JPH022404A - Device for fault diagnosis of plant - Google Patents

Abstract

PURPOSE:To easily diagnose a fault by inspecting discrepancy in an fault logic, and at the time of generating discrepancy, assuming the discrepancy as a single fault or the like of a sensor and inspecting whether the logical discrepancy can be removed or not. CONSTITUTION:The plant fault diagnosis device inputs a detecting signal from a sensor 1 to an I/O data processor 7 through a sensor signal input device 2 and an interface device 3 to process the signal. On the other hand, measuring data from an operator are inputted through a CRT 4, a keyboard 5 and an interface device 6. The diagnostic device is also provided with a diagnostic logic storage device 9. In this case, a logical discrepancy inspector 20 and a diagnosis computing element 21 are added. Whether discrepancy is generated in logic or not is checked simultaneously with normal diagnosis, and when the cause of the discrepancy can not be diagnosed, diagnosis is started from a position generating the discrepancy and advanced by an assumption or the like narrowing down an error for removing the discrepancy. At the time of diagnosing the cause, the result is displayed, and when the discrepancy can not be removed, the discrepancy is displayed as two or more errors.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a plant failure diagnosis device for diagnosing failures in various plants such as nuclear power, thermal power, chemical, and pump plants, and in particular has an observation f-ta error automatic repair function. The present invention relates to a plant failure diagnosis device equipped with a plant failure diagnosis device.

[Conventional technology]

FIG. 6 is a block diagram showing the configuration of an example of a conventional plant failure diagnosis device, in which a detection signal from a sensor 1 is sent to an input/output data processor via a sensor signal input device 2 and an interface device 3. 7 and a diagnostic calculator 8. A detection signal from the sensor 1 is input to an input/output data processor 2, where it is processed as observation data. On the other hand, observation data is input by the operator using the CRT 4, key gate 5, and interface device 6. Normally, observation data is displayed on the CRT 4, and whether or not it holds true is input in a conversational manner. The input/output data processor 7 receives observation data,
It also presents operator observation data and displays diagnostic results. The diagnostic logic storage device 9 stores in advance the fault diagnosis logic shown in FIG. 7, which will be described later.
For example, the abnormal value of potentiometer 11 is input/output data processor 1.
is automatically called when input to the diagnostic calculator 8 via. Then, based on the observation data input to the input/output data processor 7, the diagnostic calculator 8 performs the above-described diagnosis. When input from the operator is required, and when the results are displayed, the results are displayed on the CRT 4 via the input/output data processor 2 when it is necessary to display the diagnostic process (fault logic) in response to a request from the operator. or the key is -C5
Perform input from .

Hereinafter, the operation of the conventional plant failure diagnosis apparatus configured as described above will be explained with reference to FIGS. 7 to 11. FIG. 7 shows an example of a motor drive circuit to be diagnosed and its failure diagnosis logic. The upper part of FIG. 7 shows a system in which the motor 10t is driven by the battery 12. When the value of the potentiometer 11 shows an abnormal value, diagnosis is started using this as a trigger using the fault logic shown in the lower part of FIG. This failure logic is a pool algebraic expression of the causal relationship between events, observed data, and causes.
It is possible to construct logic using the graphical symbols shown in FIGS.

In FIG. 7, motor 10 is driven by battery 12. In FIG. However, switch JJ (SWJ), switch J 4
(SWJ) is usually closed on one side or both sides. A potentiometer mll is attached to the motor 10 and monitors the position of the motor. An ammeter 15 and a voltmeter 16 are added as sensors.

In FIG. 7, the true cause of failure is "battery failure." First, "motor motion" is established.

When no current is flowing, "current I=0" is established based on the value of the ammeter 15, and as a result, the event "current does not flow through the motor" is established. Furthermore, if the battery voltage is insufficient, the value of the voltmeter 16 establishes the condition ``battery voltage <IJ'', and as a result, the event ``battery voltage is abnormal'' is established, and the cause is diagnosed as ``battery abnormality''. be done.

In FIGS. 8 to 11, E1 to E8 indicate events,
AI, Aj, 9-D indicate observation data, and F, Fl
, F2 indicates the cause of the failure. Observation data AI, AZ
, B-D are attached to the main body of the diagnostic system, and whether the condition is established or not is determined based on the values of nine sensors and the input value from the operator. For example, in Fig. 7, the observed data "current I = QJ" and "current I ~ 0" are automatically determined by the value of the ammeter 15, and "battery voltage ≧ E" and "battery voltage
” is automatically determined based on the value of the voltmeter 16. The dotted line in the figure shows this relationship, and the observation data [switch BWI, S
[WJ both open] is determined by having the operator input whether it is true or not.

As mentioned above, when the diagnosis is started, the event "motor action" is established, and based on what is automatically determined from the observation data, it is established one after another according to the logic shown in Figures 8 to 11. Determine the events that will occur.

Diagnosis proceeds toward events that hold true, and if the observed data requires operator input, it is presented to the operator and the operator is asked to input whether it holds true or not. If the logic can be traced to the cause, the cause can be presented to the operator as the cause of the failure.

[Problem to be solved by the invention]

The conventional diagnostic plant failure diagnosis apparatus described above is performed on the assumption that the sensor signals and operator inputs are always correct. For this reason, the ammeter 15 and voltmeter 16 in FIG.
When a sensor such as the above malfunctions, the cause of the malfunction logic clasp may be denied and diagnosis may not be possible. for example,
In the diagnosis example shown in FIG. 7, if the true cause is "switch SW1.", but "battery voltage ≧E" is established because the voltmeter 16 is malfunctioning, the operator input "switch SW1. 8W2 both open'' will not hold true, and the diagnosis will be unreliable for any cause.

As described above, the conventional device has a problem in that if a sensor or operator input is erroneous due to a single failure, an inconsistency occurs in the logic and diagnosis becomes impossible.

The present invention can present a possible single point to the operator when diagnosis cannot be made due to a single failure in sensor or operator input data, and without having to correct the failure point and perform the diagnosis again. It is an object of the present invention to provide a brown fault diagnosis device that performs diagnosis and presents the cause to an operator.

[Means to solve the problem]

In order to achieve the above object, the present invention provides a conventional plant fault diagnosis device with: a first means for checking whether or not a contradiction has occurred in the fault logic; A second method assumes a single error in an observation data input means other than a sensor and checks whether the logical contradiction can be resolved; This is the addition of a third means of presenting the location and the diagnosis result in that case.

[For production]

Since the present invention is configured as described above, at the same time as performing normal diagnosis, it is checked whether or not a contradiction has occurred in the logic, and if the cause cannot be diagnosed, starting from the place where the contradiction has occurred, the contradiction is corrected. If the errors to be removed are narrowed down and the cause is diagnosed by proceeding with the diagnosis, the results can be presented to the operator. Furthermore, if the contradiction cannot be resolved, there will be errors in two or more places, and this can be indicated to the operator.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings. 1 is a block diagram showing the configuration of an embodiment of the present invention,
The difference is that a logical contradiction checker 20 and a diagnostic calculator 21 are added to the conventional device shown in FIG. 6 described above.

Hereinafter, these 20.21 will be explained with reference to FIGS. 2 and 3. FIG. 2 is a processing flow when a contradiction occurs in the failure logic. The logical contradiction checker 20, when the diagnostic calculator 21 is performing diagnosis in the same way as the conventional device,
Each block connected in a logical relationship as shown by the broken line or the dashed-dotted line in FIG. 3 is checked for any inconsistency. Since there is usually one cause of failure, if a discrepancy occurs, diagnosis will not be able to estimate the cause. Here, the logic contradiction refers to, for example, although the event E1 holds true in FIG. 3, the observed data Bll and B12 both do not hold, and the cause cannot be traced to the logical cause. The presence or absence of a contradiction is checked for each block, and if the cause cannot be estimated, a flow as shown in FIG. 2 is started. The logical contradiction checker 20 creates a list of failure locations for resolving contradictions. In the above example, Bll and B10 are listed as candidates. At the same time, the establishment of E1 is also suspected. AI
This is because if E1 is a failure, E1 is erroneously established. However, since A2 is not true, if AI is not true, then
This block causes a contradiction again. For this reason, A1
However, the failure is denied and the only candidates are B11 and T312. These processes are performed by the logical contradiction checker 20.

Next, the diagnosis calculation unit 21 performs a diagnosis based on the fact that B11 is not established as an error and that B11 is established. As a result, the cause of failure F1 can be estimated. The diagnostic calculator 21 informs the input/output data processor 7 and interface 6 of the CRT 4 of FIG.
displayed via. Furthermore, the diagnostic calculator 21 checks whether any error candidates remain. Since B12 remains, if C121 and C122, which are similarly diagnosed assuming that 912 holds true, will cause a contradiction again,
The establishment of B12 is denied. Since no candidates remain, the diagnosis ends. If we proceed with the diagnosis assuming an error, a contradiction will appear again if the observed data is incorrect in two or more places.
In this case, stop the diagnosis.

If all candidates are denied, the establishment of the first event itself becomes doubtful. Therefore, it is displayed to the operator that re-diagnosis is not possible.

For example, as error candidates in FIG. 7 for explaining the conventional device, "switches sw1.sw2 both open" and "
Battery voltage ≧E” is increased.

Just to be sure, "Battery voltage <EJ" means "Battery voltage ≧ E
'', the establishment of one is a form of negating the other. (These are predefined in the logic) Therefore, the following two results are displayed as a result of re-diagnosis.

■If "[Switch 8W1, SW2 both open" is not satisfied]
There is a possibility of a. In that case, the probable cause is "Switch 8W2 failure" or "Switch 8W2 failure.""There is a possibility that ``・Cut-kuri-voltage≧E'' is established incorrectly. In that case, the probable cause is ")"
is. In the embodiment of the present invention described above, if the ammeter is malfunctioning and "current I-1?OJ" is established, "motor failure" is erroneously assumed. In other words, the present invention is applicable to cases where there is redundancy between observation data of sensors (if one sensor value is incorrect, it contradicts the others). However, by expanding the diagnostic rhythm so that the operator can input that the presumed cause is an error, re-diagnosis becomes possible even if there is no redundancy between observed data. The present invention (in which the operator's input value corresponds to redundancy) includes the above-described embodiment and re-diagnosis by denying the probable cause by the operator.

Furthermore, it is easy to understand that the present invention is applicable to a wide range of applications, such as diagnosis of plants other than the embodiments described above. In order to specifically explain this, an example in the pump plant shown in FIGS. 4 and 5 will be described below.

FIG. 4 shows a part of the piping system of the pump plant, and FIG. 5 shows the fault logic for diagnosis when an abnormality occurs in which the main pump runs out of lubricating water.

Figure 4 shows the piping system 25 and its surrounding systems that cool the lubricating oil via the reducer lubricating oil cooler 26 with fresh water 27 in Bonn f7'' land and discharge the fresh water 27 to the absorption water tank 28. It is.

This is a part of the four plants and has been omitted for convenience of explanation. The flow relay 23 detects that the main four-wheel lubricating water is "cut". “ON” of reducer 70 relay 29
“OFF” and “OPEN” and “CLOSE” of the lubricating water solenoid valve 31 are automatically input. Other observation data is input by the operator. The site a-SO1 manual valve 24 is shown as an observation point. It goes without saying that a control power source for the solenoid valve 31 is provided, although it is not shown in the rounded diagram, which is not a piping system. In addition, in FIG. 4, 32 is an engine, 33
34 shows the reducer and the main pump.

As shown in the failure logic of FIG. 5, portions to which observation data from sensors such as the flow relay 23, the speed reducer flow relay 29, and the electromagnetic valve 31 are automatically input are indicated by broken lines. Furthermore, adjacent observation data are in contradictory states, and since one of them is sure to hold true, at least one cause candidate is required. This failure logic includes "denial of cause candidates by the operator" as a definition of a logical error. In other words, when the operator inputs that the estimated cause is incorrect (denial), it searches for observation data that currently does not hold true, and determines whether any of the observed data is incorrect. The cause in that case is estimated and displayed to the operator. Naturally, the observation data paired with the inverted observation data is inverted from valid to unconfirmed. For example, ■ ``Reducer flow relay ``ON'' is established (the paired ``Reducer flow relay ``OFF'' is not established) ■ ``Site flow water flow ``Yes'' is established (the paired ``NO'' is not established) If the main pump flow relay failure is diagnosed. Here, if the operator inputs that the main pump flow relay failure is incorrect, the diagnosis is continued by reversing the "reducer flow relay "disconnected" state to true (the "on" state becomes false). .

For example, if the following input is made, it is displayed that there is a possibility that the sensor of the speed reducer flow relay is malfunctioning, and that the solenoid valve itself is malfunctioning.

■r Lubrication water solenoid valve ``closed'' is established ■``Control power supply normal'' is established As stated above, the logic error definition is not limited to a narrow sense, and can be applied to the entire pump plant as a diagnostic target. This indicates that

〔Effect of the invention〕

According to the present invention, when it is impossible to diagnose the sensor or operator input data due to a single failure, the location where the single failure may occur can be presented to all operators, and the failure location can be corrected. It is possible to provide a plant failure diagnosis device that can execute a diagnosis and present the cause to an operator without having to perform the diagnosis again.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the plant failure diagnosis device according to the present invention, FIGS. 2 and 3 are diagrams for explaining the operation of FIG. 1, and FIGS.
Each figure is a diagram for explaining the scope of application of the present invention, FIG. 6 is a block diagram showing the configuration of an example of a conventional plant failure diagnosis device, and FIG. 7 is an example of the diagnosis target and failure logic of FIG. 6. and FIGS. 8 to 11 are diagrams showing examples of failure logic symbols in FIG. 6, respectively. 1...Sensor, 2...Sensor signal input device, 3...
・Interface device, 4... CRT, 5... Keyboard, 6... Interface 7Ace, 7... Input/output data processor, 20... Logical contradiction checker, 21... Diagnostic calculation 9...Diagnostic logic storage device. Applicant's agent Patent attorney Takehiko Suzue Figure 2 Figure 1 Condolences 4 Figure 6

Claims (1)

[Claims] An input device that takes in observation data from a sensor attached to a plant to be diagnosed and/or observation data from an observation data input means other than the sensor, and a failure logic in which failure diagnosis logic is stored in advance. In a plant fault diagnosis device comprising: a storage device; and a diagnostic computing unit that identifies a cause of a failure based on observed data input to the input device according to a fault diagnosis logic stored in the fault logic storage means; a first means of checking whether a contradiction has occurred; and, if a contradiction has occurred, checking whether the logical contradiction can be resolved by assuming a single failure in the sensor or a single error in an observation data input means other than the sensor; a second means for diagnosing the single error, and a third means for presenting the error location and the diagnosis result in that case when the cause can be diagnosed without logical contradiction assuming the single error; .