JPH06274784A - Plant monitoring diagnostic device and its abnormal sign identification method - Google Patents

Abstract

PURPOSE:To identify the abnormal sign of plant and its type and contents by identifying a pattern showing the abnormal sign in the change pattern of the measurement signals, set signals and control signals fluctuated in the normal fluctuation area of the measurement signal by matching. CONSTITUTION:Plural types of change pattern in time series showing the abnormal sign of a plant 30 and a plant controller 20 are set in the normal fluctuation area of the measurement signal corresponding to the normal operation range of a plant 30. The measurement signal and the set and control signals of the plant controller 20 are inputted in time series to a plant monitoring diagnosis device 10 which judges whether the change patterns shown by the measurement signal, set signal and control signal coincide with any set change patterns. Thus, the abnormal signs can be detected during the plant 30 is operated in the normal range. The type of abnormal signs can be identified from the change pattern contents.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a plant monitoring / diagnosing apparatus for monitoring an operating state of a plant within a normal operating range to detect an abnormal sign and a method for identifying an abnormal sign.

[0002]

2. Description of the Related Art There is known a plant control device which measures a physical property value indicating an operating state of a plant, for example, temperature or pressure, and automatically sets a manipulated variable of the plant based on the measurement result and a control instruction of the physical property value to be controlled. ing. In such a plant control device, it has become possible to execute from the measurement result to the diagnosis of the abnormality and the details of the abnormality. Although various proposals have been made for the abnormality detection method used in the plant control device, they have the following common points. That is, statistical processing is performed on the measurement signal to create a characteristic parameter that represents the operating state of the plant, and when the parameter value is within a predetermined allowable range, it is determined to be normal, and when the parameter value exceeds the allowable range. This is a point to judge as abnormal. Among such abnormality detection methods, those for the purpose of early detection of abnormality are disclosed in, for example, JP-A-4-3057.
There is No. 94 etc. In this proposal, the standard deviation of the command signal and
An abnormality detection method is disclosed in which a difference from the standard deviation of a measurement signal is calculated, and when the difference exceeds a certain value, it is determined that an abnormality has occurred.

In many plants, a limit check method is adopted in which a measurement signal indicating the state of the plant is input to a plant control device, and when the measurement signal deviates from a predetermined reference value, it is determined to be abnormal.

When an abnormality is detected by the plant control device,
It is important to take appropriate measures against the abnormality.
In many plants, especially chemical plants, in order to give the highest priority to safety, an interlock system that automatically shuts down an apparatus or system in which an abnormality is detected is incorporated.

In the case of a large-scale chemical plant, the shutdown of one plant may induce the shutdown of a plurality of plants connected by a pipeline, which requires extremely high reliability and safety. It is required to detect a symptom of abnormality before detecting the abnormality that the interlock system operates, and to investigate the cause and take corrective action.

Therefore, the applicant of the present application detects the abnormal sign by determining that the plant state is located in the abnormal sign detection area provided in the normal range while the plant is operating in the normal range by the measurement signal. At the same time as filing the application of the present application, a plant monitoring / diagnosing device that does this is proposed.

[0007]

However, there are various change patterns of the measurement signal toward the abnormal area, as shown in FIGS. 2, 3, and 4, and the cause of the abnormality is different depending on the change pattern. Therefore, there is room for improvement in the above proposal in terms of identifying the types of abnormal signs.

In view of the above points, an object of the present invention is to provide a plant monitoring and diagnosing device capable of identifying a change pattern of an abnormality sign from a measurement signal and a setting signal and an operation signal of the plant control device, and an abnormality thereof. It is to provide a symptom identification method.

In order to achieve such an object, the invention of claim 1 is a time-series system showing an abnormal sign of a plant and a plant control device within a normal fluctuation region of a measurement signal corresponding to a normal operating range of the plant. A plurality of types of change patterns are defined in advance, the measurement signal, the setting signal of the plant control device, and the operation signal are input in time series to the plant monitoring and diagnosis device, and the measurement signal, the setting signal, and the change pattern of the operation signal. It is characterized in that the plant monitoring and diagnosing device discriminates which one of the plurality of types matches the above, thereby identifying the abnormal sign of the plant and its type content.

According to a second aspect of the present invention, the change pattern is further accompanied by information indicating the degree of urgency up to the occurrence of an abnormality, and the urgency is indicated in the identification result of the abnormality sign of the plant and the plant control device. It is characterized by adding information.

According to a third aspect of the present invention, a plurality of types of time-series change patterns that indicate abnormal signs of the plant and the plant control device are stored in a normal fluctuation region of the measurement signal corresponding to the normal operating range of the plant. Means, input means for time-sequentially inputting the measurement signal and the setting signal and operation signal of the plant control device, and the change pattern that matches the change pattern indicated by the input measurement signal, setting signal, and operation signal And a pattern detecting means for outputting the extraction result as the type content of the abnormality sign.

[0012]

According to the first and third aspects of the present invention, a pattern indicating an abnormal sign is identified by pattern matching from the change patterns of the measurement signal, the setting signal, and the operation signal that fluctuate within the normal fluctuation region of the measurement signal.

According to the second aspect of the present invention, urgency level information is added to the change pattern to notify the user of the urgency level to prepare for the occurrence of an abnormality.

[0014]

Embodiments of the present invention will now be described in detail with reference to the drawings.

The principle of the abnormal sign identifying method of the present invention will be described.

FIG. 1 shows the areas provided for the detection of changes in the measurement signal and abnormal signs. In FIG. 1, S is
The control setting level that is the control target of the state quantity of the plant, for example, the temperature is shown. XH and XL are a first upper limit threshold value and a first lower limit threshold value for detecting abnormal signs, which are fixed threshold values.

VH and VL are a second upper limit threshold value and a second lower limit threshold value for detecting abnormal signs, which are fluctuation threshold values. As is well known, the fluctuation threshold value is obtained by adding / subtracting a constant value to the average (called a moving average) of the measured values from the present time to a fixed period before. YH and YL are an upper limit threshold and a lower limit threshold for abnormality detection. Also, level Y
Areas above H and below level YH are called abnormal areas
-XH and levels YL-XL are called high-level sudden change regions. Further, the area between the levels XH and XL will be referred to as a low level sudden change area.

In this embodiment, the measured value is the fluctuation threshold V.
It is considered that an abnormal sign has occurred when H, VL or XH, XL is exceeded, and the position of the measured value at this time is located in the low level sudden change region or the high level sudden change region. I do.

Further, the change direction of the measurement signal is discriminated based on the magnitude of two adjacent measurement values, and the change pattern of the measurement signal is shown in FIGS. 1 to 4 based on the result of the position discrimination and the result of the change direction discrimination. Identify which one corresponds.

For this purpose, the characteristic parameter D described below
V. TRS, DV. The ABS is calculated, and the change pattern is determined by selectively combining these characteristic parameters. DV. The variables used in the calculation of TRS and their formulas are as follows.

[0021]

[Equation 1] Deviation DV (t) at time t = Measured value PV (t) at time t −
Control set value SV (t) at time t

[0022]

[Formula 2] Upper fluctuation threshold value VH (t) at time t = moving average of measured values at time t + constant value

[0023]

## EQU00003 ## Lower fluctuation threshold VL (t) at time t = moving average value of measured values at time t-constant value Deviation DV (t) at time (t) and fluctuation threshold VH
(T) and VL (t), the state value DV.
TRS is

[0024]

[Equation 4] DV.TRS (t) = If VL (t) ≤ DV (t) ≤ VH (t) NOR (no abnormal sign) If VH (t) <DV (t) UP (low level) Abnormal sign of rise) If VL (t)> DV (t), it can be obtained from the logical relational expression of DOWN (abnormal sign of low level drop). As can be seen from (Equation 4), the characteristic parameter DV.TRS is a parameter that indicates that the measured value exceeds the range between the fluctuation thresholds and, if it exceeds, the direction. DV. When TRS becomes UP or DOWN, it is determined that an abnormal sign has occurred.

Characteristic parameter DV. ABS is calculated from the following mathematical formula.

[0026]

[Equation 5] DV.ABS (t) = If XH ≥ DV (t) ≥ XL, NOR
(No abnormal sign) If XH <DV (t) HIGH (abnormal sign of high level rise) If XL> DV (t) LOW (abnormal sign of low level fall) Here, XH and XL are abnormal signs It is the first threshold value for detection (see FIG. 1). DV. When ABS (t) becomes HIGH or LOW, it indicates that the measured value is located in the high level sudden change region of FIG.

FIG. 5A shows the relationship between the characteristic parameters and the deviation state (change pattern of the present invention) between the measured values. Therefore, the values indicated by the characteristic parameters are shown in FIG.
It is possible to detect whether the deviation state of the measured values matches any change pattern from normal to high level sudden change by determining what kind of combination (a) is.

Next, a plant control system for detecting and diagnosing abnormal signs by using the above-described method for identifying abnormal sign types will be described. The circuit configuration of the plant control system is shown in FIG.

In FIG. 6, the plant 30 receives an operation control signal from the plant controller (DCS) 20 and is operated with a predetermined operation amount. State quantities such as temperature and pressure (hereinafter referred to as control quantities) indicating the operating state of the plant 30 are measured by a measuring instrument, and a measurement signal indicating the measurement result is DCS2.
It is output to 0. The DCS 20 variably sets the manipulated variable based on the measurement signal so that the controlled variable fluctuates along a predetermined response characteristic (including a fixed set value). The plant monitoring / diagnosing device 10 operates as an abnormal sign detecting device, an abnormal sign diagnostic device, and an abnormal detecting device in the plant control device.

In the plant monitoring / diagnosing device 10, the following components are connected to a common bus.

Central processing unit (CPU) 11: executes abnormality sign detection processing, abnormality sign diagnosis processing, and abnormality detection processing based on the system program stored in the system memory 12. As will be described later, the CPU 11 operates as an abnormal sign pattern identifying means of the present invention.

System memory 12: A system program that defines the processing procedure executed by the CPU 11 is stored for each processing content. It also temporarily stores work data related to arithmetic processing. A program for detecting an anomaly according to the present invention and a combination of the change pattern of the measured value and the value of the characteristic parameter shown in FIG. 5A are also stored in the system memory 12, and the system memory 12 corresponds to the storage means of the present invention. To do.

Keyboard 13: Inputs operation instructions to the CPU 11.

Display device 14: Displays input information from the keyboard and calculation results of the CPU 11 under the control of the CPU 11.

Communication interface (I / F) 15: C
Signals are transferred between the PU 11 and the DCS 20 as an input means of the present invention. A control amount measurement signal, a setting signal, and an operation signal are received from the DCS 20, and a new setting signal or a new operation signal is transmitted to the DCS 20.

External storage device 16: Stores plant measurement results and abnormality sign diagnosis results in a floppy disk (FD) or hard disk (HD).

The abnormality sign identification related processing executed in such a circuit configuration will be described with reference to FIG. Figure 7 is C
The processing procedure of the abnormality detection processing and the abnormality symptom identification processing executed by the PU 11 is shown, and is stored in the system memory 12 in advance in the form of a programming language. This processing procedure is executed by the CPU 11 by interrupt processing at regular intervals.

At a certain sampling time, the CPU 1
1 inputs the plant measurement signal, the setting signal, and the operation signal via the I / F 15 and stores them in the external storage device 16 (S10). Next, the CPU 11 performs a calculation using the equations 1 to 5 and the characteristic parameter DV. TRS, DV. AB
The value of S is obtained (S20).

Next, the CPU 11 obtains the deviation state (that is, the change pattern) corresponding to the combination of the values of the two characteristic parameters from the table of FIG. 5A (S3).
0). The cause of failure of the abnormal sign corresponding to the identified change pattern and the corresponding processing are read from the system memory 12 by the CPU 11 and sent to the DCS 20 (S4).
0).

Thereafter, the CPU 11 executes the above processing procedure at regular intervals, and repeats the identification of the change state of the measurement signal, the setting signal, and the operation signal.

In addition to this embodiment, the following example can be implemented.

1) The closer the position of the measured value is to the abnormal area in FIG. 1, the higher the urgency with which the abnormality must be dealt with. Therefore, FIG.
As shown in, it is possible to set a stage for the abnormality sign, determine the degree of urgency for dealing with the abnormality according to the stage by the CPU 11, and display the determination result on the display device 14 to inform the user of the degree of urgency. Furthermore, this urgency may be determined not only in the deviation of the measured value, but also in combination with the change pattern. The information indicating the degree of urgency is attached to the change pattern prepared in advance and is displayed together with the identification result of the change pattern.

2) In the present embodiment, in order to execute the plant control in real time, the change pattern of the measurement signal is identified by using the pattern recognition method that can be executed by the four arithmetic operations, but the present invention is not limited to this embodiment. , Other recognition techniques can be used. For example, the result of comparing the change rate (or change amount) of the operation signal at the present time point and the previous time point with the change threshold value (FIG. 5).
(B) ΔMV. TRS) to the characteristic parameter DV. T
Results of comparison with a fixed threshold value used with RS (Fig. 5)
(B) ΔMV. ABS) as the characteristic parameter DV. A
When used at the same time as the BS, the abnormal sign can be detected earlier than when only the measurement signal is monitored.

3) Further, in the case where the sampling interval can be made long, the pattern matching processing of the curve showing the change state of the abnormal sign and the curve showing the time-series measurement value is performed to detect a similar change pattern. It is also possible to

[0045]

As described above, according to the present invention,
Not only can the abnormal signs be detected while the plant is operating in the normal range, the type of abnormal signs can be identified from the change pattern contents, and the urgency can also be known, so the user should make appropriate preparations. It is possible to prepare for the occurrence of abnormality.

For example, if the present invention is applied to the measurement signal and the operation signal at the same time, the abnormal sign can be detected earlier than in the case of applying the measurement signal only. Further, if the present invention is applied to the measurement signal and the setting signal at the same time, a natural fluctuation due to an operation of an operator or the like can be discarded even if it is once detected as an abnormal sign.

[Brief description of drawings]

FIG. 1 is a waveform diagram showing an abnormal sign and an abnormal sign process according to an embodiment of the present invention.

FIG. 2 is a waveform diagram showing an abnormal sign and an abnormal sign process according to an embodiment of the present invention.

FIG. 3 is a waveform diagram showing an abnormal sign and an abnormal sign process of an embodiment of the present invention.

FIG. 4 is a waveform diagram showing an abnormal sign and an abnormal sign process of an embodiment of the present invention.

FIG. 5 is a diagram showing the contents of a table used for matching change patterns.

FIG. 6 is a block diagram showing a system configuration of an embodiment of the present invention.

FIG. 7 is a flowchart showing a processing procedure executed by the CPU 11 of FIG.

8 is a diagram showing another form of the table of FIG.

[Explanation of symbols]

 10 Plant Monitoring / Diagnostic Device 11 CPU 12 System Memory 13 Keyboard 14 Display Device 15 I / F 16 External Storage Device 20 DCS 30 Plant

Claims (3)

[Claims] 1. A plurality of types of time-series change patterns indicating abnormal signs of the plant and a plant control device are predetermined in a normal fluctuation region of the measurement signal corresponding to a normal operating range of the plant, and the measurement signal and The setting signal and the operation signal of the plant control device are input to the plant monitoring and diagnosing device in time series, and the plant monitoring is performed to determine which of the plurality of types the change pattern indicated by the measurement signal, the setting signal, and the operation signal matches. An abnormality symptom identification method for a plant and a plant control device, characterized by identifying an abnormality symptom of a plant and its type contents by making a determination in a diagnostic device. 2. The change pattern is accompanied by information indicating the degree of urgency until the occurrence of an abnormality, and information indicating the degree of urgency is added to the identification result of the abnormality symptom of the plant and the plant control device. The method for identifying abnormal signs of a plant and a plant controller according to claim 1. 3. Storage means for storing a plurality of types of time-series change patterns indicating abnormal signs of the plant and the plant control device within a normal fluctuation region of a measurement signal corresponding to a normal operating range of the plant; Input means for inputting signals and setting signals and operation signals of the plant control device in a time series, and change patterns matching the change patterns indicated by the input measurement signals, setting signals, and operation signals are extracted from the storage means, A plant monitoring and diagnosing device, comprising: a pattern identifying means for outputting the extraction result as the type content of the abnormality sign.