JPH011009A - Status understanding processing mechanism for plant operation support equipment - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a situation understanding processing mechanism for a plant operation support device used in plants such as nuclear power plants and thermal power plants.

[Conventional technology]

As an example of a conventional plant operation support system, as shown in FIG. Candidate extraction knowledge memory 20,
It consists of an abnormal area candidate verification knowledge memory 21 , a countermeasure determination processing knowledge memory 22 , a working memory 1 , a display device 2 , a response input device 13 , and a data bus 14 . In addition, the situation grasp processor groove 23 in Figure 1 (fc surrounded by the broken line in the figure)
105 in FIG. 4.

In other words, the situation understanding processing mechanism 105 includes a data processing unit 101 that instructs all process variables to perform data processing necessary for subsequent processing every Δt, a memory 103 that stores knowledge for the data processing, and a plant processing unit 103 that stores knowledge for the data processing. It consists of an abnormality detecting section 102 that determines whether the abnormality is normal or abnormal, and a memory 104 that stores knowledge for detecting the abnormality.

In the conventional equipment described above, all plant variables were considered to be equally important for abnormality diagnosis.

゛ [Problem to be solved by the invention] Conventionally, all plant variables are considered and diagnosed with the same importance. 9. The calculation time required for the driving support device including the countermeasure decision processing mechanism 10 increases;
There are cases where the sampling time Δt does not fit. Furthermore, if the sampling time Δt is increased, early abnormality detection may be lost.

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a situation grasping processing mechanism for a plant operation support system that can shorten the sampling time interval and trigger the start of diagnosis at an early stage.

[Means for solving problems]

In order to achieve the above object, the present invention directly reflects abnormal events among the input process variables that are directly related to plant trips and require emergency operation operations to avoid them. A 7' A process variable selection part for selecting a process variable and a B process variable not directly related to a plant trip, a memory storing process variable selection knowledge, and a memory storing process variable selection knowledge for each sampling time Δt for collecting data of the process variable. , a data processing unit that instructs data processing of the A process variable, a memory that stores k in the A process variable data processing unit, and a memory that stores k in the A process variable data processing unit, and a memory that stores k in the A process variable data processing unit, and determines whether the A process variable is normal or abnormal, or whether it becomes abnormal after a sampling time Δt. A memory that stores abnormality detection knowledge of an abnormality detection unit and A process to determine, and data processing necessary for subsequent processing in the B process variable at the time of the abnormality or abnormality prediction and every ΔtXN (N is an integer greater than 1) time. A data processing unit that instructs and a memory that stores B7°0 process variation data processing knowledge, an abnormality detection unit that determines whether the B process variable is normal or abnormal from the data processing result of the B7' process variable, and a B process variable abnormality. It consists of a memory that stores detection knowledge.

[Effect]

By configuring as described above, the collected process variables are categorized into those related to abnormal events that are directly related to plant trips and require urgent operational actions to avoid them, and those that are related to emergency operational actions. It is divided into those related to abnormal events that do not require Then, an abnormality determination is performed every sampling time Δt using the former as a pace, and the latter data is also used to detect an abnormality at the time of this abnormality detection and every NXΔt. Therefore, the sampling time interval can be made shorter than in the past, and the diagnosis start trigger can be triggered earlier.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an example of a plant operation support system using the situation understanding processing mechanism 23 of the present invention. Situation understanding processing mechanism 2
3 is a process variable selection section 3, an A process variable data processing section 4 (in the figure, it is referred to as A variable data processing section 4), and an A process variable data processing section 4.
Variable abnormality detection unit 5, B variable data and process variable selection knowledge memory 1 that stores knowledge necessary for each process.
5, A variable data processing knowledge memory 16, A variable abnormality detection knowledge memory 17, B variable data processing knowledge memory 18, B
It is composed of a variable abnormality detection knowledge memory 19.

The A and B variable data processing units 4 and 6 calculate the slope of change in the target variable, the average value of specific time interval data, the dispersion spectrum, etc. Further, the A and B variable abnormality detection units 5 and 7 compare the target variable with a redundant variable or a set value, and compare the slope, average variance, spectrum, etc. with a threshold value, and detect that it is larger than the threshold value. An abnormality is detected, and there is no difference in processing content between A and B variables.

The configurations other than the situation grasping processing structure 23 described above are configured in the same manner as the conventional apparatus, and the functions of each configuration will be explained below.

The control mechanism 1 performs the processes shown in FIG. 3, namely, input saturation SIO of process variables, situation understanding process S11, extraction of abnormal area candidates S12, verification of abnormal area candidates S13. In order to sequentially and smoothly control the driving guide determination 814, the processing completion signal of each processing mechanism is received and the activation of the next processing is issued. That is,
The control mechanism 1 is always involved between the processes.

The input processing mechanism 2 inputs process variables such as pressure, flow rate, and valve opening degree via a detector installed in the plant, stores the values in the working memory 11 via the data bus 14, and processes them. Return a completion signal to the controller 'fsl J.

When the abnormal area candidate extraction processing mechanism 8 is activated by the control mechanism 1, it uses the knowledge stored in the abnormal area candidate extraction knowledge memo I720 and uses the situation understanding processing results as data to extract several information estimated by so-called data-driven current reasoning. Extract abnormal areas. That is, if the result of the situation grasping process is a decrease in the water supply flow rate, abnormal parts such as the water supply flow rate sensor, control system, valve drive unit, valve body, etc. are extracted. This result is stored in the working memory 11 via the data bus 14, and a processing completion signal is returned to the control mechanism 1.

Abnormal area candidate verification department! ! l! The machine groove 9 is activated by the controller +131, and uses the knowledge stored in the verification knowledge memory 21 for each abnormal region candidate extracted by the abnormal region candidate extraction processing mechanism 8 to grasp the input value and situation of the process training. Using the results as data, verify whether each candidate is the cause of the abnormality. Driving guide candidates are extracted for the area where the abnormality has been verified and stored in the working memory 11 via the data bus 14, and a processing completion signal is returned to the control mechanism 1.

When activated by the control mechanism 1, the countermeasure determination processing mechanism 10 uses the knowledge stored in the countermeasure determination processing knowledge memory 22 to determine which driving guide is optimal among the driving guide candidates for the abnormal area extracted in the verification process. Determine whether there is. This result is stored in the working memory 11 via the data bus 14 and a processing completion signal is returned to the control mechanism 1.

The abnormal area candidate extraction knowledge memory 20 identifies the abnormal area that is the cause of the process volume status indicating that an abnormality has occurred, such as "lower", "higher", "increased", or "decreased" than the normal state. It indicates which part or which device it belongs to, and is expressed and stored in production bar or frame format.

The abnormal area candidate verification knowledge memory 21 stores knowledge including the following items for each abnormal area in the form of seven frames or objects.

■ Process values and threshold values necessary to verify that the area is abnormal ■ Logical formula for determining whether the abnormality is 1 or more ■ Candidate measures for driving guides to be taken when the area is abnormal The decision processing knowledge memory 22 stores knowledge including the following items in the form of frames and objects for each maintenance method candidate extracted in the verification process.

■ Conditions for deciding whether to adopt the driving guide ■ Conditions for ordering the adopted multiple driving guides Working memory II temporarily stores the results of each process 0 Display device 12 is abnormal Area candidates, their verification results, driving guides, and questions that the system asks the user are displayed. The response input device 13 allows the user to input response data to questions posed by this system.

The specific knowledge contents of the abnormal area candidate extraction knowledge memory 20, the abnormal area candidate verification knowledge memory 21, and the countermeasure determination processing knowledge memory 22 are as follows, for example, in the m main water supply control system. That is, the memory 20 indicates that the main water supply flow t is lower than the main steam flow rate by at least a% of the rated flow rate;
If the steam generator water level is lower than the set value by b% or more, the main water supply controller or control valve, the main water supply pump controller, the main water supply pump, or the water supply piping is determined to be an abnormality candidate.

The memory 21 also shows that the main water supply control valve lift is the lift set value (
If it is 0% or more lower than the controller output), set variable U1 to True.
Let it be e.

When the booster relay output pressure is lower than the set value by more than de, variable U2 is set to True.

If both Ul and U2 are True, the confidence level is set to X.

If only Ul is True, the confidence level is set to y.

Maintenance method 1: Monitor lift operation.

Maintenance method 2: Switch the air system.

Furthermore, the memory 22 monitors the lift operation by r when the confidence level is less than or equal to y.

If the confidence is greater than y, the air system is switched.

On the other hand, the situation understanding processing mechanism 23 is controlled according to the flowchart shown in FIG. In other words, the process variables collected by the process input processing mechanism 2 are classified into process variables (hereinafter referred to as process variables) that are directly related to plant trips and that directly reflect abnormal events that require emergency operation operations to avoid them. (hereinafter referred to as A variables) and collected process variables other than this A variable (hereinafter referred to as B variables).
SL).

Next, perform appropriate data processing operations (S2) on the A variable,
Using this result, abnormality detection is performed based on the A variable (S3
). If it is abnormal, perform appropriate data processing on the Bf number (Sl
), and from this result, abnormality detection is performed based on the B variable (S5
). In this case, regardless of whether it is normal or abnormal, A variable is abnormal or B
The process advances to the next abnormal region candidate extraction processor 8 to extract abnormal region candidates estimated from the variable abnormalities. The above is performed every sampling time Δt.

Even if the Herrhison number is determined to be normal (S3), appropriate data processing is performed on the B variable every Nth data sampling time Δt (S6) (S7), and abnormality detection based on the B variable is performed from this result. Do it (S8).

If an abnormality is detected, the process proceeds to the next abnormal area candidate extraction processing mechanism 8, but if normal, the process returns to the process input processing mechanism 2.

When the A variable is determined to be normal and does not correspond to the 8th time of Δt, it is returned to the process input processing mechanism 2.

As described above, abnormality detection is performed on Δti based on the A variable, and if abnormal, abnormality detection is also performed on the B variable, and every 8th time of Δt, abnormality detection on the B variable is performed even if the A variable is normal.

In the process variable selection section 3 and the process variable selection knowledge memory 15, for example, in a PWR main water supply control system, process variables are specifically selected as follows. That is, the A variables are main feed water flow rate, steam generator water level, valve lift, controller output main steam flow rate, turbine initial stage pressure, etc., and the B variables are I/P converter output pressure, pilot positioner output pressure, booster relay. output pressure, etc. Then, as data processing using the A variable, the steam generator water level setting value is calculated from the turbine first stage pressure.

Further, abnormality detection for the A variable is as follows. That is, when the main water supply flow rate is lower than the main steam flow rate by more than e% of its rated flow rate, the main water supply flow rate is abnormal.

If the steam generator water level is lower than the set value by f% or more, the steam generator water level is abnormal.

Then, the deviation between the control valve lift and the controller output is calculated.

It is considered abnormal when the valve lift is lower than the controller output by g% full school or more.

As data processing using B variables, the estimated value of booster relay output pressure is calculated from the pilot position and output pressure.

An abnormality is detected in the B variable when the booster output pressure is lower than its estimated value by h% full scale or more.

The subsequent processing follows the memories 20, 21, and 22.

To specifically separate the A variable and the B variable, the A variable is used for signals used in the protection system, such as neutron system, coolant temperature, pressurizer pressure, water supply flow, narrow area SG water level, etc., or used for alarms. This is done based on the controller output signal, such as the pressurizer reference water level, SG reference water level, etc.

The B variable is determined by a variable other than the A variable, such as a controller output signal not directly connected to the operating end, a wide area SG water level, etc.

The sampling time Δt is set in a node-wise manner. Process input is performed with priority over internal software processing by interrupt processing based on this. However, priority is given to internal processing only when an abnormality is detected in the A variable, and if processing time exceeding Jt is required, Prevent interrupt processing.

Here, the reason why the abnormality detection of the B variable is set to N or Jt will be explained.

If an abnormality occurs in the A variable (at the time of Δt), the B variable must be checked. At this time, if the processing time exceeds Jt, this is allowed. However, an abnormality is not detected in the A variable (at Δt) and an abnormality occurs in the B variable (at NXjt), and Jt
This will not be allowed if it exceeds the limit, so in the latter case, it is thought that the former can be used.

#) B variables are defined as variables that are not directly involved in plant trips and variables that are related to abnormal events that do not require emergency measures, and a delay in abnormality detection for B variables compared to A variable can be tolerated. This is because an integer larger than 1, N=1.2..., is used.

It should be noted that it is important to note that the B variable is always considered when the A variable is abnormal, and that the B variable is not always considered when the A variable is normal.Furthermore, in this case, even if it is not considered, it will not immediately lead to an abnormal situation. This is because when an abnormal situation progresses, the A variable is always detected as abnormal. Under these circumstances, there is no problem in not allowing over Δt when the system variables are normal.

〔Effect of the invention〕

According to the present invention described above, by reducing the number of process variables that are constantly monitored, the sampling interval can be made smaller than before, and by early detection of plant abnormalities that require urgent treatment, the trigger for starting diagnosis can be made earlier. It is possible to provide a processing mechanism for grasping the situation of plant operation support equipment that can occur at any time.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an example of a plant operation support device equipped with a situation understanding processing mechanism according to the present invention, and FIG.
FIG. 3 is a flowchart for explaining the operation of the situation grasping processing mechanism shown in FIG. 3. FIG. 3 is a diagram for explaining the operation of the plant operation support device shown in FIG. be. 3... Process variable selection section, 4... A variable data processing section, 5... A variable abnormality detection section, 6... B variable data processing section, 7... B variable abnormality detection section, 15 ...Process variable selection knowledge memory, 16...A variable data processing knowledge memory, 17.-A variable abnormality detection knowledge memory, 18
...B variable data processing knowledge memory, 19...B variable abnormality detection knowledge memory. Applicant's agent Patent attorney Takehiko Suzue○
ingredient

Claims (1)

[Claims] Among the input process variables, A process variable is directly related to a plant trip and directly reflects an abnormal event that requires emergency operation to avoid it; A process variable selection unit that selects a B process variable that is not directly related to a plant trip, and a memory that stores process variable selection knowledge; a data processing unit that instructs the A process variable, and a memory that stores A process variable data processing knowledge; an anomaly detection unit that determines whether the A process variable is normal or abnormal, or whether it becomes abnormal after sampling time Δ_t; and an A process abnormality detection unit. a memory that stores knowledge; a data processing unit and a B process that instructs the B process variable to perform data processing necessary for subsequent processing at the time of the abnormality or abnormality prediction and every Δ_t×N (N is an integer greater than 1) time; A plant operation support system consisting of a memory that stores variable data processing knowledge, an abnormality detection unit that determines whether the B process variable is normal or abnormal from the data processing result of the B process variable, and a memory that stores the B process variable abnormality detection knowledge. Device status understanding processing mechanism.