CN118195431B - Method, system and application for establishing nuclear power engineering item relation model - Google Patents

Abstract

The invention provides a method, a system and an application for establishing a nuclear power engineering item relation model, and belongs to the technical field of data modeling. The method for establishing the nuclear power engineering item relation model comprises the following steps: step S1, engineering data acquisition; s2, converting model data; and S3, modeling the object relation. The system for establishing the nuclear power engineering item relation model comprises an engineering data acquisition module, a model data conversion module and an item relation modeling module. When the nuclear power engineering item relation model is applied, the item relation model obtained by the method for establishing the nuclear power engineering item relation model is adopted, and then item relation information is checked. The method is used for checking the item relation information in the nuclear power engineering digital information system by establishing the standardized item relation model so as to ensure the accuracy of the item relation information.

Description

Method, system and application for establishing nuclear power engineering item relation model

Technical Field

The invention relates to the technical field of data modeling, in particular to a method and a system for establishing a nuclear power engineering item relation model and application thereof.

Background

The nuclear power engineering is a system engineering with huge scale and extremely high complexity. Therefore, the accuracy of item relation information in the nuclear power engineering is an important foundation for ensuring the smooth implementation of the nuclear power engineering project.

At present, the construction of nuclear power engineering mainly depends on drawings and documents. Taking the design business of nuclear power engineering as an example, the information exchange is carried out on the working contents of each professional through a design information management system. However, there are some significant drawbacks to the exchange of information based on drawings and documents:

First, changes in engineering information (e.g., changes in item relationships, changes in design data, etc.) can result in a significant decrease in information exchange efficiency. For example, in the design process of a nuclear power plant, an "A valve" and a "C valve" are originally in pipeline connection, and a designer changes the "A valve" and the "D valve" into pipeline connection for some reason. The occurrence of the above-described modification behavior may cause manual modification of all drawings and documents related to the information, thereby greatly reducing engineering efficiency.

Second, due to the large number of drawings and documents, the probability of mistakes by engineering personnel is greatly increased. Therefore, related enterprises of nuclear power engineering gradually transform and develop engineering construction work in a digital form. All nuclear power engineering information is gradually converted into a computer-readable digital form from the expression forms of drawings and documents through a digital technology. All engineering item information will form item objects through a reasonable abstraction process. The object and object relationships are expressed by an object-oriented model and converted into a json file or xml file which is readable by a computer. Although the digital technology brings convenience to the construction process of engineering personnel, the engineering personnel still face a high probability of abnormal events in the design or construction process.

Therefore, the digital transformation of nuclear power engineering is in need of an item relation model for checking item relation information in a digital information system of nuclear power engineering so as to ensure the accuracy of the item relation information.

Disclosure of Invention

The invention aims to provide a method, a system and an application for establishing a nuclear power engineering item relation model, which are used for checking item relation information in a nuclear power engineering digital information system by establishing a standardized item relation model so as to ensure the accuracy of the item relation information.

The technical scheme adopted by the invention is as follows:

a method for establishing a nuclear power engineering item relation model comprises the following steps:

step S1, engineering data acquisition: acquiring object metadata and object relations in an external data source forming a database of a nuclear power engineering digital information system to obtain an object metadata database table and an object relation database table; classifying the object item relations in the object item relation database table to obtain a plurality of different relation types;

step S2, model data conversion: according to the input data form requirement input by FPgrowth algorithm, converting the relation information of the item relation under each relation type in the item relation database table into an item name two-dimensional list, and expressing the item name two-dimensional list by json format or xml format;

Step S3, modeling object item relation: reading json format files or xml format files, and analyzing the two-dimensional list of the object names under each relation type one by one; respectively carrying out list data aiming at the two-dimensional list of the object names under one relationship type, and constructing an FP tree until all object sets in the two-dimensional list of the object names under the relationship type are expressed in the FP tree, so as to obtain an FP tree model corresponding to the relationship type; and repeating the steps to obtain the FP tree model corresponding to each relation type, namely the object relation model corresponding to each relation type.

Further, in the step S1, when the metadata of the object item is collected, the metadata is implemented through a text recognition technology and/or a semantic entity recognition technology.

Further, in the step S1, when the item relationships are collected, the relationship information between the periodic items is obtained by using a semantic relationship recognition technology based on the item metadata.

Further, in the step S2, the specific process of model data conversion is as follows:

step S21, obtaining the relation information and relation information codes of the item relation under each relation type in the item relation database table;

step S22, two item name codes corresponding to two item examples in each piece of relation information are obtained;

If the relation information codes in the object relation database table are unique, converting two object name codes in the relation information codes into an object name two-dimensional list;

If two or more than two identical relation information codes exist in the item relation database table, combining the item name codes of all relation information corresponding to the relation information codes, and converting the item name codes into an item name two-dimensional list;

And S23, outputting json format or xml format data according to the input data form requirement input by FPgrowth algorithm.

Further, in the step S3, the specific process of building the FP-tree is:

step S31, scanning a two-dimensional list of object names under a relation type, finding out the occurrence frequency of each object, and sorting the objects according to the frequency;

step S32, constructing an initial node of the FP tree;

For each item set in the item name two-dimensional list, according to the sorting result in step S31, connecting the item as a node with a node in the FP tree, and marking the frequency of the node at the node;

Repeating the steps until all item sets in the item name two-dimensional list under the relation type are expressed in the FP tree.

Further, in the step S32, the specific process of node connection in the FP-tree is:

For the first node after sequencing, observing whether the node which is the same as the node exists in the next level child node of the initial node; if the node does not appear in the next level child node of the initial node, a new node is constructed, the frequency is marked as 1, and the newly constructed first node is connected with the initial node; if the node exists, the node is directly accumulated by 1 without constructing a new node;

for the second and subsequent nodes, observing whether the next level child node of the father node has the same node as the node; the father node is the node of the upper level of the node; if the child node exists, the new node is not constructed, and the node is accumulated for 1 directly; if the child node does not exist, a new node is constructed, the frequency is marked as 1, and the new node is connected with the parent node.

Based on the same inventive concept, the invention also provides a system for establishing the nuclear power engineering item relation model, so as to implement the method for establishing the nuclear power engineering item relation model, which comprises the following steps:

The engineering data acquisition module is used for acquiring object metadata and object relations in an external data source forming a database of the nuclear power engineering digital information system to obtain an object metadata database table and an object relation database table; classifying the object item relations in the object item relation database table to obtain a plurality of different relation types;

The model data conversion module is used for converting the relation information of the object relation under each relation type in the object relation database table into an object name two-dimensional list according to the input data form requirement input by FPgrowth algorithm, and expressing the object name two-dimensional list in json format or xml format;

The object relation modeling module is used for reading json format files or xml format files and analyzing the object name two-dimensional list under each relation type one by one; respectively carrying out list data aiming at the two-dimensional list of the object names under one relationship type, and constructing an FP tree until all object sets in the two-dimensional list of the object names under the relationship type are expressed in the FP tree, so as to obtain an FP tree model corresponding to the relationship type; and repeating the steps to obtain the FP tree model corresponding to each relation type, namely the object relation model corresponding to each relation type.

Based on the same inventive concept, the invention also provides application of the nuclear power engineering item relation model in item relation information inspection, and the item relation model obtained by the method for establishing the nuclear power engineering item relation model is adopted, and then the item relation information inspection is carried out.

Further, the specific process of the item relation information inspection is as follows:

Step S41, obtaining the relation and the name of the items to be searched, wherein the number of the items to be searched is more than or equal to 2;

retrieving all item names in an item metadata database;

If the object item name does not exist in the object item metadata database, outputting abnormal early warning;

If the item name exists in the item metadata database, executing the next step;

Step S42, calculating item relation support degree between every two items to be searched on the basis of an item relation model corresponding to the item relation to be searched, and outputting corresponding item relation abnormal early warning level according to the item relation support degree and a threshold value;

the object relation support degree is obtained by the following steps:

，

wherein, Namely the probability of the incidence relation between the item X ₁ and the item X ₀ under the item relation k, namely the support degree of the incidence relation between the item X ₁ and the item X ₀;

representing the number of times item X ₀ and item X ₁ occur simultaneously;

n _k represents the total number of item sets contained by relationship type k.

Further, in the step S42, when outputting the corresponding item relationship abnormal early warning level according to the item relationship support level and the threshold value, the specific process is as follows:

setting 3 thresholds as P1, P2 and P3 respectively, wherein P1 is less than P2 and P3, and forming four abnormal early warning levels of no abnormality, slight abnormality, moderate abnormality and serious abnormality in one-to-one correspondence;

outputting corresponding abnormal early warning grades according to the condition that Q falls into four interval ranges;

wherein, ；

The probability of the association between item X ₁ and item X ₀ in item relationship k is the support of the association between item X1 and item X0.

The beneficial effects of the invention are as follows:

The invention provides a modeling method, a modeling system and application of a nuclear power engineering item relation model, wherein the obtained item relation model utilizes a data model to carry out standardized expression on relations among all items of the nuclear power engineering. Through the standardized object relation model, related software and an information system can be used for checking the design information of the nuclear power engineering and the correctness and rationality of engineering behaviors so as to ensure the accuracy of relation information among objects, ensure the reliability of engineering quality and improve the construction efficiency of the nuclear power engineering.

Drawings

Fig. 1 is an extraction schematic diagram of item metadata and item relationships of an external data source of a nuclear power engineering digital information system in embodiment 1.

FIG. 2 is a schematic representation of the first set of items "M-linked" of example 1 expressed on the FP tree.

FIG. 3 is a schematic representation of the second set of items "M-linked" of example 1 expressed on the FP tree.

FIG. 4 is a schematic representation of the third set of items "M-linked" of example 1 expressed on the FP tree.

FIG. 5 is a schematic representation of the fourth set of items "M-linked" of example 1 expressed on the FP tree.

Fig. 6 is an FP-tree model corresponding to the "M-connection" relationship type in embodiment 1.

Detailed Description

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

Embodiment 1, a method for establishing a nuclear power engineering item relation model, comprising the following steps:

Step S1, engineering data acquisition: and acquiring object metadata and object relations in an external data source forming a database of the nuclear power engineering digital information system to obtain an object metadata database table and an object relation database table. Meanwhile, the object item relations in the object item relation database table are classified, and a plurality of different relation types are obtained through division.

Step S2, model data conversion: according to the input data form requirement input by FPgrowth algorithm, converting the relation information of item relation under each relation type in the item relation database table into an item name two-dimensional list, and expressing by json format or xml format.

Step S3, modeling object item relation: reading json format files or xml format files, and analyzing the two-dimensional list of the object names under each relation type one by one; respectively carrying out list data aiming at the two-dimensional list of the object names under one relationship type, and constructing an FP tree until all object sets in the two-dimensional list of the object names under the relationship type are expressed in the FP tree, so as to obtain an FP tree model corresponding to the relationship type; and repeating the steps to obtain the FP tree model corresponding to each relation type, namely the object relation model corresponding to each relation type.

In this embodiment, the external data sources that constitute the database of the digital information system of the nuclear power engineering include nuclear power design drawings and documents, construction and installation drawings and documents, operation and maintenance drawings and documents, and the like. The extract item metadata and extract item relationships are the core steps when data processing is performed by an external data source. In the metadata extraction process, each periodic object item entity can be obtained through a text recognition technology and a semantic entity recognition technology to form object item metadata. Meanwhile, in the process of extracting the object relationships, the relationship information among the periodic object items can be obtained by utilizing a semantic relationship recognition technology on the basis of the object metadata.

For example, a schematic diagram of the extraction of item metadata and item relationships of an external data source of a nuclear power engineering digital information system is shown in fig. 1. The item metadata is centrally processed and stored to form an item metadata database for each stage. Taking partial nuclear power design item metadata as an example, the item metadata database table is shown in table 1 below.

Table 1 item metadata database structure example

The item relation information is processed and stored in a centralized way to form an item relation database of each stage. The code of the object example code is composed of an object name code and an engineering digital code, for example, the object example code 'EPBAAD 0101001' is composed of the object name code 'EPBAAD' in the table 1 and the diaphragm valve engineering digital code '0101001'. Through item instance encoding, the system may decompose the encoding to obtain an item name encoding for the item. Taking partial nuclear power design item relationship data as an example, the item relationship database table is shown in table 2 below.

Table 2 item relational database structure example

In this embodiment, the specific process of model data conversion is as follows:

step S21, obtaining the relation information and relation information codes of the item relation under each relation type in the item relation database table;

step S22, two item name codes corresponding to two item examples in each piece of relation information are obtained;

If the relation information codes in the object relation database table are unique, converting two object name codes in the relation information codes into an object name two-dimensional list;

If two or more than two identical relation information codes exist in the item relation database table, combining the item name codes of all relation information corresponding to the relation information codes, and converting the item name codes into an item name two-dimensional list;

And S23, outputting json format or xml format data according to the input data form requirement input by FPgrowth algorithm.

In this embodiment, json format data is taken as an example.

The json format input by FPgrowth algorithm is expressed generally as: { relationship type: two-dimensional list of item names }.

The following are examples of refined json format data:

{

"relation type 1

[ "Item name encodes 1", "item name encodes 2" ],

[ "Item name code 1", "item name code 2", "item name code 3" ],

[ "Item name encodes 1", "item name encodes 2" ],

[ "Item name code 1", "item name code 2", "item name code 3" ]

]，

"Relation type 2

[ "Item name encodes 1", "item name encodes 2" ],

[ "Item name code 1", "item name code 2", "item name code 3" ],

[ "Item name encodes 1", "item name encodes 3" ],

[ "Item name code 1", "item name code 4", "item name code 2" ]

[ "Item name code 1", "item name code 4", "item name code 5" ]

[ "Item name code 1", "item name code 3", "item name code 5" ]

[ "Item name code 1", "item name code 4", "item name code 6" ]

]

}。

In this embodiment, the specific process of constructing the FP-tree is:

step S31, scanning a two-dimensional list of object names under a relation type, finding out the occurrence frequency of each object, and sorting the objects according to the frequency;

step S32, constructing an initial node of the FP tree;

For each item set in the item name two-dimensional list, according to the sorting result in step S31, connecting the item as a node with a node in the FP tree, and marking the frequency of the node at the node;

Repeating the steps until all item sets in the item name two-dimensional list under the relation type are expressed in the FP tree.

Further, in the step S32, the specific process of node connection in the FP-tree is:

For the first node after sequencing, observing whether the node which is the same as the node exists in the next level child node of the initial node; if the node does not appear in the next level child node of the initial node, a new node is constructed, the frequency is marked as 1, and the newly constructed first node is connected with the initial node; if the node exists, the node is directly accumulated by 1 without constructing a new node;

for the second and subsequent nodes, observing whether the next level child node of the father node has the same node as the node; the father node is the node of the upper level of the node; if the child node exists, the new node is not constructed, and the node is accumulated for 1 directly; if the child node does not exist, a new node is constructed, the frequency is marked as 1, and the new node is connected with the parent node.

The FP-tree construction process is described in detail below by taking the two-dimensional list of item names (json format) of "M-join" relationship types as an example. The two-dimensional list of item names for the "M-connection" relationship type is as follows:

{ "M connection"

[ "B valve", "A pump", "C valve" ],

[ "E valve", "D pump" ],

[ "C valve", "F pump" ],

[ "B valve", "D pump", "A pump ]

[ "B valve", "A pump", "E valve" ]

[ "B valve", "C valve", "E valve" ]

[ "B valve", "C valve", "F pump" ]

]

}。

After scanning the two-dimensional list of item names, the following frequency and ordering of 6 items can be obtained, as follows:

Valve b → 5 times, valve 2.C → 4 times, pump 3.A → 3 times, 4.E valves → 3 times, pump 5.D → 2 times, pump 6.F → 2 times.

An initial node null of the FP-tree is constructed. Taking the first item set of the 'M connection' relation type as an example, the first item set is 'B valve', 'A pump', 'C valve', and the first step sequencing result is as follows: valve B, valve C, pump A, and then the nodes B, valve C, pump A are respectively connected with null node, and the occurrence frequency is marked, as shown in figure 2. The second item set is [ "E valve", "D pump" ], and the result according to the first step ordering is: e valve→D pump. If the first node 'E valve' does not exist in the child node of null, an E valve node is constructed and connected with the null node. The second node "D pump" is not present in the next level child node of its parent node "E valve" ("E valve" has no next level child node), then the D pump node is constructed to be connected to the E valve as shown in fig. 3. The third item set is [ "C valve", "F pump" ], and the result according to the first step ordering is: c valve→F pump. If the first node 'C valve' does not exist in the child node of null, a C valve node is constructed and connected with the null node. The second node "F pump" is not present in the child node of the next level of its parent node "C valve", and the build F pump node is connected to the last node C valve, as shown in FIG. 4. The fourth item set is [ "B valve", "D pump", "A pump" ], and the result according to the first step sequence is: b valve → a pump → D pump. The first node "B valve" is in the child node of null, then the frequency of child node B valve nodes of null is accumulated by 1. The second node "a pump" is not present in the next level of child nodes of node "B valve" ("C valve" is the next level of child nodes of "B valve"), and the build a pump node is connected to the last node B valve. If the third node "D pump" does not exist in the next level child node of the previous node "a pump", the D pump node is constructed to be connected to the previous node a pump, as shown in fig. 5. And repeating the steps until all item sets in the two-dimensional list of item names under the M connection type are expressed in the FP tree, and obtaining the following FP tree model as shown in FIG. 6.

According to the modeling method for the nuclear power engineering item relation model, the obtained item relation model utilizes the data model to carry out standardized expression on the relation among all items of the nuclear power engineering, and through the standardized item relation model, related software and an information system can be used for checking design information of the nuclear power engineering and correctness and rationality of engineering behaviors so as to ensure accuracy of relation information among the items, guarantee reliability of engineering quality and improve construction efficiency of the nuclear power engineering.

Meanwhile, the data acquisition and storage mode provided in the embodiment is beneficial to data structuring and full utilization of data in the modeling process. And the tree structure is adopted to express the structure of the association relation between nuclear power engineering items, so that the model structure can flexibly express the nuclear power engineering item relation information, and the tree structure is also convenient for the realization of an IT system.

Embodiment 2 provides a system for establishing a nuclear power engineering item relationship model in this embodiment, so as to implement the method for establishing a nuclear power engineering item relationship model in the foregoing embodiment 1. The system for establishing the nuclear power engineering item relation model comprises an engineering data acquisition module, a model data conversion module and an item relation modeling module.

Specifically, the engineering data acquisition module is used for acquiring the item metadata and the item relation in an external data source forming a database of the nuclear power engineering digital information system to obtain an item metadata database table and an item relation database table. And classifying the object relationships in the object relationship database table to obtain a plurality of different relationship types. The model data conversion module is used for converting the relation information of the object relation under each relation type in the object relation database table into an object name two-dimensional list according to the input data form requirement input by FPgrowth algorithm, and expressing the object name two-dimensional list in json format or xml format. The object relation modeling module is used for reading json format files or xml format files and analyzing object name two-dimensional lists under each relation type one by one; respectively carrying out list data aiming at the two-dimensional list of the object names under one relationship type, and constructing an FP tree until all object sets in the two-dimensional list of the object names under the relationship type are expressed in the FP tree, so as to obtain an FP tree model corresponding to the relationship type; and repeating the steps to obtain the FP tree model corresponding to each relation type, namely the object relation model corresponding to each relation type.

Embodiment 3 provides an application of a nuclear power engineering item relation model in item relation information inspection, and the item relation model obtained by the method for establishing the nuclear power engineering item relation model in the embodiment 1 is adopted, and then the item relation information inspection is performed.

Further, the specific process of the item relation information inspection is as follows:

Step S41, obtaining the relation and the name of the items to be searched, wherein the number of the items to be searched is more than or equal to 2;

retrieving all item names in an item metadata database;

If the object item name does not exist in the object item metadata database, outputting abnormal early warning;

If the item name exists in the item metadata database, executing the next step;

Step S42, calculating item relation support degree between every two items to be searched on the basis of an item relation model corresponding to the item relation to be searched, and outputting corresponding item relation abnormal early warning level according to the item relation support degree and a threshold value;

the object relation support degree is obtained by the following steps:

，

wherein, Namely the probability of the incidence relation between the item X ₁ and the item X ₀ under the item relation k, namely the support degree of the incidence relation between the item X ₁ and the item X ₀;

representing the number of times item X ₀ and item X ₁ occur simultaneously;

n _k represents the total number of item sets contained by relationship type k.

For example, the item relationship support degree of "M-linked" with the item "C valve" was studied by taking the item set of "M-linked" described in example 1 as an example. From the bottom of the FP-tree, items that "M-connect" to the C-valve can be found to be a-pump, F-pump, E-valve, B-valve. The support degree of the C valve to the F pump is as follows:

，

The support degree of the association relation between the item 'C valve' and the item 'F pump';

representing the number of times item "C valve" and item "F pump" occur simultaneously;

n _{M Connection} represents the total number of item sets contained by the relationship type "M connection".

The calculation result shows that the probability of the association relation between the C valve and the F pump is 0.286 for a specific M connection relation.

Further, in the step S42, when outputting the corresponding item relationship abnormal early warning level according to the item relationship support level and the threshold value, the specific process is as follows:

Setting 3 thresholds as P1, P2 and P3 respectively, wherein P1< P2< P3, forming four interval ranges as [0, P1), [ P1, P2), [ P2, P3) and [ P3,1], and correspondingly setting four abnormality early warning levels without abnormality, slight abnormality, moderate abnormality and serious abnormality one by one, wherein the corresponding relation is shown in a table 3;

outputting corresponding abnormal early warning grades according to the condition that Q falls into four interval ranges;

wherein, ；

The probability of the incidence relation between the item X1 and the item X0 in the item relation k is that the support degree of the incidence relation between the item X1 and the item X0 is provided.

Table 3 level of anomaly early warning

For example, the item relationship information is checked by the method described above, and the result patterns displayed are shown in table 4 below.

Table 4 test results

In the embodiment, the association degree between the items can be reflected by outputting the probability value of the association relation of the items, so that the actual service requirements are met better. Meanwhile, the abnormal item relation is subjected to multi-level early warning, the result is more visual, and the use of related software products is facilitated.

Claims (9)

1. The method for establishing the nuclear power engineering item relation model is characterized by comprising the following steps of:

step S1, engineering data acquisition: acquiring object metadata and object relations in an external data source forming a database of a nuclear power engineering digital information system to obtain an object metadata database table and an object relation database table; classifying the object item relations in the object item relation database table to obtain a plurality of different relation types;

step S2, model data conversion: according to the input data form requirement input by FPgrowth algorithm, converting the relation information of the item relation under each relation type in the item relation database table into an item name two-dimensional list, and expressing the item name two-dimensional list by json format or xml format;

Step S3, modeling object item relation: reading json format files or xml format files, and analyzing the two-dimensional list of the object names under each relation type one by one; respectively carrying out list data aiming at the two-dimensional list of the object names under one relationship type, and constructing an FP tree until all object sets in the two-dimensional list of the object names under the relationship type are expressed in the FP tree, so as to obtain an FP tree model corresponding to the relationship type; repeating the steps to obtain an FP tree model corresponding to each relationship type, namely an item relationship model corresponding to each relationship type;

In the step S3, the specific process of building the FP-tree is:

step S31, scanning a two-dimensional list of object names under a relation type, finding out the occurrence frequency of each object, and sorting the objects according to the frequency;

step S32, constructing an initial node of the FP tree;

For each item set in the item name two-dimensional list, according to the sorting result in step S31, connecting the item as a node with a node in the FP tree, and marking the frequency of the node at the node;

Repeating the steps until all item sets in the item name two-dimensional list under the relation type are expressed in the FP tree.

2. The method for building a nuclear power engineering item relation model according to claim 1, wherein in the step S1, the item metadata is collected by a word recognition technology and/or a semantic entity recognition technology.

3. The method for building a nuclear power engineering item relation model according to claim 1, wherein in the step S1, when the item relation is collected, relation information between periodic items is obtained by using a semantic relation recognition technology based on item metadata.

4. The method for building a nuclear power engineering item relation model according to claim 1, wherein in the step S2, the specific process of model data conversion is as follows:

step S21, obtaining the relation information and relation information codes of the item relation under each relation type in the item relation database table;

step S22, two item name codes corresponding to two item examples in each piece of relation information are obtained;

If the relation information codes in the object relation database table are unique, converting two object name codes in the relation information codes into an object name two-dimensional list;

If two or more than two identical relation information codes exist in the item relation database table, combining the item name codes of all relation information corresponding to the relation information codes, and converting the item name codes into an item name two-dimensional list;

And S23, outputting json format or xml format data according to the input data form requirement input by FPgrowth algorithm.

5. The method for building a nuclear power engineering item relation model according to claim 1, wherein in step S32, the specific process of node connection in the FP-tree is:

For the first node after sequencing, observing whether the node which is the same as the node exists in the next level child node of the initial node; if the node does not appear in the next level child node of the initial node, a new node is constructed, the frequency is marked as 1, and the newly constructed first node is connected with the initial node; if the node exists, the node is directly accumulated by 1 without constructing a new node;

for the second and subsequent nodes, observing whether the next level child node of the father node has the same node as the node; the father node is the node of the upper level of the node; if the child node exists, the new node is not constructed, and the node is accumulated for 1 directly; if the child node does not exist, a new node is constructed, the frequency is marked as 1, and the new node is connected with the parent node.

6. A system for building a nuclear power engineering item relationship model to implement the method for building a nuclear power engineering item relationship model according to any one of claims 1 to 5, comprising:

The engineering data acquisition module is used for acquiring object metadata and object relations in an external data source forming a database of the nuclear power engineering digital information system to obtain an object metadata database table and an object relation database table; classifying the object item relations in the object item relation database table to obtain a plurality of different relation types;

The model data conversion module is used for converting the relation information of the object relation under each relation type in the object relation database table into an object name two-dimensional list according to the input data form requirement input by FPgrowth algorithm, and expressing the object name two-dimensional list in json format or xml format;

The object relation modeling module is used for reading json format files or xml format files and analyzing the object name two-dimensional list under each relation type one by one; respectively carrying out list data aiming at the two-dimensional list of the object names under one relationship type, and constructing an FP tree until all object sets in the two-dimensional list of the object names under the relationship type are expressed in the FP tree, so as to obtain an FP tree model corresponding to the relationship type; and repeating the steps to obtain the FP tree model corresponding to each relation type, namely the object relation model corresponding to each relation type.

7. An application of a nuclear power engineering item relation model in item relation information inspection, which is characterized in that the item relation model obtained by the method for establishing the nuclear power engineering item relation model according to any one of claims 1-5 is adopted, and then the item relation information inspection is carried out.

8. The application of the nuclear power engineering item relation model in item relation information inspection according to claim 7, wherein the specific process of the item relation information inspection is as follows:

Step S41, obtaining the relation and the name of the items to be searched, wherein the number of the items to be searched is more than or equal to 2;

retrieving all item names in an item metadata database;

If the object item name does not exist in the object item metadata database, outputting abnormal early warning;

If the item name exists in the item metadata database, executing the next step;

Step S42, calculating item relation support degree between every two items to be searched on the basis of an item relation model corresponding to the item relation to be searched, and outputting corresponding item relation abnormal early warning level according to the item relation support degree and a threshold value;

the object relation support degree is obtained by the following steps:

；

wherein, Namely the probability of the incidence relation between the item X ₁ and the item X ₀ under the item relation k, namely the support degree of the incidence relation between the item X ₁ and the item X ₀;

representing the number of times item X ₀ and item X ₁ occur simultaneously;

n _k represents the total number of item sets contained by relationship type k.

9. The application of the nuclear power engineering item relation model according to claim 8 in item relation information inspection, wherein in the step S42, when outputting the corresponding item relation abnormal early warning level according to the item relation support degree and the threshold value, the specific process is as follows:

Setting 3 thresholds as P1, P2 and P3 respectively, wherein P1 is less than P2 and less than P3, and forming four abnormal early warning levels of no abnormality, slight abnormality, moderate abnormality and serious abnormality in one-to-one correspondence with four interval ranges of [0, P1 ], [ P1, P2), [ P2, P3) and [ P3,1 ];

outputting corresponding abnormal early warning grades according to the condition that Q falls into four interval ranges;

wherein, ；

The probability of the association between item X ₁ and item X ₀ in item relationship k is the support of the association between item X ₁ and item X ₀.