JPS6159502A - Plant operation controller - Google Patents

Abstract

PURPOSE:To improve the reliability, to avoid an error due to the operator and to reduce the cost by providing a function deciding priority of each running operation and supervision depending on the running state in a rule type control. CONSTITUTION:In the rule type control using directly the knowledge relating to the running operation of the operator, rules relating to the object plant run ning operation and supervision are stored in advance in the rule form to a running operation supervisory knowledge storage section 5. A running operation/ supervisory deciding section 6 references a data of a running state information storage section 4, retrieves a running operation and a supervising rule to be executed in response to the running state, and the retrieved execution enable rule is fed to a dynamic priority management section 7. Thus, the priority of each rule corresponding to the running state is decided, a controller 8 receives the running operating operation command transmitted from the deciding section 6 to execute the corresponding operation. The reliability is improved by provid ing the deciding section 6, and the operator error is avoided and the running cost is reduced.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to an operation control device, and in particular, the present invention relates to an operation control device, and in particular, the knowledge regarding operation that an operator has. The present invention relates to a plant operation control device suitable for rule-based control that is used directly.

[Background of the invention]

3 CD (3tation(::o
(ordinator) is the best known. S.C.
D is especially applied in the field of FA (factory automation), and realizes a flexible control system that directly uses the know-how of skilled tools as a factory production management system. The operational rule expressed as "1-IF'-THEN rule" developed in the field of knowledge engineering is described as "1-IF'-THEN rule". Moreover, its basic execution mechanism takes the form of searching all rules having an IP part that matches data regarding the current state of the nostem, and executing the THEN part of the extracted rule. When implementing plant operation control that directly uses the knowledge of operating personnel possessed by plant operators, the SCD method can basically be applied.

However, in normal plant operation control, there are multiple items that must be operated or monitored at the same time, including operations in response to abnormalities, and among these, the highest priority operation or monitoring item must be selected depending on the operating status. There is a need.

At this time, it is necessary to add a priority to the knowledge regarding each operation and monitoring in advance, and at the same time, to modify the priority given in advance in accordance with the operating state. It is difficult to meet the above requirements with conventional rule-based control systems such as SCD.

[Purpose of the invention]

The purpose of the present invention is to provide a plant that has a function of determining the priority of each operation and monitoring according to the plant operation status in rule-based control that is implemented by directly using the knowledge of operation that all operators have. The purpose of the present invention is to provide an operation control device.

[Summary of the invention]

In rule-based control, where knowledge about driving operations is expressed as rules and rules are selected and executed according to the driving condition, when there are multiple rules to be executed at the same time, conflict resolution is traditionally used to decide which rule to execute first. It is said to be a problem. As a method to solve the problem of conflict resolution, a method is proposed in which a priority is added to each rule in advance, and knowledge (meta-knowledge) about how to apply each rule is also expressed as a rule, and when a conflict occurs, the meta-knowledge is expressed as a rule. A method is being considered in which the execution order is completely determined using In the former, the priority of each rule is fixed, making it difficult to set the priority according to the operating state of the plant. The latter method is effective for conflict resolution, but the overhead caused by the use of meta-knowledge makes it difficult to apply to rule selection that requires real-time performance, such as plant operation control. Therefore, we add a priority to each rule in advance, and at the same time embed the driving state dependence of the priority, which is treated as meta-knowledge, into the rule expression in advance, and when selecting a rule, all driving state dependencies are processed and the priority of each rule is corrected. By providing a function to do this, it becomes possible to realize a conflict resolution mechanism suitable for real-time control.

[Embodiments of the invention]

An embodiment of the present invention will be explained below with reference to FIG. 1st
The figure shows the overall configuration of a plant operation control device in an embodiment of the present invention. In this embodiment, each means is realized by one computer 2, except for a process data importing device 1 that periodically imports plant process data, and a control device 8 that actually operates the plant in response to operation commands. There is. In the computer 2, 3 is a process monitoring section, 4 is an operating state information storage section, 5 is an operating/monitoring knowledge storage section, 6 is an operating/monitoring determining section, and 7 is a dynamic priority management section.

The process monitoring unit 3 stores periodically captured process data, and at the same time monitors preset process data and converts it into a character string. Furthermore, the converted process status is stored in the operating status information storage section 4! Get into it. The operating state information storage section 4 also stores the process state obtained from the process monitoring section 3 and the operating operation and monitoring results issued by the operating operation/monitoring determining section 6. The operation/monitoring knowledge storage unit 5 is a part that stores rules regarding the operation and monitoring of the target plant in a rule format to be described later.
Each rule is stored in advance. The operation/monitoring decision unit 6 is a core part of the plant operation control device.
Referring to the data in the driving state information section 4, the driving operation/monitoring knowledge storage section 4 is searched for driving operations and monitoring rules to be executed in accordance with the driving state. The searched executable rules are sent to the dynamic priority management section 7, and the priority of each rule corresponding to the operating state is determined. The prioritized executable rules are stored in the driving/monitoring determining section 6 as an agenda list (described later). When determining driving operations and monitoring, select the rule with the highest priority in the adjuster list, and select the operating section of that rule. Execute. The determined driving operation command is sent to the control device 8. Also,
For monitoring items, process data stored in the process monitoring unit 3 is referred to. Accordingly, the determined driving operations and monitoring results are stored in the driving state information storage section 4 according to the rules. The dynamic priority management unit 7 determines the priority of the executable rule sent from the driving operation/monitoring determination unit 6 by referring to the driving state information storage unit 4, and applies the determined priority and rule to the executable rule. It is stored in the ajendar list of the driving operation/monitoring determining unit 6. The control device 8 receives the driving operation command sent from the driving operation/monitoring determining section 6 and executes the corresponding operation. In this embodiment, in the operations of the above-mentioned parts, the processes of the process monitoring part 3 and the operation/monitoring determining part 6 are performed by time-sharing processing.

The operation of the entire groove bottom and each part in this embodiment has been described above, and below, the rule expression format and detailed operation of the apparatus in this embodiment will be explained using a specific example.

As a specific example, we will take up reactor start-up operation at a nuclear power plant, especially close-to-critical operation due to control rod withdrawal from a cold shutdown state. Nuclear reactor startup operations are currently being carried out by several operators following operating procedures.

FIG. 2 shows the expression format of the rules used in this example.

As shown in fAK, each rule fi basically consists of five parts. The part 21 describes a symbol used as an index of the rule.

The part 22 is a condition part that describes the conditions under which the rule is driven, and describes the operating state under which the rule is driven.

The conditions are expressed in a list format after the symbol 1nvoke. When all the operating states described in the condition part are satisfied, the rule becomes an executable rule. Is part 25 a priority given in advance considering the importance of the rule? This is the part to be described, and a trap value is assigned after the symbol priority. 2
The part 3 is a situation part that describes the operating state that will be a factor in completely modifying the priority given in advance the next time the electrical connection rule becomes executable. In the status section, an appropriate function for modifying the priority is written together with the operating state.

The section 24 is a rule operation section. The operation section includes an 11 section 26 that describes preprocessing for operation or monitoring, and if
Describe the operation or monitoring results to be implemented based on the results of the test. A configuration formula is formed from the then part 27 and the else part 28. All rules regarding driving operations and monitoring are stored in the rule index section 21, the cutting section 22, and the operation section th.
It has an en section 24 and a priority section 25. Other parts are added and expressed as appropriate depending on the content of the rule. If multiple condition judgment is required in the operation section, use the if section and then section else-
An if section, then section, and else section may also be provided.

Note that a plurality of variables having a common meaning can be written in one rule.

Figure 3 shows a rule representation of part of the knowledge regarding operation and monitoring of a nuclear reactor from cold shutdown to near criticality. In the figure, CMI is an example of a rule regarding control rod withdrawal operation, 0
M2 to CM4 are examples of rules regarding monitoring. In the figure,
CMl's SRM stands for neutron source region neutron detector. Moreover, X in each rule indicates that it is a variable. In the figure, each symbol starting with a tatami is used to indicate a part that changes in the operating state (hereinafter referred to as an operation event and a process event). As a result, for example, if an operation event (control rod withdrawal force end) occurs and an operation event (control rod withdrawal start) occurs, the event (control rod withdrawal end) occurs. H
-Can be automatically deleted. CM2~CM4
The situation part is described, and +1 in the situation part. +2 indicates a function name that increases the corresponding priority by 1 and 2, respectively. In addition, in each rule, for the list that starts with the symbol "observation", the function that refers to the data shown in that order in the process monitoring part in Figure 1, and for the list that starts with the symbol that operates, all the specified operations are executed. A separate function is defined to do this. Note that the symbols, etc. in 0M3 must match all character strings.

The rules regarding driving operation and monitoring shown in FIG. 3 are stored in the driving operation/monitoring knowledge storage section S in FIG.
FIG. 4 shows an example of storing rules.

In this example, the rule index is an atom (one of Lisp data structures), and the condition part, situation part, operation part,
and 1nvoke and priority part respectively, with.

rule, and p-orityi attribute name. In addition, for the 1nvoke attribute, it is shown that an event corresponding to each condition has occurred (
This is indicated by adding the symbol ``active''.

Hereinafter, the detailed operations of the driving operation/monitoring determination &(16 and dynamic priority management section 7 in FIG. This will be explained using an example.

FIG. 5 shows a flowchart of the driving operation monitoring determination process in this embodiment. At 41 in the figure, it is checked whether or not there is a change in the event stored in the driving state information storage section 4 in FIG. Typically, driving operations are event-driven processes driven by the occurrence of some event.

Therefore, when determining driving operations and monitoring items, it is only necessary to focus on newly occurring events. Operation status information includes process events and operation events.
Each event is stored separately into those that have already occurred (called activated) and those that have newly occurred. FIG. 6 shows operating state information in this embodiment. This example shows a state in which the control rod withdrawal operation has been completed at a position where the reactor is not yet in a near-critical state, and an operation event (control rod withdrawal complete) has occurred.

In step 42, driving operation/monitoring rules related to the newly occurring event obtained in step 41 are found.

This process is easily realized by using a discrimination net for the condition part of each rule (Anzai et al., Learning Cognitive Psychology with ISF 2 - Problem Solving, University of Tokyo Press, 1982). That is, For the operation events and process events described in the condition part of all rules, each symbol expressing them and their order are distinguished and expressed in a tree structure, and the rule with the corresponding event at the end of the tree is The index and the digit 1!'&- in its condition part are stored.By passing a newly generated event through the discrimination net, the rule index in which the event is described can be retrieved.In this example, the discrimination net From the rule index and position information obtained using the Internet, the above-mentioned symbol (active) is added to the condition part of the corresponding rule in the driving operation/monitoring knowledge storage unit 5. (Fourth
(See figure) At this time, if a variable exists in the rule condition part, a variable combination list (a list combining variables and symbols to be assigned to them) is also added.

In step 43, executable rules are found by checking the condition part of each rule obtained in step 42.

As mentioned above, whether or not a certain rule is executable can be easily checked by simply checking whether the symbol (active) is added to all the events in the condition part.

The process of 44 is a feature fc of the present invention together with the above-described rule expression format. Usually 43 provides a plurality of executable rules. The operator determines and implements the driving operation or monitoring item with the highest priority based on the operating condition. In the present invention, the dynamic priority management unit 7 in FIG. 1 carries out all of this processing. That is, for each executable rule obtained in step 43, the entire status section and priority section are output.

For rules that do not have a situation part, the priority given to the priority part becomes the i priority, but if a situation part exists, the event stored in the driving state information storage part 4 is By referring to it, it is determined whether the priority given to the rule needs to be modified. If the event described in the status section has occurred, the priority is modified by applying the priority modification function described with the event. The above process is performed for all executable rules, and in step 45, an ajenda list is created which is stored in the form of a rule index and its priority gold list. FIG. 7 shows the ajenda list in this actual M example. In the above example, the reactor was not yet in a near-critical state, but if it was in a near-critical state, an event (operation mode number near critical) would have occurred as a process event, and what would the result be? The list is created as shown in FIG. That is, the priority of driving operations and monitoring rules corresponding to the driving state is determined.

At step 46, the rule having the highest priority is selected from the ajendar list, and at step 47, the operation section of the rule is executed. In the case of driving operation, the execution in 47 refers to
This means sending an operation command to the control device, and during monitoring, means referring to the data of the process monitoring section 2 to obtain a monitoring result.

A new process event occurs due to the execution of a driving operation, and the operation event also changes. Therefore, the process described above is repeated. Note that if no new event occurs at 41, the process moves to 46 and selects and executes the rule with the next priority.

〔Effect of the invention〕

As explained above, according to the present invention, in various plant operations, operation operation/monitoring decisions according to operating conditions can be applied to rule-based control that directly uses knowledge related to operation operations held by operators. It is possible to maintain the same functions and realize highly reliable automatic driving equipment. As a result, human errors in plant operation can be avoided and operating costs can be reduced.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the plant operation control device configuration according to the present invention, FIG. 2 is an explanatory diagram showing an example of rule expression, FIG. 3 is an explanatory diagram of an example of a rule, and FIG. 4 is a rule storage diagram. Fig. 5 is a flowchart showing the processing process, Fig. 6 is an explanatory drawing of operating state information, and Figs. 7 and 8 are explanations of the Aziendar list
C: Talk. 1... Process data import device, 2... Total n1a
, 3... Process monitoring section, 4... Operating state information storage section, 6... Operating operation/monitoring determining section, 8... Control device.

Claims (1)

[Claims] 1. A means for importing process data, a means for converting the imported process data into a character string, and plant operating status information consisting of the process data converted into a character string, operational instructions given as a character string, and plant monitoring results. means for storing knowledge expressed in rules regarding operation and monitoring; and means for determining an operation command or a monitoring command by comparing and collating the plant operating status information and the knowledge expressed in rules. and, in a plant operation control device having means for executing operation and monitoring based on the operation command and monitoring command, a condition part that describes the conditions to which the knowledge is applied, the knowledge regarding the operation and monitoring, and the knowledge. A rule consisting of a situation part that describes plant operating status information related to priority determination when applying, an operation part that describes operation operations or monitoring contents, and a priority part that describes the priority of the relevant knowledge. A plant operation control device characterized by having means for determining a priority based on information in the situation part and priority part of the rule and plant operation status information when determining an operation command and a monitoring command.