JPH08234832A - Device and method for monitoring and diagnostic plant - Google Patents

Abstract

PURPOSE: To speedily diagnose the factor of abnormality even when the influence of abnormality generated at a large scale plant is propagated between systems or subsystems in a short time and any change appears in process signals over a wide range. CONSTITUTION: This device is provided with a monitor processing part 3 for detecting the runaway of a monitor index from a normal range calculated by the process signal observed at the plant as an abnormality indication, qualitative model data base 6 registering a network model describing the characteristics of influence propagation between the respective monitor indexes, abnormality propagation route identifying part 4 for identifying the route of abnormality propagation caused by abnormality by collating the observed indication pattern with the network model, cause-and-effect table data base 7 registering a cause- and-effect table describing the indication patterns of respective monitor indexes estimated to various abnormality factors, abnormality factor identifying part 5 for identifying the abnormality factor by collating the cause-and-effect table of monitor indexes identified as the origins of abnormal changes with the observed indication patterns, and output display part 10 for outputting the identified influence propagation route.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a nuclear power plant,
The present invention relates to a plant monitoring diagnostic device and method for monitoring an operating state based on a process signal of a large-scale plant such as a thermal power plant or a chemical plant to detect an abnormality early and diagnose it.

[0002]

2. Description of the Related Art Generally, an abnormality occurring in a plant is
There are two types of mechanical abnormalities in the equipment that constitutes the plant, and abnormal operating conditions, such as unstable vibration of the flow rate of the fluid, although mechanically sound. In the case of a pump, the process signal used to detect these abnormalities is a main effect parameter indicating performance / function such as discharge pressure and flow rate, and secondary effects such as vibration, sound, and temperature that accompany operation. There are parameters and.

It is highly possible that the above-mentioned mechanical abnormality can be detected by providing a detector for measuring the secondary effect parameter for each device and monitoring it, but a large cost is required in a large-scale plant. Therefore, such a direct monitoring is performed only for major devices, and for many devices, a method is adopted in which the influence of the abnormality is detected only after the main effect parameter changes.

However, while the changes appearing in the secondary effect parameters are almost always associated with the abnormalities of a specific device, the main effect parameters include the abnormalities of many devices in addition to the above-mentioned abnormalities as the operating state. Abnormalities can affect. Also, when a change in operating conditions or a device operation confirmation test is performed, a change in the main effect parameter does not necessarily mean that the change is due to an abnormality.

From the background described above, conventionally, an abnormality is detected by the main effect parameter of the plant, and when the entire plant is viewed as one system, which of the grids or subsystems constituting the system is the source of the abnormality. Diagnostic techniques have been developed. In addition, by using the secondary effect parameter as well, the cause thereof, that is, what kind of abnormality is the abnormality in the operating state, and which equipment is abnormal if the equipment is abnormal Techniques for diagnosing whether or not there are also being developed.

[0006] These techniques are introduced, for example, in "Equipment Diagnosis Predictive Maintenance Encyclopedia" (published in 1988) supervised by Eiji Oshima, and typical examples of techniques for diagnosing abnormal sources include
There is a so-called model-based method, and a causal table-based method is a typical example of a technique for diagnosing the cause.

In the model-based method, the behavior of the main effect parameter governed by the physical law in a normal operating state is described in each subsystem by a mathematical model. Then, using this model, the output parameter is calculated from the observed value of the input parameter to the subsystem, and the abnormality is detected by comparing the result with the value of the actually observed output parameter. In other words, the subsystem that shows the change in the output parameter that cannot be explained by the change in the input parameter is used as the abnormal source. However, the model-based method requires a high-accuracy model in order to detect the occurrence of an abnormality in the initial stage, and it is difficult to construct a high-accuracy behavior model for a large-scale plant as a whole.

Therefore, a diagnostic method using a directed graph with a sign has been developed in which only the information indicating whether the phase delay of the transfer behavior between the main effect parameters is less than 180 degrees (in phase) or more than 180 degrees (negative phase) is described as a network model. ing. This method detects that the main effect parameter has increased or decreased with respect to the normal value of each, and compares the increase / decrease direction of the change with the qualitative behavior model of the entire plant system described by the signed directed graph to change the change. It identifies the parameter that became the starting point of and the transmission path of influence.

On the other hand, the method based on the cause-and-effect table uses as a symptom the judgment result of whether the value of the process signal is increased (or abnormally high), decreased (abnormally low) or normal compared to the normal value. A table in which a symptom pattern, which is a set of symptoms of each signal, is associated in advance for each possible cause of abnormality is given in advance, and the observed symptom pattern is collated with this table for diagnosis. The causal table is generally created by covering failure causes / impacts that should be assumed by the failure modes / impact evaluations of the devices constituting the plant and predicting the symptom pattern by an event tree analysis that evaluates the ripple effect.

As for the influence of the planned operation appearing in the process signal by being performed artificially and mechanically, when the state signal of the operation switch is measured,
A method is adopted in which it is detected that the operation is in progress and the monitoring diagnosis is not performed during the operation.

[0011]

By the way, in order to maintain the soundness of the plant and perform stable operation, the process signal is continuously or regularly monitored to promptly detect a change caused by the occurrence of an abnormality. , It is required to investigate the cause and deal with it appropriately. However, in many cases, it is difficult to meet the requirements of the conventionally developed technologies due to the following problems. That is, (1) The qualitative model-based diagnosis method allows the abnormalities in the plant to propagate between systems or subsystems within a short period of time, resulting in abnormalities in signal units even when a wide range of process signals change. Source identification is possible, but generally it is not possible to identify the cause, except in the special case where individual signals respond one-to-one to a particular anomalous cause.

(2) The method based on the causal table can diagnose possible causes, but since the size of the table expands exponentially with the size of the number of signals to be handled, it can be simply applied to a large-scale plant. Not applicable.

In order to deal with this problem, it is considered that the entire plant is decomposed into several systems or subsystems and a causal table is created for each of them.
This makes diagnosis difficult when the influence of an abnormality propagates between systems or subsystems in a short time.

(3) The method based on the cause-and-effect table cannot deal with a case where an unexpected abnormality occurs at the start of operation of the plant monitoring and diagnosing apparatus.

(4) The symptom pattern defined in the causal table does not always completely match the pattern actually observed depending on the degree of abnormality. In addition, if the causal table is divided by system or subsystem, changes in signals that are not included in the given symptom pattern for the identified cause may be observed, and all the symptom may be identified. It is difficult to understand whether or not it is possible to explain due to other reasons.

(5) The causal table can include not only the symptoms of the measured values of the respective signals but also the symptoms when various characteristic parameters obtained by processing the signals are used as monitoring indexes. It is difficult to understand the reason for diagnosis because it does not include information on causal relationships between indicators.

The above-mentioned (2) is a problem to be solved in order to achieve the processing time practically required as a monitoring / diagnosing device. However, the problems to be solved from the viewpoint of practicality are as follows. There are things.

(6) It is effective to monitor important abnormal events peculiar to the target plant or abnormalities actually experienced, using peculiar indices suitable for each detection. is there. As described in (5) above, it is possible to make a diagnosis by including the symptoms of such a special index in the causal table, but it is not efficient when the abnormality of the special index directly indicates the occurrence of the specific abnormality. .

(7) In addition to simple symptoms such as increase and decrease of the process signal, there are cases where the cause of abnormality can be estimated by comparing the way of change, that is, the waveform with the waveform of past case data. Since it relied on the inspiration of humans engaged in investigating the cause, diagnosis is not always possible.

(8) It has not been conventionally performed to positively identify the contents of the planned operation based on the change appearing in the process signal by performing the operation artificially and mechanically. For the signals that are not received, detailed measures such as continuing monitoring are not taken. In addition, when human operation is performed for the purpose of confirming the operation of the device, the operation of the device to be confirmed is naturally monitored, but whether the change is appropriate for other process states affected by the operation is confirmed. It is usually difficult to always monitor.

In addition to the above, it is not always possible to identify an abnormal source even in the diagnostic method using the qualitative model described in (1) above, and some problems to be solved in practical use are pointed out as follows. That is, (9) As described in the related art, the qualitative model represented by the signed directed graph expresses the propagation characteristics of signal changes only in the propagation direction and phase. That is, if the signal on the cause side increases, the signal on the result side also increases, and if it decreases, the relationship of "promoting" has a positive transfer gain, and if the signal on the cause side increases, the signal on the result side decreases. "Suppressed" with negative transfer gain, increasing, decreasing
There are only two types of relationships. Therefore, it is not possible to express the relationship that the signal on the result side is kept constant by increasing or decreasing the signal on the cause side, and it is not possible to handle the feedback effect and the like.

(10) The qualitative model does not include the concept of time in the propagation characteristics of signal changes. Therefore, it is not possible to diagnose the propagation situation of successive transient changes in consideration of the passage of time. Further, when the influence propagation path of the identified monitoring index indicating the abnormal sign has a closed loop shape, it is impossible to identify the signal that is the starting point of the change.

(11) The diagnostic method based on the qualitative model is to identify the signal that is the starting point of the change and the influence propagation path based on the initial change of the observed signal, and when the abnormality progresses and the protection sequence operates. However, it is not possible to diagnose the plant state, which changes moment by moment, in real time.

The present invention has been made in consideration of the above circumstances, and a first object thereof is to solve the problems (1), (2),
By solving (3) and (4), the cause of abnormality can be promptly detected even when the influence of abnormality that has occurred in a large-scale plant propagates between systems or subsystems within a short time and changes occur in a wide range of process signals. An object of the present invention is to provide a plant monitoring and diagnosing apparatus and method capable of diagnosing.

The second object of the present invention is to solve the above problem (5).
It is an object of the present invention to provide a plant monitoring and diagnosing apparatus and method capable of solving the above problems and providing easy-to-understand diagnosis results.

The third object of the present invention is to solve the above problem (6).
It is an object of the present invention to provide a plant monitoring / diagnosing device having a function of diagnosing a widely conceivable abnormality and a function of efficiently performing a diagnosis focusing on a specific event.

A fourth object of the present invention is to solve the above problem (7).
It is an object of the present invention to provide a plant monitoring and diagnosing apparatus and method capable of making a diagnosis by comparing the waveform of an observed process signal with the waveform of an abnormal case that has been experienced in the past.

The fifth object of the present invention is to solve the above problem (8).
On the other hand, it is another object of the present invention to provide a plant monitoring and diagnosing apparatus and method capable of detecting the occurrence of a true abnormality by identifying the influence of a planned operation that appears in a process signal.

A sixth object of the present invention is that the abnormal sign that should appear in the process signal is suppressed by the feedback effect and does not appear apparently, or the abnormal sign cannot be detected because it is hidden by normal fluctuation because the signal change is small. A plant that solves the problems (9), (10), and (11) of the diagnostic method based on the qualitative model, such as the origin of the change cannot be identified when the influence propagation path becomes a closed loop, without complicated numerical processing. It is to provide a monitoring diagnostic device and method.

Hereinafter, in the present invention, all monitoring targets including the observed value of the process signal itself and performing abnormality detection by comparison with a threshold value indicating a normal range are referred to as monitoring indices.

[0031]

In order to achieve the first object, the present invention provides means for combining the ability of identifying an abnormal source by a qualitative model and the identification of an abnormal cause by a causal table. That is, the deviation from the normal range of the monitoring index calculated from the process signal observed in the plant is detected as an abnormal sign, and the obtained sign pattern of each monitoring index is described as the influence propagation characteristic between each monitoring index. After identifying the monitoring index that is the starting point of the abnormal change and its influence propagation path by comparing with the network model, identify the various abnormal causes of the system or subsystem of the plant that the starting point of the abnormal change and the identified monitoring index correspond to. Corresponding to the causal table, at least the identified influence propagation route is output by identifying the cause of abnormality by matching the causal table describing the symptom pattern of each expected monitoring index with the observed symptom pattern. If a symptom pattern is included, the cause of the abnormality is also output, and if the corresponding symptom pattern is not included in the causal table, this is added. Providing plant monitoring diagnostic apparatus characterized by performing learning.

Further, in the identification with the above-mentioned causal table, it is observed as a causal table that describes the symptom pattern expected only for the abnormal cause that first affects the monitoring index that is the starting point of the influence propagation route due to the abnormal condition. Provided is a plant monitoring and diagnosing device characterized by identifying the cause of an abnormality by collating it with the above-mentioned symptom pattern.

In order to solve the problem (2) of the diagnosis based on the cause-and-effect table, the present invention shows the symptom pattern of each monitoring index for various abnormalities including abnormalities of other systems and subsystems for each system and subsystem of the plant. It is not affected by the other system or subsystem where the anomaly was observed, by comparing it with the causal table that describes the symptom pattern of each monitoring index assumed for the cause, and the anomaly symptom of the other system or subsystem. The present invention provides a plant monitoring and diagnosing device characterized by identifying a cause that can be explained as a current abnormal cause.

In order to achieve the second object of the present invention, in the display of the diagnostic result by the causal table, the causes with many monitoring indexes having the same signs are rearranged in the descending order of possibility as a candidate with high possibility, Further, the present invention provides a plant monitoring and diagnosing method characterized in that the monitoring indicators are displayed in a tabular form sorted in ascending order of time when an abnormal sign appears.

Further, for the abnormal cause candidate identified as a result of the diagnosis based on the causal table, the event tree analysis performed at the time of creating the causal table includes the abnormal sign of each monitoring index, and the progress of the event is tree-shaped. Provided is a plant monitoring diagnostic method characterized by displaying.

In order to achieve the third object of the present invention, a monitoring index peculiar to the abnormality is calculated from the process signal which reacts only to a specific abnormality occurring in the plant, and the deviation from the normal range is detected. According to the first diagnosis means for identifying the abnormality as the cause of the abnormality, the average value within the past fixed period from the current value of the process signal or the detection signal, a value obtained by smoothing the signal, and at least the other. Residual difference obtained by subtracting a value given as a function of at least one of the current and past values of one process signal as a predicted value is used as a monitoring index, and deviation from its normal range is detected as an abnormal sign, and a sign pattern of each monitoring index And a second diagnostic means for identifying the cause of abnormality by comparing with a causal table in which symptom patterns assumed for various causes of abnormality are described. When the second diagnosis means performs concurrently, to provide the plant monitoring and diagnosing apparatus and outputs in preference to the results of the results of the first diagnostic means second diagnostic means.

A third object of the present invention is to execute the first diagnosing means and, when the cause of the abnormality is identified, output the result and not execute the second diagnosing means. This can also be achieved by a plant monitoring and diagnosing device characterized in that the second diagnosing means is executed only when the relevant cause is not identified.

In the above description, the symptom patterns assumed for various abnormal causes described as the causal table can be expressed as, for example, a tree-type causal tree or a production-type rule.

In order to achieve the fourth object of the present invention, the waveform of the temporal change of the monitoring index which is the starting point of the change identified by the qualitative model is registered as the starting point of the change. A plant characterized by comparing and collating with the time change waveform of the monitoring index calculated from the already-existing case data, and diagnosing an abnormal cause of a case that gives a high degree of similarity above a certain value as a candidate for the current abnormal cause. Provided are a monitoring diagnostic device and method.

In the comparison and comparison between the waveforms, a fixed period before and after the change in the monitoring index, which is the starting point of the change, is taken out as time-series data, and similarly calculated with case data of the same period length. The time series data obtained by extracting the maximum value of the correlation function, or the period t0 before and after the change in the monitoring index, which is the starting point of the change, is taken as the basic function f.
(T), an evaluation function calculated between the case data x (t)

[Equation 2] The maximum value of is the similarity.

In order to achieve the fifth object of the present invention, the change of the monitoring index caused by the start and end of each planned operation manually or mechanically performed on the plant and the influence of the operation are performed. The monitoring index that may change in response to the change and the period in which the change appears are registered in advance, and when the change accompanying the start of the operation is detected, the affected monitoring index is detected for the fixed period or the change associated with the end. (EN) Provided are a plant monitoring and diagnosing apparatus and method, which are characterized by excluding from abnormal detection targets.

The present invention also provides an analog monitoring index X which changes first by the operation as means for monitoring the response characteristic of the process signal affected by the operation during the period when it is judged that the planned operation is being performed. Other analog monitoring indexes {Yi} that may change under the influence are registered in advance, from the time when the change accompanying the start of the operation is detected to the time when the change associated with the end or the end is detected. For the analog monitoring index {Yi}, a relative monitoring rate {Yi / X} is used as a new index, and a deviation from the normal range is detected as an abnormal sign.

The purpose of identifying the planned operation is to include the symptom pattern of each monitoring index, which is assumed for the planned operation performed artificially and mechanically in the plant, in the causal table for identifying the cause of abnormality. By doing so, it can be achieved by a plant monitoring and diagnostic device characterized by collating an abnormality sign pattern of a monitoring index created by a normal monitoring process with this causal table.

In addition, case data in which a planned operation that is artificially or mechanically performed on the plant is performed is prepared, and the abnormal sign pattern of the monitoring index is collated with the qualitative model to start the abnormal change. This is achieved by a plant monitoring diagnostic device that diagnoses by collating the time-varying waveform of the monitored index with the waveform of the planned operation execution case data in which the monitored index is the starting point of the change.

In order to solve the problem (9) of the diagnostic method based on the qualitative model among the sixth objects, the present invention introduces a relationship in which the influence propagation characteristics between monitoring indexes are held constant by feedback control, (EN) Provided is a plant monitoring / diagnosing device characterized in that a propagation path is considered to be established even when an apparent value of a monitoring index is not affected by a transient change suppressed by a feedback effect of a control system.

To solve the problem that the influence propagation path of an abnormality is difficult to identify when the signal change is small, a time delay is given to the influence propagation characteristics between the monitoring indexes, and the monitoring index on the cause side is abnormal due to the monitoring process. When a symptom is detected and not detected in the monitoring index on the result side, the characteristic of change of the monitoring index on the result side is extracted based on this time delay, and the abnormal range is set by limiting the time range and the range of normal values. The present invention provides a plant monitoring and diagnosing device characterized by identifying an abnormal propagation path after re-evaluating.

In order to solve the problem (10) of the diagnostic method based on the qualitative model, the present invention also provides a monitoring index in the closed loop in which an abnormal sign is first detected when the identified propagation path constitutes the closed loop. Is the starting point of the closed loop, and when different propagation paths are apparently established between the two monitoring indices, at the abnormal sign detection time of the cause-side monitoring indices,
Provided is a plant monitoring / diagnosing device characterized by estimating a true propagation path by comparing a time obtained by adding a propagation time delay between monitoring indexes given in advance with an abnormal sign detection time of a monitoring index on a result side. .

Problems of Diagnostic Method Using Qualitative Model (11)
For this, the influence propagation characteristics that the operation of the protection sequence provided for the plant abnormality affects the monitoring index are included in the qualitative model, and the abnormality of the monitoring index observed from the time when the operation of the sequence is detected. Provided as a solution means is a plant monitoring / diagnosing device characterized by monitoring a response of a plant by comparing a symptom with a symptom of a subsequent monitoring index predicted based on the qualitative model.

In addition to the above, as a means for providing the diagnosis result of the abnormal influence propagation path by the qualitative model in an easy-to-understand manner that should be included in the second purpose, the propagation characteristic with a positive gain,
Increase the observed change behavior of the monitoring index by drawing a diagram that connects the names of the process signals that correspond to the monitoring index with three types of directional line segments: the propagation characteristic with negative gain and the propagation characteristic showing the feedback effect. The result of classifying into three types of decrease, no change and no change is displayed together with the name of the process signal, the starting point of the change and the name of the identified process signal are highlighted, and the result corresponds to the identified influence propagation path. It is intended to provide a plant monitoring and diagnosing device characterized in that the color of a line segment is changed and displayed.

Further, instead of the directed line segment, the observation point and signal name of the process signal are clearly shown in the system diagram of the plant, and the change behavior of the monitoring index calculated from the process signal is increased, decreased, And the result classified without change is displayed together with the process signal name, and the process signal corresponding to the monitoring index identified as the starting point of the change and the upstream device that directly affects the signal are highlighted. Control circuits, pipes, valves, which are present in the identified impact propagation paths,
The present invention provides a plant monitoring and diagnosing device characterized in that a color of a pump or the like is changed and displayed.

[0051]

In the plant monitoring / diagnosing apparatus of the present invention according to claim 1, when an abnormal sign is detected in the monitoring index, the monitoring index which is the starting point of the change is detected by the network model describing the effect propagation characteristic between the monitoring indexes. Since the influence propagation route is identified, it is possible to easily grasp whether or not all the abnormal signs can be explained by one starting point.

Further, it becomes clear that the system or subsystem of the plant from which the process signal used for calculation of the monitoring index of the starting point is obtained is an abnormal source, and it can be compared with the abnormal sign pattern observed in the abnormal cause identification. It is possible to limit the scale of the causal table to be used to a small system or subsystem unit.

As a result, even if an abnormality occurs in which the influence propagates to a wide range of process signals in the plant within a short time, if the abnormality can be assumed, the cause is diagnosed and it is unexpected. It is possible to present at least the anomaly source and the influence propagation route even for the anomaly.

In the plant monitoring and diagnosing device of the present invention as defined in claim 2, when an unexpected abnormality is diagnosed by the plant monitoring and diagnosing device according to claim 1, the observed symptom pattern is immediately registered in the causal table. As a result, it can be learned as one of the expected events and can be provided as knowledge useful for subsequent diagnosis.

According to the plant monitoring and diagnosing apparatus and method of the present invention described in claims 3 and 4, not only the same operation as the plant monitoring and diagnosing apparatus according to claim 1 is obtained, but also the scale of the causal table used for identifying the cause of abnormality is The time required for the diagnostic processing can be further shortened because the monitoring index identified as the starting point of the change is limited to only the cause of the abnormality in which the influence of the abnormality is expected to appear first.

In the plant monitoring and diagnosing device according to the fifth aspect of the present invention, when an abnormality occurs that propagates an influence to a process signal in a wide range of the plant within a short time, a system to be collated or There will be multiple causal tables for each subsystem, but by including abnormalities in other systems and subsystems as the cause candidates in each causal table, the causal relationship between systems and subsystems can be understood, and the cause The cause of abnormality can be diagnosed by the causal table of the system or subsystem close to the side.

In the plant monitoring and diagnosing method according to the sixth and seventh aspects of the present invention, with respect to the abnormality cause candidate identified in the diagnosis based on the cause-and-effect table, the monitoring indexes in which abnormality signs are detected are rearranged in the order of detection time. By displaying or directly displaying in the event tree, it is possible to assist in grasping the relation of the abnormality sign of each monitoring index which is the basis of the identification.

In the plant monitoring and diagnosing device of the present invention as defined in claims 8 and 9, the first diagnosing means based on the monitoring index peculiar to a specific abnormality is calculated by a process which is uniformly applicable to many process signals. By providing the second diagnostic means separately based on the comparison between the abnormal sign pattern of the obtained monitoring index and the causal table, it is possible to shorten the processing time when the short-circuit diagnosis is possible. On the contrary, when it takes a long time to calculate the unique monitoring index, such separation can eliminate the influence on the diagnostic processing time using the uniform monitoring index.

In the plant monitoring and diagnosing apparatus and method of the present invention according to claims 10 to 13, only the monitoring index identified as the starting point of the abnormal change by the qualitative model is compared and collated with the case data, It is possible to easily search for similar cases and contribute to the estimation of the cause of abnormality. By performing the diagnosis based on the qualitative model first, it is possible to eliminate the need for the waveform matching for all the monitoring indexes showing the abnormal sign.

In the plant monitoring and diagnosing apparatus and method of the present invention as defined in claims 14 and 15, it is confirmed that a planned operation has been performed before the monitoring processing is performed, and the monitoring which may be influenced by the operation is performed. By excluding only the index from the target of abnormality detection, it is possible to prevent the monitoring diagnostic device from issuing an unnecessary alarm and to continue monitoring other monitoring indices.

In the plant monitoring and diagnosing device according to the present invention, monitoring processing is executed even when a planned operation is performed, and when an abnormal sign is detected in the monitoring index due to the operation. Can identify the operation by matching the observed symptom pattern with the causal pattern symptom pattern, or by matching the waveform of the monitoring index identified as the starting point of the abnormal change by the qualitative model with the case data. it can. With this, it is possible to confirm whether or not the operation of the plant monitoring / diagnosing device is normal each time a planned operation is performed.

According to the plant monitoring and diagnosing method of the present invention described in claim 18, in the plant monitoring and diagnosing apparatus and method according to claim 14 and claim 15, the plant is selected by a planned operation among the monitoring indexes excluded from the monitoring targets. Input an analog monitoring index that can evaluate the magnitude of a given disturbance,
Considering each analog monitoring index that changes under the influence of it as an output, and monitoring the transfer gain between the input and output during the disturbance occurrence period, it is as if the behavior of the plant was changed as if an active disturbance application test was performed. Can be monitored.

In the plant monitoring and diagnosing device of the present invention according to claim 19, by checking the symptom pattern of the monitoring index in consideration of the relationship introduced into the qualitative model, which can be held constant by the feedback effect, Even if there is a monitoring index in the middle of the propagation path where no change is observed due to the influence of the feedback effect, it can be identified as the propagation path.

In the plant monitoring and diagnosing device according to the twentieth aspect of the present invention, the range for confirming the abnormality sign necessary for the diagnosis based on the qualitative model has passed a given time delay from the time when the signal on the causative side has changed. By setting a limited time zone before and after the time, it is possible to detect signs of change with high sensitivity.

In the plant monitoring and diagnosing device of the present invention according to claim 21, even when the influence propagation path of the abnormality identified by the qualitative model becomes a loop, the starting point of the loop is based on the abnormality detection time of each monitoring index. Can be identified.
Further, even when a plurality of propagation paths can be established between the monitoring indexes, the true propagation path is identified by collating the time when the abnormal sign of each monitoring index is detected with the time delay given to the propagation characteristics. be able to.

In the plant monitoring and diagnosing device according to the twenty-second aspect of the present invention, the response of the process signal is predicted starting from the operation of the protection sequence provided in case of a plant abnormality, and the observed behavior is predicted. It is possible to monitor the plant state, which changes from moment to moment, in real time by sequentially checking whether or not they match.

In the plant monitoring and diagnosing device of the present invention according to claims 23 and 24, an abnormal sign is detected on the qualitative model according to claim 20 represented by a signed directed graph including a feedback effect or on the plant system diagram. It is possible to help the user to easily understand the relation between the abnormal signs of the monitoring indicators by displaying the different monitoring indicators and the identified propagation paths in different colors.

In the plant monitoring / diagnosing apparatus of the present invention as set forth in claim 25, the same action as that of the plant monitoring / diagnosing apparatus and method as set forth in claim 1-9 or 16 is shown by a cause identifying means based on a different symptom pattern matching system. You can

[0069]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a plant monitoring and diagnosing apparatus and method according to the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the construction of the first embodiment of the plant monitoring and diagnosing apparatus according to the present invention, and FIG. 2 is a flow chart showing the processing flow of the first embodiment. In FIG. 1, the control unit 1 has a function of controlling the execution of the entire apparatus, and periodically outputs a diagnosis execution command. Upon receiving this execution command, the signal input unit 2 inputs the process signal from the plant according to preset input conditions such as the signal ID and the sampling period.

The monitoring processing unit 3 calculates the monitoring index from the process signal input according to the preset monitoring conditions, such as the signal processing method for each monitoring index and the upper and lower thresholds indicating the normal range, and compares it with the upper and lower thresholds. Then, three kinds of signs are judged to be "increase" for the index exceeding the upper threshold, "decrease" for the index falling below the lower threshold, and "no change" within the normal range. The above "increase" or "decrease" is an abnormal change.

As the monitoring index, from the signal itself or the current value of the signal corresponding to the standard deviation obtained by the detection processing,
For example, an average value within a certain period in the past, a value obtained by smoothing the signal, or a value given as a function of the present and / or past values of other one or more process signals is subtracted as a predicted value. Used residuals. Since such signal processing does not focus on a specific abnormal event, it can be uniformly applied to many signals.

Further, the monitoring processing unit 3 stores the direction of change when an abnormal sign is first detected for each monitoring index as a sign. In other words, once it is determined to be “increase”,
Even if it changes to “no change” or “decrease”, the first sign “increase” is used as the sign for later diagnosis. In this way, the monitoring processing unit 3 outputs a symptom pattern that is a set of symptom determination results for each monitoring index.

The abnormal propagation path identification unit 4 collates the qualitative model in which the influence propagation characteristic between the respective monitoring indexes registered in the qualitative model database 6 is expressed by a directed graph with a sign with the observed symptom pattern, and detects the abnormal symptom. Identify the impact transmission route of. The monitoring index located at the starting point of this route is identified as the one that shows the change first, and the system or subsystem of the plant in which the process signal used to calculate the index is observed is determined to be the abnormal source. It should be noted that the function of the abnormal source identification unit is up to the identification of the monitoring index that is the starting point of the change, and as shown in FIG. 2, the function of identifying the abnormal source system or subsystem has the abnormality cause identification unit 5 described below. It may be included.

The abnormality cause identifying unit 5 is created for each system or subsystem of the plant, and is selected from a plurality of causal tables registered in the causal table database 7.
The cause-and-effect table for the system or subsystem identified as the abnormal source by the abnormal propagation path identification unit 4 is read out and collated with the symptom pattern created by the monitoring processing unit 3. In the collation, for example, an assumed cause having no contradiction in the abnormal sign and a large number of monitoring indexes having the same sign is identified as a candidate for the current abnormal cause. At this time, it is possible to identify that the operation is performed by including the symptom pattern observed by the planned operation in the causal table.

Hereinafter, the monitoring processing unit 3, the abnormal propagation path identifying unit 4, the abnormal cause identifying unit 5, and the parts of the qualitative model database 6 and the causal table database 7 surrounded by broken lines in FIG. To do. Further, a portion excluding the monitoring processing unit 3 from the monitoring diagnosis unit 8 is referred to as a diagnostic processing unit 9.

The control unit 1 sends the diagnosis result of the abnormality propagation path identification unit 4 to the output display unit 10 and also sends the result to the output display unit 10 when the cause candidate is identified by the abnormality cause identification unit 5. When the corresponding cause is not identified by the abnormality cause identification unit 5, it is determined that an unexpected abnormality has occurred, and the current symptom pattern is sent to the database management unit 11.

When the symptom pattern is sent from the control unit 1, the database management unit 11 additionally registers it in the causal table database 7 as an unregistered pattern. To the symptom pattern learned in this way, the necessary information such as the event name is added by the dialogue processing unit 12 when the cause is found later.

The output display section 10 displays various information sent from the control section 1 in an easy-to-understand format.

Hereinafter, the abnormality propagation path identification based on the qualitative model, the abnormality cause identification based on the causal table, and the output display unit 1
The features of 0 will be described in more detail.

First, the features of the means provided by this embodiment in order to solve the problems of the conventional method in identifying the abnormal propagation path based on the qualitative model will be described in detail with reference to the drawings. The monitoring index name in the specific example uses the process signal name itself for simplification.

First, the introduction of feedback characteristics into the qualitative model will be described with reference to FIG. For example, in the pressure control system of a boiling water reactor, the control system adjusts the opening degree of the regulator valve so that the pressure of the main steam becomes constant. When the reactor pressure decreases, the main steam pressure also decreases, so the control system tries to raise the pressure by narrowing the opening of the regulator valve. Therefore, there is a "promotion" relationship expressed by a solid lined segment in the figure between these observation signals.

However, as shown in FIG. 3, even if the reactor pressure is decreased, the response of the control system becomes faster when the transient change speed is slow, even if the reactor pressure decreases.
As a result, the change in main steam pressure is not visible as represented by →, and in the conventional method of the signed directed graph, the transmission path of the transient change is before and after that as indicated by the directional line segment of the dashed arrow in the figure. Will be cut by.

Therefore, in the present embodiment, by introducing a qualitative model expressing a feedback characteristic that the main steam pressure is kept constant by increasing or decreasing the opening degree of the regulator valve, the case where no apparent change appears However, the propagation route can be correctly identified.

Next, the introduction of the time delay to the influence propagation characteristic between the monitoring indexes will be described with reference to FIG. As shown in FIG. 4, when the monitoring processing unit 3 detects a sign of change such as “increase”, “decrease”, or “no change” from the behavior of the monitor index, for example, the maximum amplitude of the monitor index observed in normal time. The upper and lower thresholds are set by defining twice the normal range as the normal range. However, when the degree of transient change caused by an abnormality is small, it is hidden by normal fluctuations like the example of the signal C and accurately. It may not be detected.

Therefore, in this embodiment, when a transmission delay is introduced into the causal relationship between the monitoring indexes, and a change in the monitoring index on the cause side is detected and no change is detected on the monitoring index on the result side, The change in the monitoring index on the result side within a limited range after the transmission delay time after the change of the monitoring index on the cause side is also investigated by limiting the normal range as well, whereby accurate detection can be performed. This makes it possible to correctly estimate the route even when the degree of transient change is small.

Similarly, as shown in FIG. 4, even when the identified influence propagation path has a loop shape, the starting point of the loop is identified based on the abnormality detection time of each monitoring index, and the monitoring index is set between the monitoring indexes. Even when a plurality of propagation routes can be established, the true propagation route can be identified by comparing the time when the abnormality sign of each monitoring index is detected with the introduced time delay.

Further, a diagnosis method after the operation of the plant protection sequence will be described with reference to FIG. The conventional diagnostic method based on the signed directed graph is based on the effect propagation characteristics between plant signals from phase I under normal operating conditions to phase II where signs of anomalies are detected. It is intended to estimate the changed signal. Therefore, the state after shifting to the phase III stage, which shows different influence propagation characteristics due to the progress of anomalies, activation of equipment and system protection sequences, tripping of pumps, and scrum of the plant, is expected. It was difficult to diagnose.

Therefore, in the present embodiment, as shown in FIG. 5, by taking in a digital signal indicating the operation of the protection sequence, for example, when the pump trips, the change in the signal is predicted from that point, and the response of the plant is It is possible to diagnose whether it is normal or not.

The operation of the abnormal propagation path identifying unit 4 will be described below with reference to the flow chart shown in FIG.

When the monitoring processing unit 3 detects a sign of a signal change, the abnormal propagation path identifying unit 4 is activated. First, the increase / decrease of the observation signal extracted by the monitoring processing unit 3 is input, and the following inspection is performed on each signal. When the signal of interest is increasing or decreasing, the normal transmission model described in the qualitative model database 6 (how the increase and decrease are transmitted and time delay) is referred to, and the change in the signal on the causative side is regarded as the model. If they match, it is considered that a change has been transmitted between both signals, and the transmission path is saved. If they do not match, the signal is regarded as the starting signal. If the signal at the starting point represents the operation of the protection sequence, the sequence operation is confirmed, and if it is a normal process signal, the upstream device is saved as the abnormal point as the starting point of the change. If there is no change in the signal of interest, refer to the feedback model, if the movement of the signal on the causative side matches the model, save the feedback loop as a transmission path, and if it does not match, Alternatively, if there is no feedback loop, the signal of interest is ignored as it is irrelevant to the event.

By applying the above processing to all signals, the location of the abnormality and the route through which the influence is transmitted are identified.
Alternatively, the protection sequence operation and its response can be confirmed.

Next, the identification of the cause of abnormality based on the causal table will be described. In the first embodiment described with reference to FIGS. 1 and 2, the system or subsystem in which an abnormal change has occurred is obtained from the monitoring index identified as the starting point of the abnormal change, and the cause and effect assigned to each system or subsystem. The cause of the abnormality is identified by comparing the observed abnormality sign pattern with the table.

On the other hand, the configuration of the monitoring / diagnosis unit of the second embodiment and the causal table of monitoring indices are shown in FIG.
FIG. 8 shows the processing flow in (B). The same or corresponding portions as those of the first embodiment will be described using the same reference numerals. The same applies to the following embodiments.

As shown in FIG. 7, the abnormal propagation path identification unit 4
In this case, since the monitoring index that is the starting point of the abnormal change is identified, the causal effects of giving symptom patterns to various abnormal causes that are likely to have the first effect on the index for each monitoring index. It is also possible to register the table in the database 7a and cause the abnormality cause identifying section 5 to identify the cause based on the causal table for each corresponding monitoring index.

As described above, according to the second embodiment, by limiting the cause-and-effect table to be checked against the symptom pattern from the system or subsystem unit to the monitoring index unit,
It is expected that the diagnostic processing time by the computer will be shortened, the required memory at the time of execution will be reduced, and the diagnostic accuracy will be improved.

Third of plant monitoring and diagnosis apparatus according to the present invention
9A and 9B show the configuration of the monitoring and diagnosing unit and the causal table of subsystems in this embodiment, and FIG. 10 shows the flow of processing of the same monitoring and diagnosing unit. In the subsystem causal table database 7b, not only symptom patterns are registered for each system or subsystem of the plant, but also for various causes in which the system or subsystem is assumed to be an abnormal source, and other system Alternatively, if the influence of an abnormality that has occurred in the subsystem is assumed to be propagated, it is included and registered, and the cause of abnormality in another system or subsystem is also registered as the cause on the upstream side.

The abnormality cause identifying section 5 uses the relationship between each monitoring index given as a monitoring condition and the target system or subsystem to determine the abnormality source system or subsystem from the monitoring index showing the abnormality sign. A system candidate group {Si} is obtained.

Next, by comparing the abnormality symptom pattern created by the monitoring processing unit 3 with the causal table created for each system or subsystem, an abnormality cause candidate {Aji} is obtained for each abnormal source candidate Si. Figure 9 for each candidate
As shown in (B), the given upstream side factor {Fji} is identified, and the confirmation target cause list La including the abnormal cause candidates obtained for all the abnormal source candidates and the list Le including the upper and lower side factors are obtained.

Then, among the abnormality cause candidates included in the confirmation target cause list La, only those which are themselves included in Le and whose upstream factors are not included in La are the final abnormality cause candidates. judge. However, the abnormal source candidate is 1
In the case of only one, the La to be confirmed becomes the list of abnormality cause candidates as it is.

As a result, in the present embodiment, a system or subsystem that is not affected by another system or subsystem in which an abnormal sign is observed is determined to be an abnormal source,
Among the abnormal cause candidates identified based on the causal table for the relevant system or subsystem, the one that can explain the abnormal symptoms of another system or subsystem can be identified as the current abnormal cause.

In the example shown in FIG. 11, the abnormal signs of the second subsystem and the third subsystem are explained as the cause 2-1 and the cause 3-2, respectively. -2, the first subsystem is determined to be the abnormal source. Further, in the diagnosis result by the causal table of the first subsystem, two abnormal cause candidates 1-1 and 1-2 are identified.
Since -1 cannot explain the abnormal sign of the other subsystem, the abnormal cause 1-2 will be finally identified.

FIG. 12 is a block diagram showing the configuration of the monitoring / diagnosing unit in the fourth embodiment of the plant monitoring / diagnosing apparatus according to the present invention, and FIG. 13 is a flowchart showing the flow of the monitoring / diagnosing process. In the present embodiment, as in the case of the first embodiment shown in FIGS. 1 and 2 and the second embodiment shown in FIGS. 7 and 8, the abnormal propagation path identification unit 4 uses the monitoring processing unit 3 By comparing the abnormal symptom pattern of each monitoring index created in 1. with the qualitative model, the monitoring index that is the starting point of the change and its influence propagation route are identified.

The case data extraction unit 13 searches the time series data of the process signal at the time of abnormality occurrence registered in the case database 14 for the case data having the same monitoring index as the starting point of the change, and monitors the case data. According to the signal processing method given as the condition, the time series data of the monitoring index is calculated from the process signal for each applicable case. The comparison and collation unit 1 compares the degree of similarity between the waveform of the time-series data of the monitoring index observed as the starting point of the change and the waveform of the extracted case data.
5, the abnormality cause of the case showing a high degree of similarity above a certain value is diagnosed as the current abnormality cause.

A method of evaluating the degree of similarity will be described with reference to FIG. When the current data of the monitoring index identified as the starting point of the change and the waveform of the one-case data are as shown in the figure, a certain period t0 before and after the time when the abnormality is detected in the current data
Is used as the basic function f (t), and the next evaluation function between the case data x (t) and

(Equation 3) To calculate. The maximum value of the evaluation function as a function of the delay time τ is used as the similarity. In the above formula, the case data obtained by extracting the case data for the period t0 by the same method becomes the cross-correlation function of both data, and it is also one method to use this as the evaluation function.

As in the case of the cause identification by the causal table, it is possible to identify that the operation is performed by including the data when the planned operation is performed in the case database.

As described above, the same monitoring as usual is continued even while the planned operation is performed, and the change in the process signal associated with the operation is regarded as abnormal and the cause thereof is determined. It can be considered as an effective means for confirming the soundness of the function and operation of the diagnostic device.

However, there is a concept that the influence of the planned operation should not be regarded as abnormal and should be accurately identified. Based on such an idea, FIG. 15 shows the configuration of the monitoring processing unit in the case of the fifth embodiment of the plant monitoring / diagnosing apparatus according to the present invention. In the fifth embodiment, the operation identification unit 16 is provided in the preceding stage of the monitoring processing unit 3 of the monitoring diagnosis unit 8 of each embodiment shown in FIG. 1, FIG. 9, FIG. 9 or FIG. 1. The monitor processing means 3a is the same as the monitor processing unit 3 of the embodiments shown in the above figures.

In the operation information database 17, as shown in FIG. 16, the name of each planned operation, the condition for determining the start of the operation, the condition for determining the end, the longest operation period, and the operation i first change. When the analog monitoring index is Xi, the analog monitoring index {Yji} that changes under the influence of the operation, and the relative rate of change {Yji / Xi} that is a new index are Xi as an input variable and Yji as an output variable. It is considered that the transmission gain is, and the range of its normal gain is registered.

Next, the operation of the monitoring processing units 3a and 3b in the fifth embodiment will be described according to the flow of processing shown in FIG.

The operation identifying unit 16 shown in FIG. 15 calculates the indexes required for the start and end judgments and confirms the respective judgment conditions. Only the longest operation period or the end is confirmed from the time when the start of the operation i is confirmed. Until the condition is satisfied, Xi and {Y, which are monitoring indices that change due to the influence of the operation
ji} is used as a monitoring bypass target index to send a command to exclude from the target of abnormality detection to the first monitoring processing means 3a, while
An execution command is sent to the monitoring processing means 3b. In the second monitoring processing means 3b, as shown in FIG. 18, the relative change rate is calculated as a new monitoring index, and it is monitored by comparing it with the upper threshold Hi and the lower threshold Li given as normal gains.

By adopting the configuration of the monitoring processing section as described above, even during the planned operation, the first monitoring processing means 3a operates as usual for the monitoring index which is not affected by the operation. While continuing the monitoring process, it is possible to determine by the second monitoring processing means 3b whether or not the changing manner of the monitoring index that changes due to the operation is normal.

FIG. 19 is a block diagram showing the configuration of the sixth embodiment of the plant monitoring / diagnosing apparatus according to the present invention. . In the figure, the second diagnosing means 8b has the same function as that of the monitoring and diagnosing section 8 shown in FIG. 1, FIG. 7, FIG. 9, or FIG. 12, and assumes an abnormal symptom pattern obtained by uniform monitoring processing. Diagnosis is made by comparing the possible causes of abnormalities with the given symptom patterns in the form of a causal table. On the other hand, the first diagnosis means 8a detects the occurrence of the abnormality by calculating a monitoring index specific to the abnormality from the process signal that reacts only to the specific abnormality and comparing it with a threshold value indicating a normal range. For example, this corresponds to a case where an unstable flow phenomenon is detected by monitoring the amplitude of a vibration component of a specific cycle appearing in the flow rate signal.

Further, in the configuration of FIG. 19, the operation identifying section 16 including the operation information database 17 shown in FIG.
And all the second monitoring processing means 3b is included in the first diagnosing means 8a, and the first monitoring processing means 3a is included in the second diagnosing means 8b.
By including the execution control function of the operation identifying unit 16 in the control unit 1, it is possible to realize the function of the monitoring processing unit configured as shown in FIG.

Next, the output display method of the plant monitoring and diagnosing device of the present invention will be described.

FIG. 20 shows an example of the display of the identification result of the influence propagation route of the abnormality by the qualitative model, and the process signal name itself is used for each monitoring index for simplification. Equivalent to expressing the causal relationship between signals with a directed graph with a sign, by connecting the names of process signals with directed line segments that are represented by solid lines, broken lines, and dashed-dotted lines that show the positive / negative of the gain of transfer characteristics and the type of feedback effect. Represents the information of.

The results obtained by classifying the qualitative behavior of the observed signals into "increase", "decrease", and "no change" are displayed together with the signal names by arrow upslope, downslope, and →, By highlighting the signal name of the starting point of the transient change and changing the color and / or the shade of the directed line segment corresponding to the route identified as having transmitted the change, it is possible to easily identify the place where the transient change occurs and the ripple route. It is displayed so that it can be recognized.

FIG. 21 shows a display example of different diagnosis results by the qualitative model. In this case, the process signal observation points and signal names are clearly indicated in the plant system diagram, and the qualitative behavior of the observed signals is classified into "increase", "decrease", and "no change". It is displayed along with the name. Occurrence of a transient change by highlighting the signal at the starting point of the transient change and the equipment upstream of it, and changing the color of the control circuit, piping, valve, pump etc. existing between the signals identified as transmitting the change The location and spread route are displayed so that they can be easily recognized.

When displaying the diagnosis result based on the causal table, it is necessary to consider so that it is easy to understand the relation of the abnormality sign for each monitoring index. In the example shown in FIG. 22, the causes that are consistent with the observed symptom pattern and that have a large number of monitoring indicators having the same symptom are rearranged as high-probability cause candidates to the top, and the monitoring indicators also show abnormal changes. It is displayed by rearranging in order of earliest occurrence time.

Further, in FIG. 23, by expressing the result of the event tree analysis performed at the time of creating the causal table in a tree shape including the change of the monitoring index, the influence of the difference in the sign of each monitoring index on the diagnosis result is clearly shown. The following shows an example.

FIG. 24 is a block diagram showing the configuration of the seventh embodiment of the plant monitoring / diagnosing apparatus of the present invention. The operation identification unit 16 and the operation information database 17 in the figure are the same as those shown in FIGS. 15 and 16. In addition, the first monitoring processing unit 18a of FIG. 24 is the same as that of FIG. 1, FIG. 7, FIG.
This is the same as the monitoring processing unit 3 of the monitoring / diagnosis unit 8 in each of the embodiments shown in FIG.

The abnormal propagation path identifying unit 4 and the qualitative model database 6 of FIG. 24 are the abnormal propagation path identifying unit 4 of the monitoring and diagnosing unit 8 of each embodiment shown in FIGS. 1, 7, and 12.
And the qualitative model database 6 are the same. FIG.
First abnormality cause identifying unit 5a and cause-and-effect table database 7
Is the same as the abnormality cause identifying unit 5 and the causality table database 7 of the monitoring and diagnosing unit 8 of each embodiment shown in FIGS. 1 and 7.

Further, the second abnormality cause identifying section 5b in FIG.
Has a combined function of the case data extraction unit 13 and the comparison and collation unit 15 of the embodiment shown in FIG. 12, and the case database 14 is the same in both figures. Furthermore, the second of FIG.
The monitoring processing unit 18b is a unit that calculates a unique monitoring index that reacts to a specific abnormality and detects the occurrence of the abnormality, and is the same as the first diagnosis means 8a in FIG. Therefore, the second monitoring processing unit 18b may include the function of the second monitoring processing unit 3b of FIG. 15 as described in the description of FIG.

[0124]

As described above, according to the plant monitoring and diagnosing apparatus of the present invention as set forth in claim 1, when an abnormal sign is detected in the monitoring index, the network model describing the influence propagation characteristic between the monitoring indexes. Since the monitoring index that is the starting point of the change and its influence propagation route are identified by, it is possible to easily grasp whether or not all the abnormal signs can be explained by one starting point.

Further, it becomes clear that the system or subsystem of the plant from which the process signal used to calculate the monitoring index of the starting point is obtained is an abnormal source, and it is possible to check with the abnormal sign pattern observed in the abnormal cause identification. It is possible to limit the scale of the causal table to be used to a small system or subsystem unit.

As a result, even if an abnormality that affects the process signals in a wide range of the plant occurs in a short time, if the abnormality is a possible abnormality, the cause of the abnormality is diagnosed. It is possible to present at least the anomaly source and the influence propagation route even for the anomaly.

In this way, it becomes possible to detect and diagnose anomalies occurring in a large-scale plant at the symptom stage while distinguishing them from the effects of planned operation. Further, even when the influence of an abnormality propagates between systems within a short time and changes occur in a wide range of process signals, it becomes possible to quickly identify the abnormal source.
In any case, if it is an abnormality that can be assumed or has been experienced, it is possible to identify the cause or abnormality cause for each device.In the case of an unexpected abnormality, at least the abnormality source can be identified for each signal. It will be possible. Therefore, it is easy to understand the situation at the time of abnormality, and it is expected that early response at the symptom stage and swift restoration to normality will be expected.

According to the plant monitoring and diagnosing apparatus of the present invention described in claim 2, when the unexpected abnormality is diagnosed by the plant monitoring and diagnosing apparatus according to claim 1, the observed symptom pattern is immediately converted into a causal table. By registering, it is possible to learn as one of the expected events and provide it as knowledge useful for subsequent diagnosis.

According to the plant monitoring and diagnosing device and method of the present invention described in claims 3 and 4, not only the same operation as the plant monitoring and diagnosing device according to claim 1 is obtained, but also the scale of the causal table used for identifying the cause of abnormality. However, the time required for the diagnosis process can be further shortened because only the cause of the abnormality in which the influence of the abnormality is expected to appear first in the monitoring index identified as the starting point of the change can be further reduced.

According to the plant monitoring and diagnosing device of the present invention as defined in claim 5, when an abnormality that propagates an influence to a process signal in a wide range of the plant occurs within a short time, it should be a target for verification. There will be multiple causal tables for each system or subsystem, but by including anomalies in other systems or subsystems as cause candidates in each causal table, the causal relationship for each system or subsystem can be understood. The cause of abnormality can be diagnosed by the causal table of the system or subsystem closest to the cause.

According to the plant monitoring and diagnosing method of the present invention as set forth in claims 6 and 7, regarding the abnormality cause candidate identified in the diagnosis based on the above-mentioned causal table, the monitoring indexes in which abnormality signs are detected are arranged in the order of detection time. By rearranging and displaying or directly displaying in the event tree, it is possible to assist in grasping the relation of the abnormality sign of each monitoring index which is the basis of the identification.

According to the plant monitoring and diagnosing device of the present invention as defined in claims 8 and 9, the first diagnosing means based on the monitoring index peculiar to a specific abnormality is processed by a process which can be uniformly applied to many process signals. Since it is provided separately from the second diagnosing means based on the comparison between the abnormal sign pattern of the monitor index that can be calculated and the causal table, the processing time when the short-circuit diagnosis is possible can be shortened. On the contrary, when it takes a long time to calculate the unique monitoring index, such separation can eliminate the influence on the diagnostic processing time using the uniform monitoring index.

According to the plant monitoring and diagnosing apparatus and method of the present invention as defined in claims 10 to 13, only the monitoring index identified as the starting point of the abnormal change by the qualitative model is compared and collated with the case data. This makes it possible to easily search for similar cases and contribute to the estimation of the cause of abnormality. By performing the diagnosis based on the qualitative model first, it is possible to eliminate the need for the waveform matching for all the monitoring indexes showing the abnormal sign.

According to the plant monitoring and diagnosing apparatus and method of the present invention as set forth in claims 14 and 15, it is possible to confirm that a planned operation has been performed before performing the monitoring process, and to avoid the possibility of being affected by the operation. By excluding only a certain monitoring index from the target of abnormality detection, it is possible to prevent the monitoring / diagnosing device from issuing an unnecessary alarm and continue monitoring other monitoring indices.

According to the plant monitoring and diagnosing device of the present invention as set forth in claims 16 and 17, the monitoring process is executed even when a planned operation is performed, and an abnormal sign is detected in the monitoring index due to the operation. In some cases, the operation is identified by matching the observed symptom pattern with the symptom pattern in the causal table, or by matching the waveform of the monitoring index identified as the starting point of the abnormal change with the qualitative model with the case data. be able to. With this, it is possible to confirm whether or not the operation of the plant monitoring / diagnosing device is normal each time a planned operation is performed.

According to the plant monitoring and diagnosing method of the present invention described in claim 18, in the plant monitoring and diagnosing apparatus and method according to claim 14 and claim 15, among the monitoring indexes excluded from the monitoring target, a planned operation is performed. An analog monitoring index that can evaluate the magnitude of disturbance given to the plant is input, and each analog monitoring index that changes under the influence of that is considered as an output, and the transfer gain between the input and output during the disturbance occurrence period is monitored. The behavior of the plant can be monitored as if an active disturbance application test were performed.

According to the plant monitoring and diagnosing apparatus of the present invention as set forth in claim 19, the symptom pattern of the monitoring index is collated in consideration of the relationship introduced into the qualitative model, which can be kept constant by the feedback effect. Thus, even if there is a monitoring index in the middle of the propagation path that is not observed to change under the influence of the feedback effect, it can be identified as the propagation path.

According to the plant monitoring and diagnosing device of the present invention as defined in claim 20, the range for confirming the abnormality sign necessary for the diagnosis based on the qualitative model is the time delay given from the time when the signal on the causative side changes. By setting a limited time zone before and after the elapsed time, it is possible to detect the sign of change with high sensitivity.

According to the plant monitoring and diagnosing apparatus of the present invention as defined in claim 21, even when the influence propagation path of the abnormality identified by the qualitative model becomes a loop, it is looped based on the abnormality detection time of each monitoring index. The origin of can be identified. Further, even when a plurality of propagation paths can be established between the monitoring indexes, the true propagation path is identified by collating the time when the abnormal sign of each monitoring index is detected with the time delay given to the propagation characteristics. be able to.

According to the plant monitoring and diagnosing apparatus of the present invention, the response of the process signal is predicted from the operation of the protection sequence provided for the plant abnormality, and the observed behavior is predicted. It is possible to monitor the plant state, which changes from moment to moment, in real time by sequentially checking whether or not it matches with the movement.

According to the plant monitoring and diagnosing device of the present invention as set forth in claims 23 and 24, an abnormal symptom is displayed on the qualitative model or the plant system diagram according to claim 20, which is represented by a signed directed graph including a feedback effect. By displaying the detected monitoring index and the identified propagation path in different colors, it is possible to help the user to easily understand the relationship between the abnormal signs of each monitoring index.

According to the plant monitoring and diagnosing device of the present invention described in claim 25, the same effect as that of the plant monitoring and diagnosing device and method according to any one of claims 1 to 9 or 16 can be obtained by the cause identifying means based on the different symptom pattern matching method. Can be shown.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of a plant monitoring and diagnosing device of the present invention.

FIG. 2 is a flowchart showing the flow of processing of the first embodiment.

FIG. 3 is a diagram showing an effect of a qualitative model in which a feedback characteristic is introduced.

FIG. 4 is a diagram showing the effect of a time delay introduced on the effect propagation characteristic between monitoring indexes.

FIG. 5 is a diagram showing an example of real-time diagnosis using a qualitative model.

FIG. 6 is a flowchart showing a flow of abnormal propagation route identification processing.

7A and 7B are a block diagram showing a configuration of a monitoring / diagnosing unit of a second embodiment of the plant monitoring / diagnosing apparatus according to the present invention, and a diagram showing a causal table of monitoring indexes.

FIG. 8 is a flowchart showing a flow of monitoring diagnosis processing of the second embodiment.

FIG. 9 is a block diagram showing a configuration of a monitoring / diagnosing unit of a plant monitoring / diagnosing apparatus according to a third embodiment of the present invention, and a diagram showing a causal table of subsystems.

FIG. 10 is a flowchart showing a flow of monitoring diagnosis processing of the third embodiment.

FIG. 11 is a diagram showing a method of cause identification in the third embodiment.

FIG. 12 is a block diagram showing the configuration of a monitoring / diagnosing unit of a fourth embodiment of the plant monitoring / diagnosing apparatus according to the present invention.

FIG. 13 is a flowchart showing a flow of monitoring diagnosis processing of the fourth embodiment.

FIG. 14 is a diagram showing a method of comparing and collating with case data.

FIG. 15 is a block diagram showing the configuration of a monitoring processing unit of a fifth embodiment of the plant monitoring / diagnosing apparatus according to the present invention.

FIG. 16 is a diagram showing the contents of an operation information database.

FIG. 17 is a flowchart showing a flow of monitoring processing according to the fifth embodiment.

FIG. 18 is a diagram showing a monitoring method of a second monitoring processing means of the fifth embodiment.

FIG. 19 is a block diagram showing the configuration of a sixth embodiment of the plant monitoring / diagnosing apparatus according to the present invention.

FIG. 20 is an explanatory diagram showing a display example of an abnormal propagation route by a network between signals.

FIG. 21 is an explanatory diagram showing a display example of an abnormal propagation route by a plant system diagram.

FIG. 22 is an explanatory diagram showing a display example of a diagnosis result based on a causal table.

FIG. 23 is an explanatory diagram showing a display example of different diagnosis results based on a causal table.

FIG. 24 is a block diagram showing the configuration of a seventh embodiment of the plant monitoring / diagnosing apparatus according to the present invention.

[Explanation of symbols]

 1 control unit 2 signal input unit 3 monitoring processing unit 4 abnormal propagation path identification unit 5 abnormal cause identification unit 6 qualitative model database 7 causal table database 7a causal table database by monitoring index 7b subsystem causal table database 8 monitoring diagnostic unit 8a 1st Diagnostic means 8b Second diagnostic means 9 Diagnostic processing unit 10 Output display unit 11 Database management unit 12 Interactive processing unit 13 Case data extraction unit 14 Case database 15 Comparison collation unit 16 Operation identification unit 17 Operation information database 18a First monitoring processing unit 18b Second monitoring processor

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Shigeru Kanemoto No. 1 Komukai Toshiba-cho, Sachi-ku, Kawasaki City, Kanagawa Prefecture, Corporate Research & Development Center, Toshiba Corporation (72) Norimitsu Komai Shinsugita, Isogo-ku, Yokohama, Kanagawa Prefecture Town No. 8 Incorporated company Toshiba Yokohama Works (72) Inventor Takao Kageyama No. 8 Shinsugita-cho, Isogo-ku, Yokohama-shi, Kanagawa Prefecture Incorporated company Toshiba Yokohama Works (72) In-house Yasuo Ota 8 Shinsugita-cho, Isogo-ku, Yokohama-shi Kanagawa Address Stock Company Toshiba Yokohama Office

Claims (25)

[Claims] 1. A monitoring processing unit that detects deviation from a normal range of a monitoring index calculated from a process signal observed in a plant as an abnormal sign and creates a sign pattern of each monitoring index, and between the monitoring index. A qualitative model database in which a network model describing impact propagation characteristics is registered,
Anomaly propagation path identification unit that identifies the observed propagation pattern that matches the observed symptom pattern with a network model and the propagation path that is the starting point of the anomaly change and the influence propagation path due to the anomaly, and various anomaly causes for each system and subsystem of the plant A causal table database in which a causal table describing the symptom pattern of each expected monitoring index was registered, and a causal table for the system or subsystem of the relevant plant of the monitoring index identified as the starting point of the abnormal change was observed. Output the abnormality cause identification unit that identifies the cause of abnormality by collating with the symptom pattern and at least the identified influence propagation path, and if the symptom pattern corresponds to the causal table, also outputs the cause of abnormality A plant monitoring and diagnosing device comprising a display unit. 2. The plant monitoring and diagnosing device according to claim 1, further comprising a database management unit for performing learning by adding a symptom pattern corresponding to a causal table when the symptom pattern is not included. Plant monitoring and diagnostic equipment. 3. The causal table registers, for each monitoring index, a combination of various abnormal causes and symptom patterns in which the monitoring index is expected to be affected first. The plant monitoring and diagnosis device described. 4. The influence propagation characteristic between each monitoring index is obtained by detecting a deviation from the normal range of the monitoring index calculated from the process signal observed in the plant as an abnormal sign and using the obtained sign pattern of each monitoring index. By identifying the monitoring index that is the starting point of the abnormal change and its influence propagation path by comparing with the network model that describes the above, various abnormalities are assumed to appear first in the monitoring index that is the starting point of the change. A plant monitoring and diagnosing method characterized by identifying a cause of abnormality by comparing a causal table describing a symptom pattern for each cause of abnormality with an observed symptom pattern. 5. A monitoring processing unit that detects a deviation of a monitoring index calculated from a process signal observed in a plant from a normal range as an abnormal sign and creates a sign pattern of each monitoring index, a plant system, and a sub-system. For each system, an anomaly was observed by comparing the causal table describing the symptom pattern of each monitoring index that was assumed for various causes of anomalies including anomalies in other systems and subsystems with the observed symptom pattern. A plant that is not affected by other systems and subsystems and that is equipped with an abnormality cause identification unit that identifies a cause that can explain an abnormal symptom of another system or subsystem as a current abnormality cause. Monitoring and diagnostic equipment. 6. A deviation from a normal range of a monitoring index calculated from a process signal observed in a plant is detected as an abnormal sign, and the obtained sign pattern of each monitoring index is assumed for various abnormal causes. The results of diagnosis by checking the causal table describing the symptom pattern of each monitoring index are sorted, and the causes with a large number of monitoring indices with matching signs are sorted in the order of high probability as a possible cause candidate. In addition, the monitoring index is a method of monitoring and diagnosing a plant, which is arranged in the form of a causal table by rearranging the monitoring indexes in the order of earliest occurrence of abnormal signs. 7. A deviation of a monitoring index calculated from a process signal observed in a plant from a normal range is detected as an abnormal sign, and the resulting sign pattern of each monitoring index is assumed for various abnormal causes. As a result of diagnosis by comparing the symptom pattern of each monitoring index with the described causal table, the progress of the event including the abnormal symptom of each monitoring index caused by the identified cause of abnormality should be displayed in a tree structure. A plant monitoring and diagnosing method. 8. A first abnormality-detecting abnormality is identified by calculating a monitoring index specific to the abnormality from a process signal which reacts only to a specific abnormality occurring in a plant and detecting a deviation from a normal range thereof. From the diagnostic means and the current value of the process signal or its detection signal, the average value within a certain past period, the value obtained by smoothing the signal, and at least one of the present and past of at least one other process signal. The residual obtained by subtracting the value given as a function of the value as the predicted value is used as the monitoring index, and the deviation from the normal range is detected as an abnormal sign, and the sign pattern of each monitoring index is an assumed sign for various abnormal causes. A second diagnostic means for identifying a cause of abnormality by collating with a causal table describing a pattern, and executing the first diagnostic means and the second diagnostic means simultaneously in parallel, A plant monitoring and diagnosing device which outputs the result of the first diagnosing means prior to the result of the second diagnosing means. 9. A first abnormality-identifying cause is detected by calculating a monitoring index peculiar to the abnormality from a process signal which reacts only to a specific abnormality occurring in a plant and detecting a deviation from a normal range thereof. From the diagnostic means and the current value of the process signal or its detection signal, the average value within a certain past period, the value obtained by smoothing the signal, and at least one of the present and past of at least one other process signal. The residual obtained by subtracting the value given as a function of the value as the predicted value is used as the monitoring index, and the deviation from the normal range is detected as an abnormal sign, and the sign pattern of each monitoring index is an assumed sign for various abnormal causes. A second diagnostic means for identifying the cause of the abnormality by collating the pattern with the causal table, and executing the first diagnostic means to identify the cause of the abnormality, A plant monitoring and diagnosing apparatus, which outputs the result and does not execute the second diagnosing means, but executes the second diagnosing means only when the cause corresponding to the first diagnosing means is not identified. 10. A deviation pattern of a monitoring index calculated from a process signal observed in a plant from a normal range is detected as an abnormal sign, and a sign pattern of each of the obtained monitoring indexes is detected.
By comparing with the network model that describes the impact propagation characteristics between each monitoring index, the monitoring index that is the starting point of the abnormal change and its influence propagation path are identified, and the waveform of the temporal change of the monitoring index that is the starting point of the change is identified. , The monitoring index is compared with the temporal change waveform of the monitoring index calculated from the registered case data that is the starting point of the change, and the abnormality cause of the case that gives a high degree of similarity above a certain value is the current abnormality. A plant monitoring diagnosis method characterized by diagnosing a cause candidate. 11. A monitoring processing unit that detects deviation from a normal range of a monitoring index calculated from a process signal observed in a plant as an abnormal sign and creates a sign pattern of each monitoring index, and between the monitoring index. An abnormal propagation path identification unit that identifies the monitoring indicator that is the starting point of the abnormal change by comparing the network model describing the effect propagation characteristics with the observed symptom pattern and the impact propagation path, and the starting point of the change of the monitoring index. The case data extraction unit that searches registered registered cases that have become and calculates the relevant monitoring index, and the waveform of the temporal change of the monitoring index that is the starting point of the change and the waveform of the temporal change of the extracted case data. A plant monitoring diagnostic device characterized by comprising a comparison and collation unit for evaluating similarity and outputting the cause of abnormality of a case giving a high degree of similarity of a certain value or more as a candidate for the current cause of abnormality. Place. 12. The plant monitoring and diagnosing method according to claim 10, wherein in collation with case data, a fixed period before and after the change detection of the monitoring index which is the starting point of the change is taken out as time series data and taken out in the same manner. A plant monitoring and diagnosing method, wherein a maximum value of a correlation function calculated with case data having the same period length is used as the similarity. 13. The plant monitoring and diagnosing method according to claim 10, wherein, in collation with case data, time-series data obtained by extracting a period t0 before and after a change in a monitoring index that is a starting point of change is extracted as a basic function f (t. ), And case data x
Evaluation function calculated between (t) and A method for monitoring and observing a plant, wherein the maximum value of is the similarity. 14. When detecting a deviation from a normal range of a monitoring index calculated from a process signal observed in a plant as an abnormal sign, the deviation is detected for each planned operation artificially or mechanically performed on the plant. When the change in the monitoring index that occurs with the start and end, the monitoring index that may change under the influence of the operation, and the period in which the change appears are registered in advance, and when the change accompanying the start of the operation is detected, A method for monitoring and diagnosing a plant, characterized by excluding an affected monitoring index from an abnormality detection target until a change is detected for a certain period or with the end. 15. A monitoring processing unit that detects a deviation from a normal range of a monitoring index calculated from a process signal observed in a plant as an abnormal sign, and a planned operation that is artificially or mechanically performed on the plant. The operation information database that pre-registers the change in the monitoring index that occurs with each start and end, the monitoring index that may change under the influence of the operation, and the period in which the change appears, and the registered operation information. The monitoring processing unit issues a command to calculate the monitoring index based on the operation and to exclude the monitoring index affected by the operation from the abnormality detection target from the time when the change associated with the start of the operation is detected until the change associated with the fixed period or the end is detected A plant monitoring and diagnosing device, comprising: 16. The causal table includes a symptom pattern of each monitoring index assumed for a planned operation performed artificially or mechanically on a plant. The plant monitoring and diagnosis device according to claim 3, 4, or 5. 17. The plant monitoring and diagnosing device according to claim 11, wherein the case data includes case data in which a planned operation is manually or mechanically performed on the plant. 18. When detecting a deviation from a normal range of a monitoring index calculated from a process signal observed in a plant as an abnormal sign, the deviation is detected for each planned operation artificially or mechanically performed on the plant. The change of the monitoring index caused by the start and the end, the analog monitoring index X which changes first by the operation and the other analog monitoring index {Yi} which may change under the influence, and the period in which the change appears are set in advance. The relative change rate {Yi / X} is newly added to the analog monitoring index {Yi} from the time when the change associated with the start of the operation is detected until the change associated with the fixed period or the end is detected. A plant monitoring and diagnosing method characterized by detecting deviation from the normal range as an index as an abnormal sign. 19. The plant monitoring and diagnosing device according to claim 1, 2, 3 or 11, wherein the abnormal propagation path identifying unit introduces a relationship in which influence propagation characteristics between monitoring indexes are held constant by feedback control, A plant monitoring / diagnosing device characterized in that a propagation path is considered to be established even if the observed value of a monitoring index has no apparent effect due to the suppression of a transient change due to a feedback effect of a control system. 20. The plant monitoring and diagnosing apparatus according to claim 1, 2, 3 or 11, wherein the abnormal propagation path identifying unit delays the influence propagation characteristic between the monitoring indices by a time delay, and the monitoring processing unit causes a side error. If an abnormal symptom is detected in the monitoring index of the above and not detected in the monitoring index on the result side, the time range and the range of normal values for extracting the characteristics of changes in the monitoring index on the result side are extracted based on this time delay. A plant monitoring and diagnosing device characterized by identifying an abnormal propagation path after re-evaluating abnormal signs for a limited time. 21. The plant monitoring and diagnosing device according to claim 1, 2, 3 or 11, wherein the abnormal propagation path identification unit delays the influence propagation characteristic between the monitoring indices by a time delay and identifies the identified propagation path. When a closed loop is configured, the monitoring index in the closed loop where the abnormal sign is first detected is used as the starting point of the closed loop, and when different propagation paths are apparently established between the two monitoring indexes, the abnormality in the causal side monitoring index A plant monitoring and diagnosing device characterized by estimating a true propagation route by comparing a time obtained by adding a propagation time delay to a sign detection time with an abnormal sign detection time of a monitoring index on the result side. 22. The plant monitoring and diagnosing apparatus according to claim 1, 2, 3 or 11, wherein the abnormality propagation path identifying unit propagates an influence of a protection sequence provided for a plant abnormality on a monitoring index. The characteristics of the plant are included in the qualitative model, and the response of the plant is monitored by comparing the abnormal symptom of the monitoring index observed from the time when the operation of the sequence is detected with the symptom of the subsequent monitoring index predicted based on the qualitative model. A plant monitoring and diagnostic device characterized by the above. 23. The plant monitoring and diagnosing apparatus according to claim 1, 2, 3, 11, or 19, wherein the output of diagnostic results by the anomalous propagation path identification unit is a propagation characteristic with a positive gain and a propagation characteristic with a negative gain. , A diagram in which the names of the process signals corresponding to the monitoring index are connected by three kinds of directed line segments of the propagation characteristics showing the feedback effect, and the observed change behavior of the monitoring index is increased, decreased, or unchanged 3 The result of classification into types is displayed together with the name of the process signal, the origin of the change and the name of the identified process signal are highlighted, and the color of the directed line segment corresponding to the identified influence propagation path is changed. A plant monitoring / diagnosing device characterized by being displayed. 24. The plant monitoring and diagnosing device according to claim 1, wherein the output display of the diagnosis result by the anomalous propagation path identifying unit is an observation point of a process signal and a signal name in a plant system diagram. , The change behavior of the monitoring index calculated from the process signal is increased, decreased,
And the result classified without change is displayed together with the process signal name, and the process signal corresponding to the monitoring index identified as the starting point of the change and the upstream device that directly affects the signal are highlighted. A plant monitoring and diagnosing device characterized in that control circuits, pipes, valves, pumps and the like existing in the identified influence propagation path are displayed in different colors. 25. The plant monitoring and diagnosing device according to claim 1, wherein the relationship between various abnormal causes and planned operation and the symptom pattern of each expected monitoring index is a tree-type causal tree or production. A plant monitoring and diagnostic device characterized by being expressed as a format rule.