JPH01103712A - Operation backup device for process plant - Google Patents

Abstract

PURPOSE:To quickly recognize the state of a process plant in a wide range to quickly code with the abnormality of the plant by using the knowledge engineering to diagnose the abnormality of the plant and make and synthesize countermeasure plans. CONSTITUTION:A physical process or a chemical process are used to store plant data from the process plant, which produces materials of energy, in a plant data base 3. An inference mechanism 5 reads in plant data and a knowledge for abnormality diagnosis of the process plant and making and synthesis of countermeasure plans from the plant data base 3 and a knowledge base 4. The inference mechanism 5 refers to plant data to make inference for abnormality diagnosis and making and synthesis of countermeasure plans based on the knowledge for abnormality diagnosis of the process plant and making and synthesis of countermeasure plans. The result of this inference is reported to an operator by a man-machine interface.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Application Field) The present invention identifies the causes of abnormalities based on the knowledge engineering concept from data on disturbances in process plants such as power plants and oil refinery plants. The present invention relates to an operation support system for a process plant that diagnoses the state of a plant, selects a new operation target based on the diagnosis results, and forms and synthesizes a countermeasure plan necessary to achieve the new operation target.

(Prior art) Generally, when a process disturbance occurs during the operation of a process plant due to a failure or malfunction of equipment, operator error, external disturbance, etc., if the disturbance is left unattended, substances will build up locally in the plant. Creates balance and concentration of energy,
There is a risk of damage to the plant and release of hazardous substances into the environment.

Therefore, it is necessary to detect the disturbance at an early stage, diagnose the state of the plant, and take appropriate measures. Therefore, in conventional process plants, necessary information about the plant is generally collected in a central control room and abnormalities are detected and diagnosed by monitoring this information.

In other words, in conventional abnormality detection diagnosis, a normal operating range is determined in advance for measured values from the plant, and an alarm device is installed to issue a warning to operators when the measured value deviates from this range. The entire plant can be broadly monitored. Furthermore, when an alarm occurs, abnormality diagnosis and countermeasures for the abnormality are handled by referring to a procedure manual regarding the alarm, or by educating operators through operational training in advance.

(Problem to be solved by the invention) However, as process plants generally become larger and more complex, the scale of these alarm devices also increases.
The contents of procedure manuals and training are also becoming more complex. Therefore, in the event of a disturbance, a number of alarms may be generated that far exceeds the operator's processing ability, and in such a case, it is difficult to appropriately process the alarms.

Therefore, we developed a system that uses computer information processing technology to reduce the number of alarms and provide only truly valuable and appropriate information to operators, and an abnormality support system that provides operators with operations to respond to abnormalities. has been done. Such alarm suppression systems include, for example, a system that selects the earliest alarm from a large number of alarms, and a system that uses relatively simple logic, such as a system that directly links to the cause when multiple alarms occur due to the same cause. A system has been proposed that suppresses unnecessary alarms by logic such as excluding one alarm and erasing other alarms. Although these alarm suppression systems are somewhat effective, they are generally partial responses and cannot be considered as a fundamental countermeasure.

On the other hand, some abnormality support systems predetermine response operations for anticipated abnormalities and provide them to operators as operating procedures when an abnormality occurs; has been pointed out as a problem. Other systems have been proposed that provide response plans that are not based on abnormality diagnosis results, but although these systems are effective from the perspective of ensuring plant safety, they do not necessarily result in optimal response operations. This is natural because the optimal operating procedure cannot be determined unless the cause of the abnormality and the event are clear.

Incidentally, in recent years, an abnormality diagnosis/driving operation guide system based on a new concept using knowledge engineering methods has been developed. This abnormality diagnosis/operation guide system includes a knowledge base containing knowledge about the causes of abnormalities and abnormal changes that appear in plant measurement signals as a result, as well as knowledge that determines operational procedures to respond to abnormalities, and this knowledge. An inference mechanism is provided that reads the above knowledge from the base and infers the cause of the abnormality and corresponding operation procedures.

In this abnormality diagnosis/operation guide system, for example, when an abnormal pattern is obtained from plant measurement values, the inference mechanism searches the knowledge base and determines whether there is knowledge that matches that pattern. Pick up what you are looking for and what meets your needs. If there is a plurality of pieces of matched knowledge, the cause of the abnormality and the corresponding operation procedure are estimated by selecting the most probable one from the plurality of pieces of knowledge.

However, in the above-mentioned abnormality diagnosis/driving operation guide system, although it is possible to diagnose abnormal causes and abnormal patterns that have been considered in advance and stored in the knowledge base,
It is not possible to diagnose or derive a corresponding operation procedure for abnormal causes and abnormal patterns that are not anticipated in advance. In addition, individual knowledge regarding the causes of abnormalities, abnormal patterns, and response operation procedures are essentially the same and difficult to structure, so the number of knowledge items increases dramatically as the system becomes larger. . Therefore, there is a problem in that the processing time becomes enormous and it becomes difficult to quickly respond to abnormalities.

The present invention has been made in consideration of the above circumstances, and is capable of diagnosing abnormalities in process plants and formulating and synthesizing countermeasure plans.
By using knowledge engineering methods, we can broadly and quickly grasp the status of the plant and respond quickly to abnormalities in the plant by systematically carrying out this from the viewpoint of production and ensuring safety, which is the original purpose of the plant. The purpose is to provide an operation support device for process plants that is capable of

Another object of the present invention is to provide an operation support system for a process plant that is capable of diagnosing unknown abnormality causes and abnormality patterns that have not been anticipated in advance and dealing with the abnormalities.

[Structure of the invention]

(Means for Solving the Problems) The process plant operation support device according to the present invention has a plant database that stores plant data from process plants that produce substances and energy using physical processes and chemical processes. , a knowledge base that has the knowledge for diagnosing abnormalities in the process plant and formulating and synthesizing countermeasure plans, and plant data and data necessary for diagnosing abnormalities in the process plant and formulating and synthesizing countermeasure plans from the above plant database and knowledge base. It is equipped with an inference mechanism that reads knowledge and performs inferences for diagnosing abnormalities and formulating and synthesizing countermeasure plans, and a man-machine interface that displays the results of inferences made by this inference mechanism.

(Operation) Plant data from process plants that produce substances and energy using physical and chemical processes is stored in a plant database. Further, the knowledge base is provided with knowledge for diagnosing abnormalities in the process plant and formulating and synthesizing countermeasure plans.

The inference mechanism first reads plant data and knowledge for diagnosing abnormalities in the process plant and formulating and synthesizing countermeasure plans from the plant database and knowledge base. Then, the inference mechanism makes inferences for diagnosing abnormalities and formulating and synthesizing countermeasure plans based on knowledge for diagnosing abnormalities in the process plant and drafting and synthesizing countermeasure plans while referring to the plant data.

The results of the inference by the inference mechanism are then provided to the operator through the man-machine interface.

(Embodiment) An embodiment of the operation support device for a process plant according to the present invention will be described with reference to the drawings.

The present invention is suitably used as an operation support device 2 for a nuclear power plant 1, and as shown in FIG. A knowledge base 4 that has knowledge for formulating and synthesizing plans, and the plant data and knowledge necessary for diagnosing abnormalities in the nuclear power plant 1 and formulating and synthesizing countermeasure plans are read from the plant database 3 and knowledge base 4. It is provided with an inference mechanism 5 that performs inference for diagnosing abnormalities and formulating and synthesizing countermeasure plans, and a man-machine interface 6 that displays the results of inference by this inference mechanism 5.

The plant database 3 includes the nuclear power plant 1
Information regarding the operating status of the plant, such as process measurement values from the plant, is input and stored as plant data. ' In addition, the knowledge base 4 includes knowledge describing functional hierarchical structures of multiple nuclear power plants 1 corresponding to multiple operating states, as knowledge for diagnosing abnormalities in the nuclear power plants 1 and formulating/synthesizing countermeasure plans. (hereinafter referred to as hierarchical structure knowledge) and knowledge for diagnosing abnormalities for each node in the functional hierarchical structure (hereinafter referred to as node diagnosis knowledge)
knowledge to perform abnormality diagnosis for the entire plant (hereinafter referred to as "overall diagnosis knowledge"); knowledge to select new operational targets based on the abnormality diagnosis results (hereinafter referred to as operational target selection knowledge); knowledge to formulate countermeasure plans necessary to achieve operational goals (hereinafter referred to as planning knowledge), and knowledge to synthesize this countermeasure plan into a consistent overall plan for the entire plant (hereinafter referred to as planning synthesis knowledge). ).

The contents of each knowledge will be explained in detail below.

Hierarchical structure knowledge is a knowledge that understands the nuclear power plant 1 as a hierarchical structure in terms of its functions, based on a knowledge engineering approach.

In other words, an artificial object such as a process plant is generally constructed with a clear purpose/function, and its structure includes a system for achieving the final goal of the process plant, and various subgoals to achieve the final goal. It is thought that a hierarchical structure is formed, with subsystems having subsystems, and subsystems related to those subsystems.

Further, it is considered that the above-mentioned subsystems are equipped with auxiliary systems for configuring those subsystems and maintaining and exhibiting their functions.

Such a hierarchical structure can be clearly recognized especially when a process plant is observed from a functional perspective, since physical structures generally bridge many functions.

In this example, in creating a knowledge base for the functional hierarchical structure of the nuclear power plant 1, we will focus on the movement of materials within the plant, which is the basis of a process plant, and the movement of energy using material movement as a medium. , creating a clear functional hierarchy of knowledge.

Hierarchical knowledge about the nuclear power plant 1 is created with respect to two basic purposes of the nuclear power plant 1: producing power and maintaining safety. For example, a functional hierarchical structure for electricity production is created as shown in Figure 2, and a functional hierarchical structure for safety maintenance is created for at least three types of purposes as shown in Figures 3 to 5. Ru.

In Figures 2 to 5, 70 functions such as storage, transportation, branching, barrier, etc.
on) is expressed using the symbols shown in FIG.

First, the functional hierarchical structure for power production shown in Figure 2 is created for the purpose of the plant to normally produce power, and is made up of multiple nodes, and these multiple nodes are connected to the RAMS structure. lry.

be done. A node in this hierarchical structure is called a flow structure and has a specific function.

The power production flow structure 11 at the top has a function for achieving power production and is composed of a plurality of flow functions. That is, nuclear power plant 1
is a system that generates electricity using nuclear fission reactions, so as a flow function to achieve that purpose, source 1 that generates energy through nuclear fission reactions is used as a flow function.
2. A system 13 that converts and transports the energy, and a user 14 that consumes the transported energy are provided.

The power production management flow structure 15 at the bottom of the power production flow structure 11 includes a nuclear energy generation system 16 as a subsystem of the source 12 that generates energy through the nuclear fission reaction, and a system that converts and transports the energy. Thermal energy transport system 17 and energy conversion system 18 as subsystems of
, electrical energy conversion system 19, waste heat transport system 2
0 is provided, and furthermore, the power grid 2 as said user 14
1. An environment 22 is provided.

Each system 16 to 20 also has support 2 as an auxiliary system necessary to maintain its function.
3 to 27 are provided. These supports 23 to 27 are provided with a 70-structure at the bottom of the grooves,
The systems within these 70-structures maintain the functionality of each support 23-27.

For example, flow structures that maintain the functions of the nuclear energy generation system support 23 include a control rod drive flow structure 30 for controlling nuclear fission reactions, and a core recirculation flow structure for circulating cooling water into the reactor. 31, a feed water heating flow structure 32 for heating feed water into the reactor, and a pressure control flow structure 33 for controlling the pressure of the reactor.

A subsystem and a support (not shown) are provided inside each of these flow structures 30 to 33, and a flow structure that maintains the function is provided below each support.

On the other hand, among the functional hierarchical structures for maintaining safety, the functional hierarchical structure for preventing radioactive release shown in Figure 3 is created for the purpose of preventing nuclear fuel radioactivity from leaking into the surrounding environment. A radiation release prevention flow structure 35 is provided as the top node.

This radioactive release prevention flow structure 35 includes a radioactive source 36 as a flow function, and this radioactive source 3
It consists of a radioactivity barrier 37 that shields radioactivity from 6 and an environment 38 around the radioactivity barrier 37.

The radioactive release prevention flow structure 40 as a lower node includes a radioactive source 41 as a flow function, a fuel cladding tube 42, a secondary coolant boundary 43, a relief safety valve 44, a reactor containment vessel 45, an emergency gas treatment System 46, rare gas hold up system 47
and environment 48 are provided. Furthermore, each flow function 41-47 is provided with supports 50-57 as auxiliary systems. Each of these supports 50~
57 is also provided with a flow structure (not shown) below it.

The functional hierarchical structure for maintaining safety functions shown in Figure 4 is as follows:
It is created for the purpose of suppressing heat generation from the core and removing generated heat during scram etc., and a safety function maintenance flow structure 60 is provided as the top node. This safety function maintenance flow structure 60 includes a heat source 61 as a heat source, and a heat source 61 as a heat source.
A heat removal system 62 for removing heat from the heat exchanger 1 and a heat sink 63 for discharging the removed heat are provided.

The core cooling function flow structure 64 at the bottom of this safety function maintenance flow structure 60 includes a core thermal output 65
, a secondary coolant 66, a condenser system 67, a containment vessel 68, a heat sink 69, and other flow functions are provided, and each of the flow functions 65 to 68 is provided with supports 70 to 73 as auxiliary systems, respectively. Also,
A control rod drive flow structure 74, a boric acid water injection function flow structure 75, and many other 70-structures (not shown) are provided below these supports 70 to 73.

Furthermore, the functional hierarchical structure for maintaining containment vessel integrity shown in Figure 5 is created with the aim of maintaining the integrity of the reactor containment vessel and reactor building, which serve as the final barrier in the event of a core meltdown, etc. It is something. The containment vessel integrity maintenance flow structure 77 as the top node includes a radioactive source 78, a radioactive barrier 79 for shielding radioactivity from the radioactive source 78, and a radioactive barrier 79 for shielding the radioactive source 78.
9 is provided as a flow function.

The containment vessel isolation function flow structure 81 as a lower node includes flows such as a radioactive source 82, a reactor containment vessel 83, a reactor containment vessel isolation 84, a reactor building 85, a reactor building isolation 86, and an environment 87. No action is provided. Each of the flow functions 83 to 86 is provided with supports 88 to 91 as respective auxiliary systems in the same way as other hierarchical structures, and a flow structure (not shown) is further provided below each of the supports 88 to 91.

Furthermore, each of the 70-structures in each of the functional hierarchical structures described above includes basic countermeasures against abnormalities in each flow function and support as knowledge data.

FIG. 7 and FIG. 8 are diagrams showing an example of the above-mentioned basic measures.

FIG. 7 shows an example of the basic measures provided in the feed water heating flow structure 32, and describes countermeasures and measures to be taken when condensate leaks from the feed water tubes etc. in the feed water heater. The purpose of the countermeasure in this case is to prevent leakage of primary cooling water. When condensate leaks from a feedwater heater, possible countermeasures to achieve the objective include reducing the reactor output and then isolating the feedwater heater system where the leakage occurred.

FIG. 8 shows an example of basic countermeasures provided in the power production management flow structure 15, and describes the purpose and countermeasures to be taken when the feed water temperature decreases and the reactor output increases. In this case, the objective of the countermeasure is to maintain core integrity. For example, if one system of the feedwater heater is isolated, the core inlet subcooling will increase due to the decrease in the feedwater temperature, resulting in an increase in reactor power, and the limit value for maintaining the health of the reactor core will be monitored. The nuclear energy generation system support 23 that is currently running becomes abnormal. One possible countermeasure to this abnormality is to lower the reactor output by lowering the core recirculation flow rate.

In this case, since it is necessary that the lower core recirculation flow structure 31 having the functions necessary for implementing the countermeasure is normal, it is stated that the recirculation control system is normal as a constraint.

Note that the functional hierarchical structures shown in FIGS. 2 to 5 are roughly expressed for the purpose of explanation, and in reality, a more detailed functional hierarchical structure is created.

Further, if the type of nuclear power plant 1 is different, a functional hierarchical structure according to the type is created.

The node diagnostic knowledge describes a method for diagnosing an abnormality in each node constituting the functional hierarchical structure.

This node diagnosis knowledge includes knowledge to judge the health of each node (hereinafter referred to as health judgment knowledge), and abnormalities in nodes judged to be unhealthy based on this health judgment knowledge, taking into account the structure related to material or energy balance. It is divided into knowledge for performing diagnosis (referred to as g, lower node abnormality diagnosis knowledge). The above-mentioned soundness judgment knowledge is based on the knowledge engineering concept, and is explained below.

That is, an abnormality (failure) that occurs in a physical structure such as a process plant generally affects some of the functions performed by that structure, and the effect spreads from the bottom to the top of the hierarchical structure. In other words, the propagation of anomalies is bottom-up (
(Bottom up) process.

On the other hand, when diagnosing such an anomaly, humans usually check from the top of the hierarchy, and as they catch up on the anomaly, they gradually move down the hierarchy and narrow the scope of the search, using a top-down approach. .

This way of thinking is considered natural since the greatest concern for humans is whether or not a process plant that is operated with a predetermined purpose always achieves its final purpose. Therefore, in the present invention, similar to the way humans think, the hierarchical structure is searched top-down to diagnose abnormalities.

Specifically, this abnormality diagnosis is performed as follows. First, for each flow function and support of each flow structure of the hierarchical structure knowledge, parameters such as generation amount and storage amount are specified in advance for determining the soundness of its performance.

This parameter is plant data from the nuclear power plant 1 or a calculated value calculated based on the plant data.

Then, at any given time, while referring to the plant data, starting from the top of the hierarchical structure and traversing the hierarchy in order, the performance health of each flow function and support is sequentially determined based on the specified parameters. , to obtain pass/fail judgment results for each flow function and support.

Next, the node abnormality diagnosis knowledge is knowledge for performing an abnormality diagnosis on a flow structure including a flow function or support whose health is determined to be abnormal based on the judgment result of the health judgment knowledge, and will be described next. It looks like this.

First, abnormal states are classified according to the type of flow function as shown in FIG. 9.

That is, the above-mentioned health judgment is performed based on, for example, the source (5our
ce) is determined based on performance parameters regarding the amount generated or stored.

In this case, the amount of generation is determined by performance parameters in units of (Ks /sec) and (cal/5ec), and the amount of storage is determined by (Ky) and (Cal/5ec).
It is determined based on a performance parameter in units of Kcal). The causes of the abnormality diagnosed for each flow function from the determination result are abnormal generation amount, abnormal amount of heat generation, abnormal rate of change due to abnormal transport, and abnormal amount of change.

Furthermore, the health of transport is determined based on performance parameters regarding flow rate. Regarding the flow rate, (K9 /sec) and (
The cause of the abnormality as a diagnosis result is a transportation abnormality or a branching abnormality. In addition, storage (Stor-age)
), distribution, sink, and barrier, and the cause of the abnormality is diagnosed based on the classification.

In this case, it is possible that the storage amount is due to an abnormal rate of change due to an abnormality in transport, but this abnormality is diagnosed not as an abnormality in the flow function itself, but as a cause of the abnormality in the transport or barrier connected to the flow function. This is the case. At the time of this diagnosis, each 70
- Knowledge of flow functions and flow function connections within the structure is utilized.

Next, based on FIG. 9, when the results of diagnosing each flow function are summarized for an arbitrary flow function, it is classified into one of the cases shown in FIG. 10. FIG. 10 shows the results of diagnosing a specific flow function, support for that flow function, and flow functions adjacent to that flow function, and also describes the judgment results for the diagnosis results. In FIG. 5, normal cases are indicated by O, and abnormal cases are indicated by ×.

Further, Δ indicates an abnormality in the performance parameters, but it is determined that the abnormality is due to the surrounding flow function.

The abnormal cases shown in FIG. 10 are classified into 10 cases. For example, case 1 indicates a case in which a specific flow function, its support, and adjacent flow functions are all normal, and the determination thereof is also normal. Case 2 shows a case where there is no abnormality in a specific flow function or adjacent flow function, but there is an abnormality in the support, and the judgment result is [Potential failure, an abnormality in the corresponding flow function will appear sooner or later due to an abnormality in the support] "in·
be.

Cases up to case 10 are classified in the same manner below.

Note that Case 5.6 is a judgment result based on the assumption that there is movement of matter, but if there is almost no movement of material and there is a problem with accuracy, a judgment like Case 5.6 should not be performed. Good too.

In this way, by determining and classifying the abnormality cases for each flow function of a particular flow structure, one of the following judgment results can be obtained regarding the source of the abnormality of the flow structure.

1. There is an abnormality in a specific flow function, but there is no abnormality in its support.

2. There is an abnormality in a specific flow function, and there is also an abnormality in its support.

3. There is an abnormality in a specific support, but there is no abnormality in the corresponding flow function.

4. Multiple abnormalities W consisting of any combination of the above three
A (for multiple flow functions).

From this determination result, a diagnosis result regarding the propagation range of the abnormality within the flow structure and the performance evaluation of the flow structure can be obtained.

The overall diagnosis knowledge is knowledge that describes a method for diagnosing the entire plant from the diagnosis results of each node based on the node diagnosis knowledge, and is as described below.

When abnormal flow structures are sequentially diagnosed from the top of the plant function hierarchical structure according to the node diagnosis knowledge, it is possible to determine in which nodes in the hierarchical structure the abnormality exists, how these abnormalities are connected, and how they are scattered. It becomes clear what to do.

If there is a connection between the node abnormalities and the connection matches the support relationship between the 70 structures in the hierarchical structure, it is confirmed that the abnormality at the bottom has clearly propagated based on the relationship between the support functions. can. If the match is incomplete, the magnitude of the anomaly is still small, and it is expected that the support relationship has not become apparent. Even in this case, it becomes clear which function in the hierarchical structure is found to be abnormal. A similar judgment is made in the case where abnormalities are scattered.

The driving target selection knowledge is knowledge that describes a method for selecting a new driving target based on the result of the abnormality diagnosis described above and selecting a functional hierarchical structure corresponding to the new driving state.

In general, plant operators detect anomalies, search for causes, and evaluate effects based on the functional hierarchy based on the pattern of process values that serve as a norm for a certain operating state.
Based on the evaluation results, a new driving goal is selected, and corresponding operations are performed to achieve the new driving goal, while at the same time, abnormality monitoring is continued in accordance with the new driving state.

In other words, the ultimate goal of operating a process plant is to continue production and maintain safety during normal operation, but when an abnormality occurs in the plant, ensuring safety may be prioritized even at the cost of some production. be. For example, in a nuclear power plant, the operational goals during normal operation are to produce electricity and ensure safety, but if some abnormality occurs and it becomes difficult to continue operation at the rated output, it is necessary to disconnect the abnormal part or reduce the output. A driving target is selected with emphasis on ensuring safety so as to shift to a safer state, such as switching, and abnormality monitoring is performed in accordance with the driving state.

The present invention incorporates the thinking process of plant operators, selects new operational goals based on abnormality diagnosis results, and creates a functional hierarchical structure corresponding to a new operating state from among a large number of functional hierarchical structures. is selected and anomaly diagnosis is performed based on its functional hierarchical structure.

Therefore, as shown in Figure 11, the operational goal selection knowledge classifies events that occur in the plant based on abnormality diagnosis results, and describes the operational goals after the occurrence of the event and the functional hierarchical structure corresponding to the operating state. You will be equipped with the knowledge that you have learned.

Events that occur in a plant are classified into equipment failures with backup equipment, equipment failures without backup equipment, scram events, etc. based on the abnormality diagnosis results. For example, during normal operation, abnormality diagnosis is performed using the functional hierarchical structure for power production shown in Figure 2 and the functional hierarchical structure for radioactive release prevention shown in Figure 3. In the event of a minor radioactivity leak, we select the operational goal of disconnecting the faulty equipment and continuing steady operation while maintaining the integrity of the radioactivity barrier, and establish a functional hierarchy for power production and radioactivity release prevention. Continue to use the structure for abnormality diagnosis.

In addition, in the event of equipment failure for which there is no spare equipment or leakage of safety relief valves, such as an abnormality in the core recirculation flow structure 31 shown in Figure 2, the operational goals of reducing output and maintaining the integrity of the radioactive barrier will be set. Similarly, the functional hierarchy of power production and radioactivity release prevention is used for abnormality diagnosis. However, when output reduction operation is set as an operation target, some of the set values of parameters used in the soundness judgment knowledge are changed.

Next, when a scram occurs, the operation goal during normal operation of producing electricity disappears, and the operation goal of safe shutdown (wet shutdown/cold shutdown) and maintaining the integrity of the radioactive barrier is selected. The fourth mode corresponds to the new operating condition.
Abnormality diagnosis is performed using the functional hierarchical structure for maintaining safety functions and the functional hierarchical structure for preventing radioactive release shown in the figure. In this case, ensuring safety is given top priority, and abnormality diagnosis regarding power production is not performed.

Furthermore, in the event that the reactor core is damaged or melts due to the failure of the emergency core cooling system, etc., in order to prevent the release of radioactivity into the environment as much as possible, which is the final operational goal, The operational objectives of ensuring vessel integrity and radioactivity barrier are selected, and an abnormality diagnosis is performed using the functional hierarchical structure for maintaining containment vessel integrity and preventing radioactivity release that corresponds to the new operating state. In this case as well, ensuring safety is given top priority, and abnormality diagnosis regarding power production is not performed.

The knowledge shown in FIG. 11 is an example, and events that occur in the plant may be further classified to create even more detailed operating conditions and operating targets.

In this case, it is necessary to provide even more types of functional hierarchical structures.

The planning knowledge is knowledge describing a method for formulating a countermeasure plan necessary for achieving the new driving goal selected by the driving goal selection knowledge.

In this planning knowledge, it is first determined whether the new functional hierarchical structure is the same as the functional hierarchical structure before the abnormality diagnosis. If the new functional hierarchical structure is the same as the functional hierarchical structure before the abnormality diagnosis, a countermeasure plan is created based on the abnormality diagnosis results that served as the basis for selecting the new functional hierarchical structure. In addition, if the new functional hierarchical structure is different from the functional hierarchical structure before the abnormality diagnosis, the abnormality diagnosis is performed again using the abnormality diagnosis knowledge corresponding to the new functional hierarchical structure, and countermeasures are taken based on the abnormality diagnosis results. Develop a plan.

Next, the planning knowledge derives basic measures prepared in advance in the interlayer structure knowledge for each 70-structure based on the operational objectives, and then determines whether each basic measure is executable based on constraints and plant conditions. Make a judgment.

The derivation of such basic measures is performed sequentially from the top 70-structure to the lower 70-structures of the flow structure in a top-down process. As a result, as shown in FIG. 12, a hierarchical structure is formed in accordance with the functional hierarchical structure, from the abstracted basic measures located at the upper level to the basic measures that are sequentially broken down. In this N-layer structure, the contents of the basic measures become more specific from the upper level to the lower level.

In principle, basic countermeasures are derived from the top down to the lowest 70-structure level, but during this process, there may be cases where the plant condition does not comply with the stoppage provisions or where the countermeasures cannot be implemented due to some other constraints. Sometimes it comes. In such cases, basic measures should be derived only to a practicable level. As a result of such processing, appropriate basic countermeasures are derived for every event, and a hierarchical structure of concrete basic countermeasures that is as practical as possible is drawn up as a countermeasure plan.

The planning synthesis knowledge is knowledge that describes a method for synthesizing a countermeasure plan as a hierarchical structure of basic measures drawn up using the planning knowledge into an overall plan that is consistent with the entire plant.

The hierarchical structure of basic measures created according to the planning knowledge described above may contain basic measures that are inconsistent with each other, so it is necessary to maintain consistency as a whole. Contradictions are resolved based on the objectives of the measures listed in the basic measures for each layer.

In other words, since the upper level in the hierarchical structure of basic measures is an abstracted basic policy, a dependency relationship is set in advance for the countermeasure objectives described in the basic measures between each N layer, and the upper level Delete lower-level basic measures that describe countermeasure objectives that are contrary to the countermeasure objectives.

In this way, after resolving the contradictions in the llll1l structure of basic measures, the basic measures at the lowest level of this hierarchical structure are aggregated for each measure purpose, and the final countermeasure plan as an overall plan is synthesized. In this case, the countermeasure objectives listed in the basic countermeasures located at a higher level than the synthesized countermeasure plan indicate the reasons and effects for implementing the countermeasure plan, and are used to help operators evaluate the provided countermeasure plan. This is useful data that can be used as a reference.

Next, the above knowledge base 4 and the above plant database 3
An inference mechanism 5 connected to the plant database 3 and the knowledge base 4 periodically reads plant data and knowledge for diagnosing abnormalities in the nuclear power plant 1 and formulating and synthesizing a countermeasure plan, and uses the read plant data. Diagnose abnormalities and draw up/synthesize countermeasure plans while referring to them.Inferences are made based on knowledge to diagnose plant abnormalities, select new operational targets, and formulate/synthesize countermeasure plans. .

First, the inference mechanism 5, while referring to the plant data,
Among the hierarchical structure knowledge, the functional hierarchical structure of power production and radiation release prevention is used to determine the health of each node in the functional hierarchical structure at that point in time, according to the health judgment knowledge. For nodes determined to be unhealthy, abnormality diagnosis is performed according to the node abnormality diagnosis knowledge, with reference to knowledge regarding the internal structure of the node and plant data. Thereafter, the inference mechanism 5 integrates the abnormality diagnosis results of each node according to the overall diagnosis knowledge, and performs an abnormality diagnosis for the entire plant.

Thereafter, the inference mechanism 5 selects a new driving target based on the abnormality diagnosis result and in accordance with driving target selection knowledge. and,
The functional hierarchical structure corresponding to the new operating state is read from the knowledge base, and the functional N-layer structure is used to further perform abnormality diagnosis for each node and the entire plant. Further, the inference mechanism 5 formulates a countermeasure plan to achieve a new operational goal according to the planning knowledge, and then synthesizes the drafted countermeasure plans into an overall plan that is consistent with the entire plant according to the plan synthesis knowledge.

The inference mechanism 5 is designed to periodically repeat such abnormality diagnosis, selection of a new driving target, and inference regarding the formulation and synthesis of a countermeasure plan.

A man-machine interface 6 connected to the inference mechanism 5 is equipped with an input device such as a keyboard and a display device such as a CRT, and this display device displays the inference results by the inference mechanism 5.

Examples of displayed inference results include the location of an abnormal node on the plant's functional hierarchy and its connections, the location of abnormal flow functions and abnormal supports in the flow structure corresponding to the abnormal node, and their connections, In other words, there are the causal relationship and the range affected by the abnormality, the newly selected operational target, the functional hierarchical structure name used for the abnormality diagnosis, the countermeasure plan necessary to achieve the new operational target, its purpose, the reason for deriving it, etc.

Next, the operation of the above embodiment will be explained.

When an abnormality occurs in the nuclear power plant 1, the abnormality is input to the plant database 3 as plant data through various detectors, amplifiers, signal transmission lines, etc., and is stored in the plant database 3. Then, the inference mechanism 5 extracts plant data including data about abnormalities from the plant database 3 and knowledge base 4, diagnoses abnormalities in the nuclear power plant 1, and formulates a countermeasure plan.
Load the knowledge for performing synthesis.

Thereafter, the inference mechanism 5 performs a top-down judgment on the health of each flow function and each support from the top of the functional hierarchical structure of power production and radiation release prevention according to the health judgment knowledge while referring to the plant data. This determination result is temporarily stored in a working memory within the plant database, for example.

Then, the inference mechanism 5 diagnoses the abnormality of the flow structure including the flow function or support whose health is determined to be abnormal in accordance with the node abnormality diagnosis knowledge. After diagnosing a node containing an abnormality, the abnormality diagnosis for the entire plant is performed based on the diagnosis result and in accordance with the overall diagnostic knowledge.

After that, the inference mechanism 5 selects a new driving goal and functional hierarchical structure based on the abnormality diagnosis result and in accordance with the driving goal selection knowledge. Then, the inference mechanism 5 determines whether the new functional layer structure is the same as the functional M layer structure before the abnormality diagnosis, and if the same, based on the planning knowledge and the abnormality diagnosis results, a new operational goal is set. A countermeasure plan is created to achieve the goal, and the countermeasure plans are synthesized into an overall plan that is consistent with the entire plant according to the plan synthesis knowledge.

In addition, if the new functional hierarchical structure is different from the functional hierarchical structure before the abnormality diagnosis, perform the abnormality diagnosis again using the new functional hierarchical structure, and then formulate a countermeasure plan based on the abnormality diagnosis results. Perform synthesis. In addition, if the plant status changes and the operational goals change while the countermeasure plan is being drafted and synthesized, and the functional hierarchical structure changes accordingly, the countermeasure plan that is being inferred at that point will be drafted and synthesized. The abnormality may be diagnosed and a countermeasure plan formulated and synthesized so as to give top priority to the new operation target that has been switched.

The inference results by the inference mechanism 5 are displayed on the display device by the man-machine interface 6, and the operator is provided with the abnormality diagnosis results, selected operation targets, countermeasure plans, etc.

In the above embodiment, based on the final goal of the nuclear power plant 1, which is to produce electricity and ensure safety, the hierarchical structure knowledge that hierarchically represents the functional structure of the nuclear power plant 1 is used to diagnose abnormalities and formulate and synthesize countermeasure plans. Therefore, it is possible to perform abnormality diagnosis and formulate and synthesize countermeasure plans in order of importance, broadly and quickly, and without omissions.

In addition, since diagnosis is performed based on typical performance parameters for highly abstracted functions, it is possible to quickly diagnose plant abnormalities and formulate and synthesize countermeasure plans starting with important functions without being confused by trivial information. Can be done.

Furthermore, since the hierarchical structure knowledge is created based on the essential and universal mass and energy balance for the nuclear power plant 1, the abnormality diagnosis results, countermeasure plans, and their objectives/derives can be determined using the hierarchical structure knowledge. The reasons are consistent and easy to understand.

In addition, the system selects operation targets according to the plant status and uses a functional hierarchy structure corresponding to the operating status to diagnose abnormalities and formulate and synthesize countermeasure plans. can be done.

In particular, it is highly effective in that it is not necessary to envisage individual abnormality causes and abnormal patterns in advance as in the past, and it is possible to diagnose unknown abnormality causes and formulate and synthesize countermeasure plans.

In the above embodiment, it has been explained that the driving target is automatically selected according to the driving target selection knowledge, but it is also possible to reflect the intention of the operator via the man-machine interface 6 when selecting the driving target. good. For example, this may be the case when cooling under reduced pressure from a wet stop state to a cold stop state.

Further, in the above embodiment, the abnormality diagnosis device 2 for a nuclear power plant 1 was described, but the present invention is not limited to this, and can be applied to a thermal power plant, a chemical industrial plant, etc. using a physical process or a chemical process. It can be widely applied as an abnormality diagnosis device for process plants that produce substances and energy.

〔Effect of the invention〕

As explained above, the process plant operation support device according to the present invention includes a plant database that stores plant data from a process plant that produces substances and energy using physical processes and chemical processes;
A knowledge base that has the knowledge to diagnose abnormalities in the process plant and formulate and synthesize a countermeasure plan, and plant data and knowledge necessary to diagnose abnormalities in the process plant and formulate and synthesize a countermeasure plan from the above plant database and knowledge base. It is equipped with an inference mechanism that reads the data and performs inferences for diagnosing abnormalities and formulating and synthesizing countermeasure plans, as well as a man-machine interface that displays the results of the inferences made by this inference mechanism. By searching for abnormalities, diagnosing them, and formulating and synthesizing countermeasure plans based on the structure, systematic abnormality diagnosis and countermeasure planning and countermeasure planning can be carried out widely and quickly, allowing operators to quickly detect abnormalities in the plant. We can help you respond. In particular, it is highly effective in that it is not necessary to envisage individual abnormality causes and abnormal patterns in advance as in the past, and it is possible to diagnose unknown abnormality causes and formulate and synthesize countermeasure plans.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the operation support device for a process plant according to the present invention, and FIG. 2 shows a functional hierarchical structure for power production in a nuclear power plant provided in the knowledge base of the above embodiment. Figure 3 is a diagram showing the functional hierarchy structure regarding radiation release prevention in nuclear power plants provided in the knowledge base of the above embodiment, and Figure 4 is a diagram showing the safety of nuclear power plants provided in the knowledge base of the above embodiment. FIG. 5 is a diagram showing the functional hierarchical structure for maintaining the integrity of the containment vessel of a nuclear power plant, which is provided in the knowledge base of the above embodiment, and FIG. FIG. 5 shows the symbols used, FIGS. 7 and 8 show examples of basic measures provided in the knowledge base of the above embodiment, and FIG. 9 shows the symbols used in the above embodiment. FIG. 10 is a diagram that summarizes the judgment results for each flow function in the above embodiment for an arbitrary flow function. FIG. FIG. 12 is a diagram showing criteria for selecting a new driving target, and is a diagram showing a hierarchical structure of basic measures created using the planning knowledge of the above embodiment. 1... Nuclear power plant, 2... Abnormality diagnosis device,
3...Plant database, 4...Knowledge base,
5... Reasoning mechanism, 6... Man-machine interface.

Claims (1)

[Claims] 1. A plant database that stores plant data from process plants that produce substances and energy using physical processes and chemical processes, and the formulation and synthesis of abnormality diagnosis and countermeasure plans for the process plants. A knowledge base that has the knowledge to perform abnormality diagnosis and countermeasure plan formulation/synthesis is loaded by reading the plant data and knowledge necessary for abnormality diagnosis of process plants and formulation/synthesis of countermeasure plans from the above-mentioned plant database and knowledge base. 1. An operation support device for a process plant, comprising: an inference mechanism that makes an inference; and a man-machine interface that displays the results of the inference made by the inference mechanism. 2. The knowledge base is knowledge for diagnosing abnormalities in process plants and formulating and synthesizing countermeasure plans.
Knowledge that describes the functional hierarchical structure of multiple process plants that correspond to multiple operating states, knowledge for diagnosing abnormalities for each node of those functional hierarchical structures, and knowledge for diagnosing abnormalities for the entire plant. , the knowledge to select a new operational target based on the abnormality diagnosis results, the knowledge to formulate a countermeasure plan necessary to achieve the new operational target, and the knowledge to ensure that this countermeasure plan is consistent with the entire plant. 2. The process plant operation support system according to claim 1, having knowledge for synthesizing into an overall plan.