CN-121998061-A - Intelligent knowledge graph construction method for construction engineering examination

CN121998061ACN 121998061 ACN121998061 ACN 121998061ACN-121998061-A

Abstract

The invention discloses an intelligent knowledge graph construction method for construction engineering examination, which relates to the technical field of construction engineering examination knowledge graphs, and comprises the following steps: s1, acquiring geometric feature data and attribute feature data of building components in a building information model, acquiring a building engineering examination specification, and analyzing the specification into a rule set, wherein the rule set comprises space constraint conditions; s2, determining a spatial topological relation among building components based on the geometric characteristic data, and generating a first relation pair set; the knowledge graph takes building components as nodes and takes a space topological relation and a standard semantic relation as edges, so that the space association between each pair of components and the corresponding examination conclusion thereof are stored in the graph in the form of an explicit relation, and when the basis of a certain illegal conclusion needs to be traced, the original rule entry for generating the conclusion can be directly positioned through graph traversal.

Inventors

XIA LINGLING
ZHANG QIN
HAN WENHAI
HAN FANG
ZHU XIANGDONG
Li Yelu
LUO XIAOLING
YU JU

Assignees

云南业信规划发展集团有限公司

Dates

Publication Date: 20260508
Application Date: 20260408

Claims (9)

1. The intelligent knowledge graph construction method for the construction engineering examination is characterized by comprising the following steps of: S1, acquiring geometric feature data and attribute feature data of building components in a building information model, acquiring a building engineering examination specification, and analyzing the specification into a rule set, wherein the rule set comprises space constraint conditions; S2, determining a spatial topological relation among building components based on the geometric characteristic data, and generating a first relation pair set; S3, matching the space constraint condition in the rule set with the geometric feature data, identifying a component combination meeting the space constraint condition, and generating a second relation pair set according to a matching result; s4, building an architectural engineering examination knowledge graph by taking the building component as a node and taking the relation pair in the first relation pair set and the second relation pair set as an edge.
2. The method for constructing the intelligent knowledge graph for inspecting the building engineering according to claim 1, wherein in the step S1, component type, spatial position coordinate and outline dimension parameter of each building component are extracted from a building information model, spatial range representation of each building component is generated based on the spatial position coordinate and outline dimension parameter, and the standard treaty is analyzed into a rule set.
3. The method for constructing the intelligent knowledge graph for building engineering examination according to claim 2, wherein the method for constructing the intelligent knowledge graph for building engineering examination is characterized by carrying out semantic analysis on the specification strip and identifying the spatial relationship description in the specification strip, and specifically comprises the following steps: Obtaining text data of a standard article, performing word segmentation on the text data, segmenting continuous natural language text into word sequences consisting of independent words, performing part-of-speech tagging on each word in the word sequences, and determining part-of-speech categories corresponding to each word, wherein the part-of-speech categories comprise nouns, verbs, adjectives, prepositions and conjunctions; Acquiring a preset spatial relationship keyword library, wherein a plurality of keywords representing spatial positions or directions are stored in the spatial relationship keyword library, traversing word sequences, comparing each word with the keywords in the spatial relationship keyword library, and marking successfully matched words as candidate spatial relationship words; For each candidate spatial relationship word, capturing a segment containing the candidate spatial relationship word and a preset number of words adjacent to the candidate spatial relationship word from the word sequence, as a context segment corresponding to the candidate spatial relationship word, performing dependency syntax analysis on the context segment, and generating a syntax dependency relationship tree of the context segment, wherein the syntax dependency relationship tree marks a main-predicate relationship, a dynamic guest relationship, a mediate relationship and a modification relationship among the words; Identifying a first noun word and a second noun word which are directly syntactically associated with the candidate spatial relationship word based on the syntactic dependency tree, wherein the first noun word and the second noun word respectively describe two building components related to the spatial relationship; And combining the candidate spatial relationship words, the first noun words and the second noun words to generate a spatial relationship description triplet, wherein the spatial relationship description triplet is the identified spatial relationship description.
4. The method for constructing an intelligent knowledge graph for examining construction engineering according to claim 1, wherein in S2, a first building component is selected from a set of building components, and a second building component is selected from the rest of the building components to form a pair of components to be judged, and a first spatial range representation corresponding to the first building component and a second spatial range representation corresponding to the second building component are obtained; The minimum Euclidean distance between the first spatial range representation and the second spatial range representation is calculated based on the first spatial range representation and the second spatial range representation, space intersection detection is performed based on the first spatial range representation and the second spatial range representation, whether an intersection region exists between the first spatial range representation and the second spatial range representation is judged, inclusion detection is performed based on the first spatial range representation and the second spatial range representation, and whether one spatial range representation is completely located inside the other spatial range representation is judged.
5. The method for constructing an intelligent knowledge graph for examining a building engineering according to claim 4, wherein if the minimum euclidean distance is zero and no intersection area exists, the pair of components is determined to satisfy the adjacent relation, if the minimum euclidean distance is greater than zero and smaller than a preset adjacent distance threshold value and no intersection area exists, the pair of components is determined to satisfy the adjacent relation, and if the intersection area is a non-empty set, the pair of components is determined to satisfy the intersection relation; if all points on the first spatial range representation are inside the second spatial range representation and the volume represented by the first spatial range representation is less than the volume represented by the second spatial range, determining that the member pair satisfies the inclusion relationship and that the first building member is included by the second building member; If all points on the second spatial range representation are inside the first spatial range representation and the volume represented by the second spatial range representation is less than the volume represented by the first spatial range, determining that the member pair satisfies the inclusion relationship and that the second building member is included by the first building member; for a pair of members determined to satisfy at least one spatial relationship type of an adjacent relationship, an intersecting relationship, or a containing relationship, a relationship record is generated, the relationship record containing a first member identification of a first building member, a second member identification of a second building member, and the determined spatial relationship type, all relationship records are aggregated, and a first set of relationship pairs is generated.
6. The method for constructing an intelligent knowledge graph for building engineering inspection according to claim 1, wherein in the step S3, each rule in the rule set is traversed, each rule includes a condition part and a conclusion part, the condition part includes at least a first component type, a second component type, and a target spatial relationship type between the first component type and the second component type, the conclusion part includes a logical conclusion corresponding to the rule, and the first component type, the second component type, and the target spatial relationship type are extracted from the condition part of the rule traversed currently; A first component set conforming to a first component type is screened out from building components based on the first component type, a second component set conforming to a second component type is screened out from building components based on the second component type, each first component in the first component set is traversed, and each second component in the second component set is traversed, so that a component pair to be queried is formed.
7. The method for constructing an intelligent knowledge graph for examining building engineering according to claim 6, wherein for each member pair to be queried, querying whether a spatial topological relation record corresponding to the member pair exists in the first relation pair set, and if so, obtaining a logic conclusion from a conclusion part of a rule currently traversed, and generating a second relation pair between a first member and a second member of the member pair, wherein the second relation pair comprises a first member identifier, a second member identifier and a relation type determined by the logic conclusion, and collecting all generated second relation pairs to form a second relation pair set.
8. The method for constructing the intelligent knowledge graph for inspecting the building engineering according to claim 1, wherein a building component set is obtained, the building component set comprises a plurality of building components, each building component corresponds to a component identifier, the building component set is traversed, a graph node is created for each building component, the graph node comprises a node identifier and a node attribute set, the node identifier adopts the component identifier of the building component, and the node attribute set at least comprises the component type, the spatial position coordinate and the overall dimension parameter of the building component; Traversing a first relation pair set, wherein for each first relation pair in the first relation pair set, the first relation pair comprises a first component identifier, a second component identifier and a spatial relation type, a first graph edge is created by taking a graph node corresponding to the first component identifier as a starting point and a graph node corresponding to the second component identifier as an ending point, a relation type attribute is added for the first graph edge, and the relation type attribute takes a value as the spatial relation type.
9. The method for constructing an intelligent knowledge graph for building engineering inspection according to claim 8, wherein a second relation pair set is traversed, for each second relation pair in the second relation pair set, the second relation pair comprises a first component identifier, a second component identifier and a relation type determined by a logic conclusion, a second graph edge is created by taking a graph node corresponding to the first component identifier as a starting point and a graph node corresponding to the second component identifier as an ending point, and a relation type attribute is added for the second graph edge, the relation type attribute is a relation type determined by the logic conclusion, and all the created graph nodes, the first graph edge and the second graph edge are collected to obtain the knowledge graph for building engineering inspection.

Description

Intelligent knowledge graph construction method for construction engineering examination Technical Field The invention relates to the technical field of construction of building engineering examination knowledge maps, in particular to an intelligent knowledge map construction method for building engineering examination. Background The construction engineering inspection is a key link for ensuring that design achievements meet national specifications, industry standards and safety requirements, and directly relates to engineering quality, construction safety and reliability of later operation and maintenance, the traditional construction engineering inspection mainly depends on a manual mode, and an inspector checks component arrangement and size relation in a construction information model or a two-dimensional drawing one by one according to self professional knowledge and paper or electronic version specification regulations, and the defects of low efficiency, inconsistent standard execution, easy omission, hidden problem and the like are gradually exposed in a manual inspection mode along with the increasing of the construction scale, the rapid increase and frequent updating of the specification regulations. In order to improve the inspection efficiency, an automatic inspection technology based on a building information model appears in recent years, the technology firstly extracts component attribute data from the building information model, then converts a specification rule into a logic expression which can be processed by a computer, and finally carries out traversal matching on the model data by compiling a specific query script or a rule engine to output a component list violating the specification, however, the conversion process of the specification rule of the method is highly dependent on manual coding, requires a professional to manually compile the specification described by natural language into a program code or a structured query statement, has low conversion efficiency and is easy to introduce human errors, each specification is usually independently processed, the inspection result is output in a discrete violation list form, and a correlation network between components and between the specifications cannot be established, so that the inspection result is difficult to trace back and comprehensively analyze. Disclosure of Invention Aiming at the defects of the prior art, the invention provides an intelligent knowledge graph construction method for building engineering examination. The intelligent knowledge graph construction method for the construction engineering examination comprises the following steps: S1, acquiring geometric feature data and attribute feature data of building components in a building information model, acquiring a building engineering examination specification, and analyzing the specification into a rule set, wherein the rule set comprises space constraint conditions; S2, determining a spatial topological relation among building components based on the geometric characteristic data, and generating a first relation pair set; S3, matching the space constraint condition in the rule set with the geometric feature data, identifying a component combination meeting the space constraint condition, and generating a second relation pair set according to a matching result; s4, building an architectural engineering examination knowledge graph by taking the building component as a node and taking the relation pair in the first relation pair set and the second relation pair set as an edge. Preferably, in the step S1, component type, space position coordinate and outline dimension parameter of each building component are extracted from a building information model, space range representation of each building component is generated based on the space position coordinate and outline dimension parameter, and standard treatises are analyzed into rule sets, specifically, semantic analysis is carried out on the standard treatises, space relation description in the standard treatises is identified, the space relation description is mapped into preset space relation types, rule items which take the space relation types as conditions are constructed, and the rule items form the rule sets; carrying out semantic analysis on the specification strip, and identifying the spatial relationship description in the specification strip, wherein the method specifically comprises the following steps: Obtaining text data of a standard article, performing word segmentation on the text data, segmenting continuous natural language text into word sequences consisting of independent words, performing part-of-speech tagging on each word in the word sequences, and determining part-of-speech categories corresponding to each word, wherein the part-of-speech categories comprise nouns, verbs, adjectives, prepositions and conjunctions; Acquiring a preset spatial relationship keyword library, wherein a plurality of ke