JPS63302318A - Support apparatus for forming operation program of nuclear power plant - Google Patents

Abstract

PURPOSE:To enable reduction in required time for preparation of an operation program according to a demand for the preparation of an urgent and quick operation program, by performing a learning using an achievement data for optimization of programs with respect to knowledge for the preparation of programs. CONSTITUTION:When an urgent corrected operation program is prepared, a rough operation program formed at an operation program preparing section 1 is provided to a user through an interface section 4. An operation program is optimized by a program evaluating/correcting section 2 and the results are provided to the user. Then, the correcting section 2 analyzes the condition of a core using a core simulator 5 at each of operating points serving as reference for alteration of an outputs of the rough operation program and the results are processed in terms of various quantities of state, thermal margin and the like of the core. Subsequently, the results are converted into an evaluation index using an evaluation correcting knowledge base 7 from these quantities of state to calculate a desired function for optimization. Moreover, knowledge of program evaluation and correction methods is learned 3 from an achievement data for the optimization of programs to improve the evaluation and correction methods automatically to reduce the frequency of repeated calculation thereby enabling the shortening of the required time.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to the creation of an operation plan for a nuclear power plant, and in particular,
The present invention relates to an operation plan creation support device that can reduce the time required for repeated calculations of core analysis, evaluation, and correction for operation plan creation.

[Conventional technology]

An example of an invention related to creating an operation plan for a nuclear power plant is Japanese Patent Application Laid-Open No. 57-37296. In this example,
A method and apparatus configuration for determining an optimal outlet cover pattern for determining a load change pattern that satisfies core operation restriction conditions during load following operation are shown.

[Problem that the invention seeks to solve]

In a known example, when determining the optimal reactor discharge cover turn, a core simulator is used to perform predictive calculations on the state of the core, and a load change pattern that is closest to the load change request pattern is determined while satisfying the limiting conditions such as the thermal margin of the core and the operating range. In order to achieve the desired optimization, the operation plan is evaluated and revised, and this process is repeated to create the final operation plan. However, with this conventional method, emergency and quick operations such as an abnormality in the plant's output control device or an unscheduled load change request from intermediate supply, etc.
There is a problem in that when there is a request to create a single drawing, the time required to create a plan is not necessarily short enough to satisfy the requirement. The main reason for this is the calculation time required to repeat the core state analysis using a core simulator for multiple reference operating points on the operation plan. In order to reduce the calculation time of the simulator, a simplified core model such as a -dimensional core simulator can be used, but compared to a detailed three-dimensional core simulator, the calculation accuracy of the core thermal margin etc. is slightly lower. It becomes a sacrifice.

An object of the present invention is to provide an operation plan creation support device for a nuclear power plant that solves the problem of shortening the time required to create a plan in order to meet the demand for urgent and quick operation plan creation. .

[Means for solving problems]

The above purpose is.

1) When the urgency of the plan creation request is extremely high, plan optimization using a core simulator that repeatedly analyzes, evaluates, and corrects core characteristics is not performed, but only the knowledge of planning experts prepared in advance is used. Create and provide operation plans using

2) Reduce the number of repeated calculations by learning the knowledge that determines the plan evaluation/correction method from plan optimization performance data and automatically improving the evaluation/correction method.

This can be achieved by using two methods.

In particular, by training plan creation knowledge using actual plan optimization data, it is possible to improve the quality of operation plans created using only knowledge without using a core simulator. Functions that can withstand practical use can be realized.

[Effect]

When implementing an unscheduled load change in a plant, experienced core management engineers say that an operational plan (i.e., core flow rate and control rod pattern Change plans) can be created. Therefore, if this knowledge of operation planning can be further improved by learning the process of plan optimization, plans can be created to quickly provide operation plans without using a core simulator, depending on the urgency of the request. Support equipment can be constructed. Furthermore, although there are various methods for learning knowledge, learning by analogy is applicable because the knowledge used in plan creation, evaluation, and modification is organized into specialized knowledge that can be expressed in a specific format. In the specific application of this learning method, the following steps are required: 1) Reflection of the modification contents affected by plan optimization in the planning knowledge, 2) Modification by evaluating the effect of the knowledge used at each iterative modification step of the optimization process, 3) We have devised three methods for generating and adding knowledge by aggregating performance data from the optimization process to determine optimization strategies.

〔Example〕

An embodiment of the nuclear power plant operation plan creation support device of the present invention will be described below. FIG. 1 shows the basic configuration of the embodiment, which utilizes knowledge processing technology called knowledge engineering and separates the planning algorithm into an inference part and a knowledge base. The driving plan creation unit 1 receives a user's plan creation request from the interface unit 4 and creates a rough driving plan using the planning knowledge base 6. Types of operation plan creation include plant restart, load following, control rod pattern exchange/adjustment, and creation of a revised operation plan in response to the occurrence of a control device abnormality, and the user inputs required specifications. If the user's request is to create an urgently revised driving plan, the rough driving plan created by the driving plan creation unit 1 is provided to the user through the interface. Other than this, the plan evaluation correction unit 2 optimizes the operation plan,
Provide the results to the user.

The plan evaluation modification unit 2 first analyzes the core state using the core simulator 5 for each of a plurality of operating points that serve as a reference for changing the output of the rough operating plan. The analysis results are based on various states of the core, gffi (for example, power distribution).

(quality, local peaking coefficient, etc.), thermal margin (for example, minimum critical heat output ratio, margin for maximum linear power density limit value), etc. Next, these state quantities are converted into evaluation indicators using the evaluation correction knowledge base 7, and an objective function for optimization is calculated. Plan optimization is achieved by modifying the plan in a direction that maximizes this objective function, and repeating analysis, evaluation, and modification of the core state.

A detailed configuration block of the learning section 3 is shown in FIG.

The procedure for learning planning knowledge is as follows: First, the plan correction point extraction unit 11 compares and matches the general operation plan 18 and the detailed operation plan 19 that has been optimized for each operation point, and the plan is corrected by optimization. Extract the points. Next, planning knowledge learning part 1
2, the knowledge used in the above-mentioned plan modification points is searched, the knowledge is modified to match the plan modification contents, and the plan creation knowledge base 6 is replaced. These learning procedures are shown in FIG.

Figure 4 shows the procedure for learning evaluation correction knowledge. After completing optimization by repeatedly evaluating and modifying the operation plan.

The evaluation index calculation unit 13 calculates the core state evaluation index for each operating point of the general operation plan 18 and detailed operation plan 19, and determines the amount of change. On the other hand, the state quantity change calculation unit 14 calculates the change in the core state quantity before the plan modification and the core state quantity after the plan modification for each operating point in each iterative calculation step of plan optimization, and stores it in the storage unit 15. do. Plan revision knowledge learning part 1
In step 7, the knowledge of plan revisions used in each iterative calculation step is searched, and the knowledge of plan revisions is corrected by statistically processing the correspondence actual data between changes in core state quantities and changes in plan revisions. do. Next, the plan evaluation knowledge learning section 16
calculates the sum of changes in the core state quantity that occurred during the plan optimization process, and calculates this, the evaluation index at the start of plan optimization output from the evaluation index calculation unit 13, the objective function, and the sum of changes associated with these plan optimizations. The changes are combined to generate plan evaluation knowledge that provides correction targets for core state quantities when optimizing the plan, and is added to the evaluation correction knowledge base 7.

FIG. 5 shows how the learning section of the operation plan support device of the present invention
This figure shows how the speed of plan optimization is improved when plan evaluation/correction knowledge is learned.

If the plan modification is learned, it is possible to determine the plan modification that accurately achieves the core state quantity change target (i.e., modification of the core flow rate and control rod pattern manipulated variable), which speeds up the optimization process. , the number of repeated calculations is reduced. Furthermore, when learning of planning evaluation knowledge is added, knowledge of optimization strategy regarding how to determine the core state modification target in order to maximize the objective function is provided based on the actual results, so the figure As shown, the optimization speed increases and the number of iterative calculation steps decreases.

〔Effect of the invention〕

According to the present invention, depending on the urgency of the operation plan creation request,
It is possible to provide a plan creation support system that can quickly create an operation plan.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of an embodiment of the present invention, Fig. 2 is a calculation block diagram inside the learning section, Fig. 3 is a diagram showing the learning procedure of plan creation knowledge, and Fig. 4 is plan evaluation/correction knowledge. FIG. 5 is a diagram showing how the speed of plan optimization is improved when plan evaluation/correction knowledge is learned.

Claims (1)

[Scope of Claims] 1. An operation plan creation unit that takes in a load change request pattern through an interface and creates target values for control rod patterns and core flow rate manipulation as a function of time using a plan creation knowledge base; and the operation plan. The rough operation plan output from the creation section is taken in, the core state at multiple reference operating points is analyzed using a core simulator, and the operation plan is evaluated and corrected using the knowledge base and the plan creator through the interface. A plan evaluation and correction unit that uses evaluation and correction input to evaluate and correct the operation plan, and then repeats the same process from core state analysis to optimize the operation plan, as well as a plan creation knowledge base and evaluation and correction knowledge. The invention is characterized by comprising a learning section that adds to and corrects the plan creation knowledge base and the evaluation/correction knowledge base by taking in the operation plan evaluation/correction data generated in the plan optimization process of the plan evaluation/correction section. A support device for creating operation plans for nuclear power plants. 2. In claim 1, the general operation plan and the optimized operation plan are compared and matched to extract plan correction points, and then the general operation plan of the plan correction point portion is created. Search for the planning knowledge used,
An operation plan creation support device for a nuclear power plant, characterized in that the content of modification of the planning knowledge retrieved from the content of modification of the plan is determined, and the content of modification of the planning knowledge is replaced with the content of modification. 3. In claim 1, for changing the core state quantity by incorporating changes in the core state quantity before and after the modification used in the repetition of evaluation and modification of the operation plan by the plan evaluation modification unit. The knowledge of the plan modification used to determine the manipulated variable change in the core flow rate or control rod pattern is modified to give a manipulated variable change that brings about a change in the core state quantity, and then the rough operation plan and the evaluation index of the optimized operation plan, and from the sum total of changes in the core state quantity that occurred in the plan optimization process, the evaluation index of the rough operation plan, and the change in the evaluation index due to optimization. 1. A nuclear power plant operation planning support device characterized by adding new knowledge to plan evaluation knowledge for determining correction targets for core state quantities based on plan evaluation using a combination of data.