JP2011253491A - Plant abnormality detector, method for the plant abnormality detector, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a plant abnormality detector, a method for the plant abnormality detector, and program for preventing false detection.SOLUTION: The plant abnormality detector includes: a reference value decision part for deciding a reference value for detecting abnormality of a plant on the basis of a statistical value of first plant data to which moving average processing has been performed; a second moving average processing part for performing moving average processing to second plant data; and an abnormality detection part for detecting abnormality of the plant by comparing the second plant data to which the moving average processing has performed with the stored reference value.

Description

  Embodiments described herein relate generally to a plant abnormality detection device, a plant abnormality detection device method, and a program.

  A plant abnormality detection device that detects abnormality of a plant (a power plant, a chemical plant, etc.) is used. In the plant abnormality detection device, a measurement value (plant data) from a sensor installed in the plant is compared with a reference value to detect a plant abnormality or its sign.

Japanese Patent No. 2772179

By the way, plant data may fluctuate due to various factors, and there is a possibility that a plant abnormality is erroneously detected due to this fluctuation.
An object of the present invention is to provide a plant abnormality detection device, a plant abnormality detection device method, and a program that prevent erroneous detection due to variations in plant data.

  The plant abnormality detection device according to the embodiment includes a reference value determination unit that determines a reference value for detecting a plant abnormality based on a statistical value of the first plant data that has been subjected to moving average processing, and second plant data. A second moving average processing unit that performs a moving average process, a second plant data that has been subjected to the moving average process, and an abnormality detection unit that detects an abnormality of the plant by comparing the stored reference value; , Provided.

1 is a block diagram showing a plant abnormality detection system 1 according to a first embodiment. It is a flowchart showing an example of the operation | movement procedure of the plant abnormality detection system. It is a schematic diagram which shows the content of a moving average process. It is a schematic diagram which shows the content of the double moving average process. It is a schematic diagram which shows the example of the determination conditions defined in the plant condition file. 6 is a diagram illustrating an example of a display state on a display unit 111. FIG. 6 is a diagram illustrating an example of a display state on a display unit 111. FIG. 6 is a diagram illustrating an example of a display state on a display unit 111. FIG. It is a block diagram which shows the plant abnormality detection system 1a which concerns on 2nd Embodiment. It is a schematic diagram showing the object period in the case of updating a reference value. It is a figure which shows the example of a display of the fluctuation | variation of the moving average value A (t) and T (i) on the display part 111. FIG. It is a block diagram which shows the plant abnormality detection system 1b which concerns on 3rd Embodiment. It is a schematic diagram which shows an example of the combination of sampling period (DELTA) t and the number n of data to average. It is a schematic diagram which shows an example of the plant data with the large fluctuation rate y. It is a schematic diagram which shows an example of plant data with the fluctuation rate y intermediate. It is a schematic diagram which shows an example of plant data with small fluctuation rate y. It is a block diagram which shows the plant abnormality detection system 1c which concerns on 4th Embodiment. It is a figure which shows the example which overlapped and displayed the some reference value in the graph format. It is a block diagram which shows the plant abnormality detection system 1d which concerns on 5th Embodiment. It is a figure which shows the example which displayed the moving average value with the graph of the time-axis of logarithm expression. It is a figure which shows the example which displayed the moving average value with the graph of the time-axis of logarithm expression. It is a block diagram which shows the plant abnormality detection system 1e which concerns on 6th Embodiment. It is a figure showing designated period AC. It is a figure which shows the example which displayed the reference value Ta based on the moving average value of the period A, the moving average value Ab (t) of the period B, and the moving average value Ac (t) of the period C superimposed.

Hereinafter, embodiments will be described in detail with reference to the drawings.
(First embodiment)
FIG. 1 is a block diagram showing a plant abnormality detection system 1 according to the first embodiment. The plant abnormality detection system 1 includes a plant 10 and a plant abnormality detection device 20. The plant 10 is a plant such as a power plant or a chemical plant, for example. The plant abnormality detection device 20 is a device that monitors the operation state of the plant and detects an abnormality, and includes a plant data input unit 102, a plant data history file unit 103, a moving average processing unit 104, a moving average data file unit 105, a reference. A value calculation unit 106, a reference value file unit 107, a plant condition recording file unit 108, a plant state determination unit 109, an abnormality detection unit 110, a display unit 111, an external notification unit 116, a voice output unit 117, and an electronic mail unit 118 are included.

  The plant abnormality detection system 1 can be configured by a combination of a computer (CPU (Central Processing Unit) and peripheral devices (input device, output device, main storage device, auxiliary storage device)) and a program.

  The plant data input unit 102 sequentially inputs plant data (measured values) from various sensors (detectors that detect the state of the plant (for example, current, voltage, pressure, temperature)) installed in the plant 10. The plant data input unit 102 functions as an input unit that sequentially inputs the second plant data.

  The plant data history file unit 103 sequentially stores the plant data input by the plant data input unit 102 as a history.

The moving average processing unit 104 performs a moving average process on the plant data stored in the plant data history file unit 103. The moving average processing unit 104 functions as a “first moving average processing unit that performs a moving average process on first plant data” and a “second moving average processing unit that performs a moving average process on second plant data”.
The moving average data file unit 105 sequentially stores the moving average processed plant data.

The reference value calculation unit 106 calculates a reference value for detecting a plant abnormality based on the moving averaged plant data. The reference value calculation unit 106 functions as a reference value determination unit that determines a reference value for detecting a plant abnormality based on the statistical value of the first plant data that has been subjected to the moving average process.
The reference value file unit 107 stores the reference value calculated by the reference value calculation unit 106. The reference value file unit 107 functions as a reference value storage unit that stores the determined reference value.

The plant condition record file unit 108 stores conditions for determining the state of the plant (for example, the plant is in stable operation, is being stopped, etc.).
The plant state determination unit 109 determines the state of the plant from the conditions stored in the plant condition record file unit 108 and the current plant data. The plant state determination unit 109 functions as a plant state determination unit that determines the state of the plant based on the second plant data.

  The abnormality detection unit 110 determines whether there is a plant abnormality based on the moving average processed plant data, the reference value, and the determination result in the plant state determination unit 109. The abnormality detection unit 110 functions as an abnormality detection unit that detects the abnormality of the plant by comparing the second plant data subjected to the moving average process with the stored reference value.

The display unit 111 is a display device that displays the detection result of the abnormality detection unit 110, for example, a liquid crystal display.
When the abnormality detection unit 110 detects an abnormality, the external notification unit 116 notifies the outside via the voice output unit 117, the electronic mail unit 118, and the like.

(Operation of the plant abnormality detection system 1)
Hereinafter, the operation of the plant abnormality detection system 1 will be described. FIG. 2 shows an example of the operation procedure of the plant abnormality detection system 1.

  Hereinafter, plant abnormality detection (steps S2 to S6) and determination (step S1) of the reference value Ta (upper limit reference value Ta1, addition / subtraction reference value Ta0) prior to this detection are classified into the plant abnormality detection system 1. The operation will be described.

A. Plant abnormality detection (1) Input of plant data (step S2)
The plant data input unit 102 takes in (inputs) plant data (measured values by sensors) from the plant 10, adds time information, and stores and stores the information in the plant data history file unit 103. As described above, this input is performed sequentially (for example, periodically).

(2) Moving average processing of plant data (step S3)
The moving average processing unit 104 takes out the plant data from the plant data history file unit 103, performs the moving average process, and saves and accumulates the moving average processed plant data in the moving average data file unit 105.
Note that it is not necessary to perform a moving average process on all plant data. It is sufficient to perform moving average processing only on plant data from a predetermined abnormality detection sensor.

FIG. 3 shows the contents of the moving average process.
In the moving average process, the moving average value A _i is sequentially generated from the measured value X _i of the plant data as follows.

1) the plant data history file unit 103, the n-number of plant data (measured _value) X 1 to X _n are removed at the sampling period Delta] t, and the average value and the moving average value _{A n.} That is, as shown in the following equation, plant data X _{1 to} X _n are added, and a result obtained by dividing by the number n is set as a moving average value A ₁ .
_{A n = (X 1 + X} 2 + X 3 + ····· X n) / n
The calculated moving average value _An is added with data at a time corresponding to the measured value X _n and stored in the moving average data file unit 105.

2) Next, after time (sampling period) Δt, n pieces of plant data (measured values) X _{2 to} X _{n + 1} are taken out at the sampling period Δt, and the average value thereof is set as a moving average value _{An + 1} . That is, as shown in the following equation, plant data X _{2 to} X _{n + 1} are added, and a result obtained by dividing by the number n is set as a moving average value _{An + 1} .
A _{n + 1} = (X ₂ + X ₃ +... X _n + X _{n + 1} ) / n
The calculated moving average value _{An + 1} is added to the time data corresponding to the input time of the plant data _{Xn + 1} , and stored in the moving average data file unit 105.

3) Thereafter, similarly, after time Δt, n pieces of plant data (measured values) X _{3 to} X _{n + 2} are taken out at the sampling period Δt, and the average value is set as the moving average value _{An + 2} .
A _{n + 2} = (X ₃ +... X _n + X _{n + 1} + X _{n + 2} ) / n

  As described above, for each sampling period Δt, the plant data to be processed is shifted and the average value is calculated (moving average). This is a moving average process.

  The sampling period Δt and the number n of data to be averaged are determined as appropriate. Details of this will be described later.

Here, the moving average processing unit 104 may further perform moving average processing on the moving average processed plant data (moving average value A _i ) to calculate a double moving average value AA _i (double moving average). processing).

FIG. 4 shows the contents of the double moving average process.
In the double moving average process, double moving average values AAi (AA _n , AA _{n + 1} , AA _{n + 2} ,...) Are sequentially generated from the moving average values Ai as follows.
AA _n = (A ₁ + A ₂ + A ₃ +... A _n ) / n
AA _{n + 1} = (A ₂ + A ₃ +... A _n + A _{n + 1} ) / n
AA _{n + 2} = (A ₃ +... A _n + A _{n + 1} + A _{n + 2} ) / n

In this case, the double moving average value AA _i is stored in the moving average data file unit 105 instead of the moving average value A _i . That is, a plant abnormality can be detected using the double moving average value AA _i instead of the moving average value A _i . By using the double moving average value AA _i , it is possible to more reliably cope with plant data fluctuations than when the moving average value A _i is used.

(3) Plant state determination (step S4)
The plant state determination unit 109 determines the state of the plant. In other words, the plant state determination unit 109 extracts from the plant condition record file unit 108 the determination conditions necessary for determining the plant state (for example, during stable plant operation and during start / stop). In addition, the plant state determination unit 109 extracts the latest plant data from the plant data history file unit 103. The plant state determination unit 109 determines the current plant state from the latest plant data and determination conditions. The plant state determination unit 109 notifies the abnormality detection unit 110 of the determination result.

FIG. 5 shows an example of determination conditions defined in the plant condition file. Here, a condition for determining whether or not the plant is in stable operation (rated load) is shown.
Plant data from a predetermined sensor (point ID sensor) is taken out from the plant data history file unit 103, and it is determined whether the condition is satisfied or not. In this example, when the generator output is 480 to 520 [MW] and the load is in the ON state, it is determined that stable operation is being performed (rated load).

(4) Plant abnormality detection (step S5)
The abnormality detection unit 110 compares the moving average value A _i of the plant data with a reference value Ta (upper limit reference value Ta1, addition / subtraction reference value Ta0) for determining whether the plant 10 is abnormal, and whether there is an abnormality in the plant 10 Determine. If the moving average value of greater than _{A i} is an upper limit reference value Ta1 _(A i> Ta1), when the moving average value Ai is lower reference value Ta0 smaller than _(A i <Ta0), it is determined that an abnormality has occurred in the plant . In other words, when the moving average value A _i is not less than the lower limit reference value Ta0 and not more than the upper limit reference value Ta1, it is determined that no abnormality has occurred in the plant.

  The reference value Ta (upper limit reference value Ta1, addition / subtraction reference value Ta0) is extracted from the reference value file unit 107. A method for calculating the reference value Ta will be described later.

Here, the plant state information determined by the plant state determination unit 109 can be used as in the following 1) to 3).
1) Switching of the reference value Ta By switching the reference value Ta according to the state of the plant, the plant abnormality can be detected appropriately. In this case, the reference value file unit 107 stores a plurality of reference values Ta corresponding to different plant states.

2) Execution and stop switching of plant abnormality detection It is conceivable that a plant abnormality is detected in a specific plant state, and a plant abnormality is not detected in other plant states. For example, plant abnormalities are detected only during stable operation (rated load).

3) Temporal reference of reference value Ta (t) As will be described later, the reference value Ta can be a function of time t instead of a constant value (reference value Ta (t)). When the state of the plant changes dynamically (for example, when shifting from a stopped state to a stable operation state), if an occurrence of an abnormality in the plant is to be detected, the reference value Ta may be a function of time t. preferable. In such a case, an event in the plant (for example, start-up of the plant) is detected from a change in the state of the plant, and this detection time can be used as a temporal reference for the reference value Ta (t).

(5) Notification of abnormality detection (step S6)
The abnormality detection unit 110 displays the detection result on the display unit 111. Further, when a plant abnormality is detected, the abnormality detection unit 110 notifies the external notification unit 116. The external notification unit 116 notifies the outside that a plant abnormality has occurred by using the voice output unit 117, the electronic mail unit 118, or the like. The e-mail unit 118 transmits an e-mail to an administrator address registered in advance.

B. Determination of reference value Ta (upper limit reference value Ta1, adjustment reference value Ta0) (step S1)
The reference value calculation unit 106 takes out the moving average value A _i of the plant data from the moving average data file unit 105 and performs statistical processing to determine the reference value Ta (upper limit reference value Ta1, addition / subtraction reference value Ta0).

Specifically, the average value a and the standard deviation σ of the moving average value A _i can be obtained by the following equation. An average value a and a standard deviation σ are calculated from the n1 moving average values A _i .
a = (ΣA _i ) / n1
σ = (√ (Σ (A _i −a) ² )) / n1

A reference value width (upper reference value Ta1 and lower limit reference value Ta0) is determined based on the obtained average value a and standard deviation σ.
Ta1 = a + m · σ
Ta0 = a−m · σ
m: Predetermined constant (1 to 3 etc. are appropriately set according to the degree of fluctuation of the plant 10)

  The determined reference value Ta (upper limit reference value Ta1 and lower limit reference value Ta0) is stored in the reference value file unit 107. The stored reference value Ta is used for detecting plant abnormalities as described above.

Here, the moving average value A _i corresponding to a specific plant state is used to determine the reference value Ta. That is, the reference value calculation unit 106 extracts plant data corresponding to a specific plant state based on the plant state determination condition extracted from the plant condition record file unit 108. Then, a reference value Ta is determined using the moving average value A _i of the extracted plant data. This is because the plant abnormality is detected using the reference value Ta corresponding to the plant state.

Instead of the moving average value A _i, using a double moving average value AA _i, when detecting an abnormality of the plant, by statistically processing a double moving average value AA _i, the reference value Taa is determined The

C. The display state of the moving average value A display of _i and the reference value Ta then the moving average value A _i and the reference value Ta is described.
FIG. 6 shows an example of the display state on the display unit 111. Moving average value A (t), measured value X (t), moving average value A (t), reference value Ta corresponding to each of measured value X (t) (upper limit reference value Ta1, lower limit reference value Ta0), Tx (Upper limit reference value Tx1, lower limit reference value Tx0) are displayed in an overlapping manner.

  When the measured value X (t) fluctuates temporarily (range P in FIG. 6), when the measured value X (t) is compared with the reference value Tx, a temporary abnormality is detected. On the other hand, when the moving average value A (t) is used, the influence of temporary fluctuation is reduced, and erroneous detection of abnormality is prevented.

  FIG. 7 shows an example of a display state on the display unit 111. Here, it is determined that an abnormality has occurred when the moving average value A (t) falls below the lower limit reference value Ta0 and then a certain time α has passed. Thus, it can be determined that an abnormality has occurred when the moving average value A (t) deviates from the reference value Ta for a certain time α. The time α can be appropriately defined from 0 minute to n minutes for each plant data (sensor).

As described above, the reference value Ta (upper limit reference value Ta1, lower limit reference value Ta0) can be a function of time t (Ta (t)). FIG. 8 shows an example of a display state on the display unit 111.
Here, the presence / absence of an abnormality in the plant is detected until a time β elapses from a specific event (for example, ignition event) (monitoring period). In this case, the reference value Ta (t) (the upper limit reference value Ta1 (t), the lower limit reference value Ta0 (t)) is statistically processed from the past moving average value A0 (t) starting from this specific event. , Calculated. When the current moving average value A (t) deviates from the reference value Ta (t) during the monitoring period, it is determined that an abnormality has occurred.

  In the present embodiment, after moving average processing of past plant data, statistical processing is performed to determine a reference value Ta (reference range). Then, the presence or absence of abnormality is detected by comparing the moving average value A (t) of the current plant data with the reference value Ta. By performing moving average processing on plant data, noise components are removed from the plant data and false detection of abnormalities is reduced.

(Second Embodiment)
FIG. 9 is a block diagram showing a plant abnormality detection system 1a according to the second embodiment. A reference value recalculation unit 112 is added, and a reference value Ta necessary for detecting plant abnormality is periodically determined and updated.

  The reference value recalculation unit 112 outputs a reference value update instruction to the reference value calculation unit 106. This update instruction includes a target period (one month ago, six months ago, one year ago, etc.) of the reference value calculation as a parameter (update parameter).

Based on this update parameter, the reference value calculation unit 106 extracts the moving average value A _i within the target period from the moving average data file unit 105, determines the reference value Ta (upper limit reference value and lower limit reference value), Saved in the value file unit 107. The reference value recalculation unit 112 outputs a reference value update instruction to the reference value calculation unit 106 at an update cycle (1 day, 1 week, 1 month, 3 months, 6 months, etc.). As a result, the reference value is updated to the latest value.

At this time, for example, in the following cases 1) and 2), the reference value is not updated.
1) When the state of the plant within the target period is not the detection target (when there is no plant data to be detected)
2) When the number of corresponding plant data is less than the specified value (number of data to be averaged n)

  10 and 11 show specific examples in which the reference value is updated every day. As shown in FIG. 10, the reference value T (i) (upper limit reference value T1 (i), lower limit reference value T0 (i)) is calculated daily from the moving average value A (t) for the past month starting from the current time. Is done. In addition, as shown in FIG. 11, fluctuations in the moving average values A (t) and T (i) are displayed on the display unit 111.

  In the present embodiment, the reference value T (i) is periodically updated. As a result, for example, it is possible to cope with changing plant conditions due to environmental temperature (atmosphere, seawater, etc.) or deterioration over time, and erroneous detection of plant abnormality is reduced.

(Third embodiment)
FIG. 12 is a block diagram showing a plant abnormality detection system 1b according to the third embodiment. In the plant abnormality detection system 1 according to the first embodiment, a moving average number calculation unit 113 for calculating an optimal average number (number n of data to be averaged) for moving average plant data is added.

  The fluctuation of the plant data varies depending on the state of the plant (during stable plant operation and start / stop) and the type of plant data (sensor type (temperature, pressure, etc.)). For this reason, the sampling period Δt and the number n of data to be averaged are adjusted according to the state of the plant and the type of plant data. For example, as shown in FIG. 13, a combination of the sampling period Δt and the number n of data to be averaged is determined.

Hereinafter, a method for determining the sampling period Δt and the number n of data to be averaged will be described.
(1) Determination of sampling period Δt The sampling period Δt is determined based on the state of the plant (when the plant is stable or when it is started and stopped) and the type of plant data (type of sensor). For example, when the plant data is stable and the type of plant data is current or voltage, the sampling period Δt is 5 seconds, and when the type of plant data is temperature, the sampling period Δt is 60 seconds. The current / voltage changes quickly and the temperature corresponds to a relatively slow change.

  On the other hand, when the plant starts and stops, the fluctuation of plant data becomes larger than when the plant is stable. For this reason, even a temperature sensor determines a fast sampling period Δt, such as 10 seconds.

The above is summarized as 1) and 2) below.
1) Sampling cycle (Δt) when the plant is stable
5 seconds: current, voltage 30 seconds: pressure 60 seconds: temperature

2) Sampling cycle (Δt) at plant start / stop
5 seconds (10 seconds also possible): Current, voltage, pressure, temperature

(2) Calculation of fluctuation rate (change rate) y of plant data The fluctuation rate y of plant data is calculated. The moving average number calculation unit 113 takes out past plant data from the plant data history file unit 103 at the determined sampling period Δt. The moving average number calculating unit 113 calculates a plurality (k) of change amounts ΔX _i of the plant data to obtain an average change amount ΔXa.

ΔX _i = X _{i + 1} −X _i
ΔXa = ΣX _i / k

Further, the fluctuation rate y of the plant data is calculated from the average change amount ΔXa by the following equation.
y = (ΔXa / (UL)) * 100
U: Upper limit of plant data (sensor upper range)
L: Lower limit of plant data (sensor lower limit range)

  Thus, the fluctuation rate y is a percentage of the average change amount ΔXa with respect to the sensor range width of “0-100%”.

  FIG. 14A to FIG. 14C show examples of plant data having different fluctuation rates y. The plant data in FIG. 14A has a large variation rate y, and the plant data in FIG. 14C has a small variation rate y. The plant data in FIG. 14B has an intermediate variation rate y of the plant data in FIGS. 14A and 14C.

(3) Determination of the number of averaged data (average parameter) n from the variation rate y The number n of averaged data is determined from the variation rate y. For example, the number n is determined in the range of 10 to 60 so that the number n increases if the variation rate y is large. For example, the relationship between the fluctuation rate y and the number n can be determined as in the following 1) to 4).

1) Fluctuation rate y: 0 to 1% (case 1), number n: 10 2) Fluctuation rate y: 1 to 1.5% (case 2), number n: 20 3) Fluctuation rate y: 1.5 to 2% (case 3), number n: 30 4) Fluctuation rate y: 2% or more (case 4), number n: 60

  The relationship between the fluctuation rate y and the number of data n can be determined as appropriate. For plant data that changes quickly and has large fluctuations, the sampling cycle is shortened and the number n of data to be averaged is increased. For plant data that changes slowly and has little fluctuation, increase the sampling period and decrease the number of data to be averaged.

(4) Notification of Sampling Period Δt and Number of Data to be Averaged n The moving average number calculation unit 113 notifies the moving average processing unit 104 of the sampling period Δt and the number of data to be averaged n.

Based on the notified sampling period Δt and the number of data (average parameter) n, the moving average processing unit 104 calculates the moving average value A _i and stores it in the moving average data file unit 105.

  In the present embodiment, the parameters for calculating the moving average value (sampling period Δt, number of data n) can be tuned according to the magnitude of fluctuations in plant data. As a result, it is possible to calculate a moving average value that matches the characteristics of the plant data and improve the accuracy of plant abnormality detection.

(Fourth embodiment)
FIG. 15 is a block diagram showing a plant abnormality detection system 1c according to the fourth embodiment. In the plant abnormality detection system 1a according to the second embodiment, a multiple reference value display control unit 114 that superimposes and displays a reference value obtained from a plurality of periods, current plant data (measured value), and a moving average value is added. ing. The multiple reference value display control unit 114 is a first display control unit that displays a plurality of reference values that are periodically determined by the reference value determination unit and the second plant data that has been subjected to the moving average process. Function as.

  As described above, the reference value recalculation unit 112 and the reference value calculation unit 106 calculate a plurality of reference values starting from the current time for each sensor and store them in the reference value file unit 107. For example, a reference value T1 (upper limit reference value T11, lower limit reference value T10) from the moving average value of the nearest month before and a reference value T2 (upper limit reference value T21, lower limit reference) from the moving average value up to one year ago. Value T20).

  The multiple reference value display control unit 114 determines the abnormality detection by the abnormality detection unit 110, a plurality of reference values T from the reference value file unit 107, and the moving average value A from the moving average data file unit 105 to the present. , The measured value X up to the present is taken out from the plant data history file unit 103. The multiple reference value display control unit 114 causes the display unit 111 to display the extracted data in a graph format.

  FIG. 16 shows a specific example in which a plurality of reference values are superimposed and displayed in a graph format. Reference value T1 (upper limit reference value T11, lower limit reference value T10) from the moving average value one month ago and reference value T2 (upper limit reference value T21, lower limit reference value T20) from the moving average value up to one year ago, The moving average value A (t) and the measured value X (t) are superimposed and displayed.

  In this embodiment, a plurality of reference values T1 and T2 corresponding to plant data of a plurality of periods such as a short period and a long period are calculated, and a plant abnormality is detected using these reference values T1 and T2. As a result, it is possible to detect a plurality of failure factors such as aged deterioration and partial failure.

(Fifth embodiment)
FIG. 17 is a block diagram showing a plant abnormality detection system 1d according to the fifth embodiment. In the plant abnormality detection system 1 according to the first embodiment, a logarithmic graph display control unit 115 that superimposes and displays the reference value and the moving average value from the past to the present in the long-term period is added. The logarithmic graph display control unit 115 functions as a second display control unit that displays the second plant data that has been subjected to the moving average processing, using a graph having a time axis as a logarithmic display.

The logarithmic graph display control unit 115 receives the determination result of the abnormality detection from the abnormality detection unit 110, and extracts the reference value from the reference value file unit 107 and the moving average value from the past to the present from the moving average data file unit 105.
The logarithmic graph display control unit 115 displays these moving average values on the display unit 111 as a logarithmic expression time axis graph. 18 and 19 show the display state of the display unit at this time. In FIG. 18, the moving average value from 100 days ago to the present is displayed. In FIG. 19, the moving average value from 1000 days ago to the present is displayed. In this way, it is easy to grasp the temporal variation of the moving average value by switching and displaying the scale of the time axis.

  In the present embodiment, a logarithmized time axis is displayed together with the reference value for the moving average data. As a result, it is possible to easily grasp the degree of deterioration over time leading to plant abnormality detection or before abnormality detection.

(Sixth embodiment)
FIG. 20 is a block diagram showing a plant abnormality detection system 1e according to the sixth embodiment. A reference value evaluation unit 119 having a function for comparing and confirming the created reference value and the moving average value from the past to the present of the specified period is added to the plant abnormality detection system 1 according to the first embodiment. ing.

  In order to evaluate the validity of the reference value created by the reference value calculation unit 106, the reference value evaluation unit 119 designates a plurality of accumulated moving average value periods (at the start time and at the end time). Notify the detection unit 110.

  The abnormality detection unit 110 extracts the moving average value for the specified period from the moving average data file unit 105, extracts the specified reference value from the reference value file unit 107, and causes the display unit 111 to display it.

  FIG. 21 shows designated periods A to C. FIG. 22 shows the display state of the display unit 111. As shown in FIG. 22, a reference value Ta (upper limit reference value Ta1, lower limit reference value Ta0) based on the moving average value of period A, moving average value Ab (t) of period B, moving average value Ac (t ) Is superimposed and displayed.

  It is possible to determine whether the past data (moving average values Ab (t), Ac (t)) for several times when the plant is healthy is within the reference value Ta. When there is a case where the reference value deviates from Ta, the reference value calculation unit 106 can recalculate the reference value or manually correct the reference value.

  In the present embodiment, the validity of the reference value can be confirmed in advance from a plurality of past data before starting the daily operation with the established reference value.

(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.

  For example, the reference value recalculation unit 112 of the second embodiment, the moving average number calculation unit 113 of the third embodiment, the multiple reference value display control unit 114 of the fourth embodiment, the logarithmic graph of the fifth embodiment A plant abnormality detection device can be configured by appropriately combining all or part of the display control unit 115 and the reference value evaluation unit 119 of the sixth embodiment.

  DESCRIPTION OF SYMBOLS 1 ... Plant abnormality detection system, 10 ... Plant, 20 ... Plant abnormality detection apparatus, 102 ... Plant data input part, 103 ... Plant data history file part, 104 ... Moving average process part, 105 ... Moving average data file part, 106 ... Reference value calculation unit 107 ... Reference value file unit 108 108 Plant condition record file unit 109 109 Plant condition determination unit 110 110 Abnormality detection unit 111 Display unit 112 Reference value recalculation unit 113 Moving average Number calculation unit, 114 ... multiple reference value display control unit, 115 ... logarithmic graph display control unit, 116 ... external notification unit, 117 ... voice output unit, 118 ... e-mail unit, 119 ... reference value evaluation unit

Claims (8)

A first moving average processing unit for moving average processing the first plant data;
A reference value determining unit for determining a reference value for detecting an abnormality of the plant based on the statistical value of the first plant data subjected to the moving average process;
A reference value storage unit for storing the determined reference value;
An input unit for sequentially inputting second plant data;
A second moving average processing unit that performs a moving average process on the second plant data;
An abnormality detection unit that detects the abnormality of the plant by comparing the second plant data that has been subjected to the moving average processing and the stored reference value;
A plant abnormality detection device comprising: A plant state determination unit that determines the state of the plant based on the second plant data;
The reference value storage unit stores a plurality of reference values;
The abnormality detection unit switches the plurality of reference values based on the result of the determination to detect an abnormality in the plant.
The plant abnormality detection device according to claim 1. The reference value determining unit periodically determines and changes the reference value based on the second plant data sequentially input;
The reference value storage unit stores the changed reference value;
The abnormality detection unit detects the abnormality of the plant by comparing the second plant data subjected to the moving average process and the changed reference value;
The plant abnormality detection apparatus according to claim 1 or 2. The 1st display control part which superimposes and displays the some reference value periodically determined by the said reference value determination part, and the 2nd plant data by which the moving average process was carried out is provided. The plant abnormality detection device according to any one of the above. The input unit inputs the second plant data at a predetermined cycle,
A number determining unit for determining the number of plant data to be subjected to moving average processing based on a variation rate of the second plant data in the predetermined cycle;
The plant abnormality detection device according to any one of claims 1 to 4, wherein the first and second moving average processing units respectively perform a moving average process on the determined number of first and second plant data. . A second display control unit for displaying the second plant data that has been subjected to the moving average process, using a graph with a logarithmic display of the time axis;
The plant abnormality detection device according to any one of claims 1 to 5, further comprising: A process of moving average processing the first plant data;
Determining a reference value for detecting an abnormality of the plant based on a statistical value of the first plant data subjected to the moving average process;
Sequentially inputting second plant data;
A process of moving average processing the second plant data;
Comparing the second plant data subjected to the moving average process with the stored reference value to detect an abnormality of the plant;
A plant abnormality detection method comprising:   The program for functioning a computer as a plant abnormality detection apparatus of any one of Claims 1 thru | or 6.

Images