unit 134. There are several methods for detecting this change. Hereinafter, change detection methods 1 to 3 will be described.

(1) Change detection method 1
Changes in data can be detected by comparing averaged values with instantaneous values. For example, as shown in the following equation (1), the data (signal) sig is averaged using a predetermined time constant tc, and a deviation “sig− between the average value v and the data (instantaneous value) sig is shown. By providing a threshold value (lower limit, upper limit (−Va, Vb)) for “v”, occurrence of a change can be detected. For example, it is determined that the data value decreases when “sig−v <−Va”, and the data value increases when “sig−v> Vb”.

v = v + wt (sig (i) -v) (1)
sig (i): signal (data)
v: Filter (averaging) processing value wt (= MAX (1 / w, 1 / i)): Weight w = tc / dt
tc: Time constant dt: Sampling cycle i: Number of data points

Hereinafter, it will be described that the expression (1) means a kind of weighted averaging process for the data sig. Here, if “wt = 1 / i” is assumed, Expression (1) becomes “v = v + (sig (i) −v) / i”.
When i = 1, v = sig (1)
When i = 2, v = sig (1) + (sig (2) −sig (1)) / 2
= (Sig (1) + sig (2)) / 2
……………………………………………………………………………………
When i = n, v = (sig (1) +... + sig (n)) / n (2)
Expression (2) represents that the filter processing value v is an average value of the data sig. If the number of data points i is in the range of “1 / w <1 / i” (i <w), “wt = 1 / i”, and the filter processing value v is the average value of the data sig itself.

  If the number of data points i exceeds the range of “1 / w <1 / i” (i ≧ w), the weight wt is kept constant as “wt = 1 / w”. That is, if “wt = 1 / i”, the weight wt decreases and becomes constant. Thus, Expression (1) represents an averaging process in which the weight wt to the data sig close to the current time is increased.

  As described above, equation (1) means a weighted averaging process in which data sig is emphasized in w ranges (range of time constant tc) from the current time (when instantaneous value sig is measured). For example, if the time constant tc is 1 minute and the sampling period dt is 0.1 second, w = 600, and data sig is averaged with emphasis on 600 data within about 1 minute from the present time. means. The reason for performing such weighted average processing is to detect plant abnormalities based on data in the range of the time constant tc. That is, the abnormality of the plant 110 is detected ignoring the data fluctuation (long-term or instantaneous fluctuation) in a time sufficiently larger or smaller than the time constant tc. In other words, the time constant tc that makes it easy to detect an abnormality in the plant 110 is set. Note that the time constant tc may be changed according to the type of abnormality.

(2) Change detection method 2
A change in data may be detected only by the instantaneous value sig. Depending on the signal, the averaging process as shown in equation (1) is not necessary. For example, for a signal that does not change if the normal state is maintained during plant monitoring, the change is detected by comparing the values measured sequentially with the upper and lower threshold values without averaging. May be.

(3) Change detection method 3
A change in data may be detected based on the maximum value and the minimum value of the instantaneous value sig within a predetermined time (for example, a time width corresponding to the test period). That is, the maximum value and the minimum value are respectively compared with threshold values, and a change in data within a predetermined time is detected.

  In abnormality monitoring, if such a process is performed, a time delay occurs in detecting the abnormality. However, in order to detect data changes caused by regular tests, it is only necessary to detect data changes that occur between the start and end of the test. For this reason, data can be divided by a time width corresponding to such a test period.

  The history data storage unit 134 stores a change history of plant data. As shown in FIG. 2, in the history data storage unit 134, when a change is detected by the data processing unit 133, the direction of the change (positive / negative (increase / decrease) of the change) and the time at which the change occurred are stored.

The judgment formula creation unit 135 creates a judgment formula (judgment condition, signal list) for judging whether or not there is a driving operation. The determination formula creation unit 135 functions as the following (1) and (2). Details of the determination formula creation unit 135 will be described later.
(1) Deriving unit for deriving periodic change patterns corresponding to each other from the selected data set (2) Creation unit for creating a judgment formula for excluding derived change patterns

  The monitoring processor 136 determines whether the plant 100 is normal or abnormal. Prior to this determination, the monitoring processing unit 136 determines whether or not there is a driving operation using the determination formula created by the determination formula creation unit 135. When a driving operation is detected, a signal that may change during the driving operation is bypassed with reference to the signal list (determination formula) created by the judgment formula creation unit 135, and an abnormality is judged using the remaining signals Is made. That is, the monitoring processing unit 136 functions as a determination unit that excludes the derived change pattern from the selected data set and determines whether there is an abnormality in the plant 110. As an abnormality determination method, the above-described equation (1) can be used. However, other methods may be used for abnormality determination.

(Details of judgment formula creation unit 135)
Hereinafter, details of the determination formula creation unit 135 will be described. The judgment formula creation unit 135 searches for a change similar to the past when new data is stored in the history data storage unit 134. As shown in the following (1) to (3), the judgment formula creation unit 135 extracts a change pattern having the same range (change pattern is almost constant) and periodicity (occurrence interval is almost constant). . This change pattern can be represented by changed data and a combination of positive and negative changes.

(1) Change simultaneity Summarize changed data in consideration of change simultaneity. That is, considering that there is a slight time lag in the data change, the data considered to have started changing almost simultaneously are collected as one change pattern.

  Consider a case where a change history as shown in FIG. 2 is stored in the history data storage unit 134. In FIG. 2, the increase in signal SA and the decrease in signal SB are expressed as changing at the same time t1 (or time t2). However, it is also conceivable that the change times t1 of the signals SA and SB do not completely coincide with each other due to the change propagation time and the like. Therefore, in order to determine that the change is due to the same test, the time required for the test is taken into account, and an estimated time for determining the change due to the same test is set. Then, a change detected within a certain time set as an estimated time from the first change of any signal is determined as a change due to the same operation. That is, when the signal SB changes within a predetermined time from when the signal SA first changes, it is handled that the changes in the signals SA and SB occur simultaneously. Further, the change time of the signal SA is defined as the change time of the signals SA and SB. In this way, change patterns in which signals (data) have changed can be collected within the range of change simultaneity.

(2) Identity of change pattern Ideally, the same change occurs in the same data (signal) if the same driving operation is performed (the change pattern is the same). However, there are cases where the magnitude of the disturbance is slightly different for each test, or there is a case where a change that should be detected is not detected due to a change determination threshold value. For this reason, it is necessary to devise a device that can be regarded as the same driving operation even if the change patterns do not completely match. That is, when the changing signal and the direction of the change (increase / decrease) match a predetermined number (for example, two) or more, the change pattern is in the same range (possibility of change due to the same driving operation Is high).

  This will be described based on the example of FIG. In this case, the change pattern is the same at time t1 and time t2, but the change pattern is different at time t3. That is, a decrease in the signal SC is added. At these three times t1, t2, and t3, an increase in signal SA and a decrease in signal SB are common. Thus, since there are always change patterns of the two signals SA and SB, it is determined that the change pattern is the identity range.

(3) Periodicity of change Here, the intended operation is mainly a regular test, and the change occurs at regular intervals. This will be described based on the example of FIG. Here, if the difference between time t1 and time t2 and the difference between time t2 and time t3 match within a certain range, it is determined that the three changes have periodicity.

  As a result, for example, change pattern 1 (increase in signal SA and decrease in signal SB) and change pattern 2 (increase in signal SA, decrease in signal SB, and decrease in signal SC) are in the same range and periodic. It is extracted as a characteristic change pattern.

  Based on the extracted change pattern, a judgment formula (signal list) is created. For example, when change pattern 1 (increase in signal SA and decrease in signal SB) and change pattern 2 (increase in signal SA, decrease in signal SB, and decrease in signal SC) are extracted, this change pattern is corresponded. Thus, a judgment formula (signal list) for the actual operation is created. In this way, a list of signals in which changes are seen is created and used to define signals to be bypassed in the monitoring process. FIG. 3 shows an example of this signal list. Here, basically, based on the change pattern of “increase in signal SA, decrease in signal SB”, and in addition to this, the change pattern in which signal SC changes, it is determined whether or not it is a driving operation.

  In this way, it is possible to monitor plant abnormalities except for signal changes due to human operations performed in regular tests. Thereby, the occurrence frequency of erroneous determination can be reduced, and the reliability of the plant monitoring device 130 can be improved.

  The history data providing unit 137 provides the operator with history data in the case where the determination result of the change pattern identity and periodicity is intermediate. It functions as a data providing unit that provides a data set in which the positive and negative changes coincide within a predetermined range.

  As described above, the judgment formula creation unit 135 automatically creates a judgment formula for judging a regular test from the signal change pattern stored in the history data storage unit 134. However, there may be cases where it takes a long time to determine that a regular test is performed for a long time, or when the test interval is not constant for some reason and cannot be determined as a regular test. Therefore, the change history data stored in the history data storage unit 134 is presented to the operator, and the operator confirms the change due to the execution of the regular test by collating with the test results. Thereby, even in a situation where automatic determination by the history data providing unit 137 is difficult, it is possible to correctly determine the driving operation by interposing the operator's determination.

The judgment formula editing unit 138 enables editing of the judgment formula created by the judgment formula creation unit 135 and functions as an editing unit for editing the judgment formula.
The judgment formula creation unit 135 automatically creates a judgment formula for determining a regular test from the signal change pattern stored in the history data storage unit 134. This judgment formula is a logical product of changes in a plurality of signals. This is because they do not have knowledge of the causal relationship between the test and signal change. Therefore, there is a possibility of making a duplicate determination when detecting a change. For example, when the signal SA and the signal SD both show a change that increases when a certain test is performed, if the two signals are multiplexed signals, the change of the signal SA and the signal SD is described by a logical sum. Is natural. Even if one of the signal SA and the signal SD does not change, the test execution can be correctly determined. Considering such points, the judgment formula automatically created by the judgment formula creation unit 135 can be appropriately edited by the operator, so that the plant monitoring device 130 can make a more accurate judgment.

(Operation procedure of the plant monitoring system 100)
FIG. 4 is a flowchart showing an example of the operation procedure of the plant monitoring system 100.
(1) Data input (step S11)
A set of data (including a plurality of data) is continuously input from the measuring instrument 120 to the input unit 131 at a constant period (sampling period dt). The input data sets are stored in chronological order.

(2) Selection of changed data (step S12)
The data processing unit 133 processes a plurality of data as necessary, and obtains the presence / absence of data change and the sign of the change. For example, the data (instantaneous value sig) is averaged based on equation (1), and the presence or absence of data change is detected based on whether or not the difference between the instantaneous value sig and the average value v is within a predetermined range. For example, the presence / absence of data change is detected based on whether the maximum value and minimum value of data (instantaneous value sig) within a predetermined time are within a predetermined range.

  With respect to data determined to have changed, the sign of the change and the time when the change occurred are stored in the history data storage unit 134 as history data.

(3) Derivation of change pattern (step S13)
The judgment formula creation unit 135 extracts a change pattern of data having identity and periodicity from the history data. Prior to this, data considered to have started changing almost simultaneously may be collected as one change pattern.

(4) Creation of judgment formula (step S14)
The judgment formula creation unit 135 creates a judgment formula based on the extracted change pattern. The history data providing unit 137 provides the history data to the worker as necessary.

(5) Editing judgment formula (step S15)
Using the judgment formula editing unit 138, the worker edits the judgment formula.

(6) Plant monitoring (step S16)
The plant monitoring device 130 monitors the plant 110 using the edited determination formula.

(Other embodiments)
Embodiments of the present invention are not limited to the above-described embodiments, and can be expanded and modified. The expanded and modified embodiments are also included in the technical scope of the present invention.

It is a block diagram showing the plant monitoring system concerning one embodiment of the present invention. It is a schematic diagram showing an example of the history data preserve | saved at a history data preservation | save part. It is a schematic diagram showing an example of the list | wrist of the signal extracted by the determination formula preparation part. It is a flowchart showing an example of the operation | movement procedure of a plant monitoring system.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100 ... Plant monitoring system, 110 ... Plant, 120 ... Measuring device, 130 ... Plant monitoring apparatus, 131 ... Input part, 132 ... Data storage part, 133 ... Data processing part, 134 ... History data storage part, 135 ... Creation of judgment formula Part, 136 ... monitoring processing part

Claims (9)

An input unit for inputting a data set composed of a plurality of data representing the state of the plant;
A selection unit for selecting, from the data set, a data set in which at least one of a plurality of data constituting the data set has changed;
A derivation unit for deriving a periodic change pattern corresponding to each other from the selected data set;
A determination unit that excludes the derived change pattern from the selected set of data and determines whether there is an abnormality in the plant;
A plant monitoring apparatus comprising: The derivation unit is
An extraction unit for extracting a data set from which the changed data and at least a part of the positive and negative of the change coincide with each other from the selected data set;
A determination unit for determining presence or absence of periodicity of change in the extracted data set;
An output unit that outputs changed data and a positive / negative combination of the changes in the set of changes determined to have periodicity in the change as the derived change pattern;
The plant monitoring apparatus according to claim 1. 3. The data set according to claim 1, wherein the selection unit selects a data set based on whether or not a difference between an average value and an instantaneous value of each of the plurality of data is within a predetermined range. The plant monitoring device described in 1. The plant monitoring apparatus according to claim 3, wherein the averaged value is a value based on a weighted average that places importance on data that is temporally close to the instantaneous value. 3. The data sorting unit according to claim 1, wherein the selection unit selects a data set based on whether or not a maximum value and a minimum value within a predetermined time of each of the plurality of data are within a predetermined range. The plant monitoring device described in 1. A creation unit for creating a judgment formula for excluding the derived change pattern;
The determination unit excludes the derived change pattern from the selected data set using the determination formula;
The plant monitoring device according to any one of claims 1 to 5, wherein An editing section for editing the judgment formula;
The plant monitoring apparatus according to claim 6, further comprising: A data providing unit for providing a data set in which the changed data and the sign of the change coincide in a predetermined range from the selected data set;
The plant monitoring apparatus according to any one of claims 1 to 7, further comprising: Inputting a data set composed of a plurality of data representing the state of the plant;
Selecting, from the data set, a data set in which at least one of a plurality of data constituting the data set has changed;
Deriving corresponding and periodic change patterns from the selected set of data;
Excluding the derived change pattern from the selected set of data and determining whether there is an abnormality in the plant;
The plant monitoring method characterized by comprising.

Images