CN-121981234-A - Dynamic construction method of elevator safety risk knowledge graph integrating space-time multifactor

CN121981234ACN 121981234 ACN121981234 ACN 121981234ACN-121981234-A

Abstract

The invention discloses a dynamic construction method and a system for a safety risk knowledge graph of an elevator by fusing space-time multifactor, wherein the method comprises the following steps: and carrying out space-time reference unification processing on the multi-source heterogeneous space-time multi-factor data to finish attribute normalization reconstruction, generating a standardized data set and realizing data dimension alignment and integration. And extracting the elevator safety field entity based on the standardized data set, extracting the space-time and field attribute, and constructing the space-time and field association relationship among the entities to form an entity-relationship system integrating the space-time multiple factors. Based on the system, a heterogeneous information network is built, an inference rule base is embedded, a knowledge graph ontology framework is built, concept levels, attribute constraints and semantic association are defined, and basic version elevator safety risk knowledge graph construction is completed. And establishing a space-time dynamic updating mechanism by relying on the basic map, and iteratively optimizing entity attributes, association relations and reasoning rules according to real-time data to realize dynamic iterative construction of the knowledge map.

Inventors

HUANG QING
CHEN ZHENG
ZHANG SONGCHENG
LEI YANG
ZHANG RUI
RUAN YIHUI
ZHANG FUSHENG
GE YANG
CHEN WEIBI
DONG HAOMING

Assignees

武汉市特种设备监督检验所
苏州工学院
江苏省特种设备安全监督检验研究院
中国地质大学（武汉）

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (10)

1. A dynamic construction method of an elevator safety risk knowledge graph integrating space-time multifactor is characterized by comprising the following steps: S1, performing space-time reference unification on multi-source heterogeneous space-time multi-factor data, mapping various data to the same geographic space coordinate system and a time dimension scale, synchronously completing normalized reconstruction of data attribute dimensions, generating a standardized data set associated with space-time attributes, and completing dimension alignment and association integration of the space-time multi-factor data; S2, based on a standardized data set, identifying and extracting full-type entities related to space-time multifactorial in the elevator safety field, synchronously extracting space-time attribute features and field attribute features of each entity, defining and constructing space-time association relations and field association relations among the entities according to association rules and field management logic of the space-time multifactorial and the elevator safety, and forming an entity-relation system fused with the space-time multifactorial features; S3, building an elevator safety field heterogeneous information network based on an entity-relation system fused with space-time multi-factor characteristics, embedding a space-time multi-factor association reasoning rule base and building a knowledge graph body framework, defining concept levels of the body, attribute constraint and semantic association relations, completing mapping fusion of space-time multi-factor and elevator safety field knowledge, and completing building of a basic version elevator safety risk knowledge graph; S4, building a space-time dimension dynamic updating mechanism by relying on a basic version of elevator safety risk knowledge graph, and iteratively optimizing entity attributes, association relations among entities and reasoning rules according to space-time multi-factor data acquired in real time to complete dynamic iteration construction of the elevator safety risk knowledge graph.
2. The dynamic construction method of the elevator safety risk knowledge graph integrating space-time multifactorial according to claim 1, wherein the space-time reference unification processing in the step S1 is characterized in that: for the original geospatial coordinates of various space-time multi-factor data, a space affine transformation formula is adopted to realize the mapping conversion to a preset urban geospatial coordinate system, and the formula is that Wherein To unify two-dimensional coordinate values under the urban geospatial coordinate system after conversion, For the horizontal coordinate value of the drawing, For the vertical coordinate value, Is a 3X 3 space affine transformation matrix, is obtained by calculating the characteristic point coordinate pairs of the urban geographic space reference, The two-dimensional coordinate values of the geographic space originally acquired for the space-time multi-factor data, For the original abscissa value of the value, Is the original ordinate value; Aiming at the original time stamps of various space-time multi-factor data, a time axis linear mapping formula is adopted to realize the calibration to a preset standard time dimension scale, and the formula is that Wherein For the timestamp value under the calibrated standard time dimension scale, Is the reference point in time of the standard time dimension scale, For the time stamp values of the original acquisition of the spatiotemporal multi-factor data, Is the reference time point of the original time axis, Is the time metering interval of a standard time dimension scale, A time metering interval which is an original time axis; and after the geographic space and the time dimension reference are unified, performing reconstruction operation on the data attribute dimension by adopting an extremum normalization formula.
3. The method for dynamically constructing the elevator safety risk knowledge graph with the integration of space-time multifactorial functions according to claim 1, wherein the construction process of the space-time association relationship and the domain association relationship among the entities in the step S2 is as follows: First extracting space-time attribute feature set of each entity And domain attribute feature set Wherein The number of the entity is given to the person, As a feature of the geospatial coordinates of an entity, Is a time dimension feature of an entity, is based on a geospatial coordinate feature Calculating the spatial association degree between any two entities , Is an entity And entity Based on time dimension characteristics Calculating the time association degree between any two entities Obtaining the space-time association relation between the entities through the coupling operation of the space association degree and the time association degree Realizing the quantitative construction of the space-time association relation between the entities; And then based on the domain attribute feature set of each entity Calculating the similarity of domain attributes between any two entities The qualitative construction of the domain association relationship among the entities is realized through the threshold value judgment of the attribute similarity, wherein the threshold value is a preset elevator safety domain attribute association judgment threshold value; Finally, the space-time association relation between the entities is fused Association with a Domain And forming an entity-relation system with space-time quantization characteristics and field qualitative characteristics, and completing the global construction of the association relation among the entities fusing the space-time multi-factor characteristics.
4. The method for dynamically constructing the knowledge graph of the elevator safety risk fused with the space-time multifactor according to claim 1, wherein the construction process of the knowledge graph body frame in the step S3 is as follows: The method comprises the steps of firstly building an elevator safety field heterogeneous information network by using an entity-relation system fused with space-time multi-factor characteristics as a data base, taking full-type entities as network nodes, and taking space-time association relations and field association relations among the entities as network edges to form a heterogeneous information network fused with space-time multi-factor characteristics by both nodes and edges; Embedding a space-time multi-factor association reasoning rule base, developing ontology framework construction based on ontology concept level, attribute constraint and semantic association relation, and mapping entity nodes in the heterogeneous information network to the ontology concept level In the corresponding concept class of the entity, realizing the hierarchical binding of the entity and the ontology concept; Configuring preset attribute constraint for each concept class The time-space attribute and the field attribute of each concept class meet the requirements of the value taking domain and the coupling constraint; mapping network edges in heterogeneous information networks into semantic association relations among ontology concept classes Mapping of association relation between entities and semantic association of ontology is achieved, wherein 、 Numbering the ontology concept classes; After mapping is completed, a space-time multi-factor association reasoning rule base is embedded into an ontology framework semantic association layer to form a unified logic system of rule and semantic association, and the elevator safety domain knowledge, an ontology concept level, attribute constraint and semantic association relationship are subjected to global fusion to realize the integrated binding of space-time multi-factor characteristics and elevator safety domain knowledge, so that a knowledge map ontology framework is constructed.
5. The method for dynamically constructing a space-time multifactor-fused elevator safety risk knowledge graph as claimed in claim 4, wherein the ontology concept hierarchy in S3 is a hierarchy formed by performing conceptual hierarchy division on all types of entities in the elevator safety domain based on a space-time multifactor-fused entity-relationship system, and is recorded as Wherein Is a global set of ontology concept levels, For the number of conceptual layer levels, For a preset maximum number of levels of the elevator safety domain concept, The system is a top-level core concept class, and corresponds to a full-type entity of equipment class, space-time environment class, responsibility main body class, management rescue class and fault risk class; To the point of As an intermediate subclass, the core concept class of each vertex is characterized by dimension according to time-space attributes Dimension of attribute of field The sub-class of concepts formed by the progressive subdivision, For the bottom instance class, the layering mapping of the concept of the elevator safety field and the space-time multi-factor entity is realized for the specific entity instance corresponding to each intermediate subclass.
6. The method for dynamically constructing a space-time multifactor-fused elevator safety risk knowledge graph according to claim 4, wherein the attribute constraint in S3 is a global constraint set of space-time attribute constraint and domain attribute constraint set for each concept class of the ontology, and is recorded as Wherein For space-time attribute constraint, satisfy , As the geospatial coordinate attribute value of the concept class, Taking a value domain for preset urban geographic space coordinates, As the time dimension attribute value of the concept class, Taking a value domain for a preset standard time dimension; For the constraint of the field attribute, satisfy , As the domain-attribute feature value of the concept class, For a preset minimum value of the domain attribute in the same dimension, Is a preset maximum value of the domain attribute under the same dimension, and the space-time attribute constraint and the domain attribute constraint meet the coupling constraint relation And realizing consistency constraint of concept class attributes and entity space-time multi-factor characteristics and domain characteristics.
7. The method for dynamically constructing a space-time multifactor-fused elevator safety risk knowledge graph as set forth in claim 4, wherein the semantic association in S3 is based on a space-time association between entities Association with a Domain The constructed semantic association set between the ontology concept classes is recorded as Wherein 、 The ontology concept class is numbered, Is the first 、 The space-time association degree of the corresponding entity of each concept class is obtained by multiplying the reciprocal of the Euclidean distance of the geographic space of the entity by the reciprocal of the time difference, Is the first 、 The domain attribute similarity of the entity corresponding to each concept class is obtained by the ratio of intersection and union of the attribute feature sets of the entity domain, and semantic fusion of space-time association and domain association is realized through intersection operation, so that the semantic association relationship among the ontology concept classes completely characterizes the association rule of space-time multifactor and elevator safety domain.
8. The method for dynamically constructing the elevator safety risk knowledge graph with the fused space-time multifactorial as claimed in claim 4, wherein the process of embedding the space-time multifactorial association reasoning rule base in the S3 is as follows: Through a rule mapping formula Completing the bi-directional mapping of inference rules and semantic associations, wherein For space-time multi-factor correlation reasoning rule set, bidirectional mapping makes the reasoning rule form semantic support for the ontology framework, and finally through the field knowledge fusion formula Completing mapping fusion of space-time multifactor and elevator safety domain knowledge, wherein For the fused global knowledge set, And (3) realizing the global binding of the domain knowledge and the ontology concept level and semantic association relation for the elevator safety domain knowledge set through union operation.
9. A dynamic construction system for an elevator safety risk knowledge graph integrating space-time multifactor is characterized by comprising the following components: the data unification and standardization unit is used for carrying out space-time reference unification treatment on the multi-source heterogeneous space-time multi-factor data, mapping various data to the same geographic space coordinate system and a time dimension scale, synchronously completing the normalization and the reconstruction of data attribute dimensions, generating a standardized data set associated with space-time attributes, and completing the dimension alignment and association integration of the space-time multi-factor data; The entity and relation construction unit is used for identifying and extracting full-type entities related to space-time multifactorial in the elevator safety field based on the standardized data set, synchronously extracting space-time attribute characteristics and field attribute characteristics of each entity, defining and constructing space-time association relations and field association relations among the entities according to association rules of the space-time multifactorial and the elevator safety and field management logic, and forming an entity-relation system fused with the space-time multifactorial characteristics; The basic map construction unit is used for constructing an elevator safety field heterogeneous information network based on an entity-relation system fusing space-time multi-factor characteristics, embedding a space-time multi-factor association reasoning rule base and constructing a knowledge map body framework, and simultaneously defining a body concept level, attribute constraint and semantic association relation to finish mapping fusion of space-time multi-factor and elevator safety field knowledge and finish basic version elevator safety risk knowledge map construction; And the map dynamic iteration unit is used for building a space-time dimension dynamic updating mechanism by relying on the basic version elevator safety risk knowledge map, and completing dynamic iteration construction of the elevator safety risk knowledge map according to real-time acquired space-time multi-factor data iteration optimization entity attributes, association relations among entities and reasoning rules.
10. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program is executed by a processor for a method for dynamically constructing a spatiotemporal multifactor-fused knowledge-graph of the safety risk of an elevator according to any one of claims 1 to 8.

Description

Dynamic construction method of elevator safety risk knowledge graph integrating space-time multifactor Technical Field The invention belongs to the technical field of elevator safety monitoring, and particularly relates to a dynamic construction method and a system for an elevator safety risk knowledge graph integrating space-time multifactor. Background The current elevator conservation volume continuously grows, the elevator operation environment is complex and various, factors influencing elevator safety present multidimensional and spatiotemporal characteristics, cover a plurality of aspects such as the elevator equipment self running state, the urban meteorological power and other external environments, the whole flow safety management and the like, various influence factors are mutually associated and dynamically changed, and great challenges are brought to elevator safety risk management and control. The existing elevator safety management is dependent on traditional modes such as manual inspection, fault repair and the like, and lacks of effective integration and utilization of multi-source heterogeneous space-time multi-factor data. The elevator related data from different sources are stored in different platforms in a scattered manner, obvious isomerism exists in the geospatial coordinates, the time dimension scale and the attribute expression, effective fusion and association analysis of the data are difficult to achieve, and the elevator safety risk identification hysteresis and the control pertinence are insufficient. The existing elevator safety-related knowledge graph construction method is mainly focused on the application of single-dimension data, is not fully integrated with space-time multi-factor characteristics, and cannot completely characterize the association rule and dynamic change trend of elevator safety risks in space-time dimensions. Meanwhile, the existing knowledge graph is in a static construction mode, and cannot be dynamically updated according to space-time multi-factor data acquired in real time, so that the knowledge graph content is disjointed from the actual running state of the elevator, and the actual requirement of dynamic management and control of the safety risk of the elevator is difficult to meet. Therefore, how to integrate multisource heterogeneous space-time multifactor data constructs a knowledge graph capable of dynamically responding to data changes and accurately representing the association relation of elevator safety risks, achieves accurate identification, early warning and dynamic management and control of elevator safety risks, becomes a key problem to be solved urgently in the current elevator safety management field, and has important practical significance for improving the elevator safety operation level and guaranteeing public safety. Disclosure of Invention The invention aims to solve the problems that multisource spatiotemporal data are difficult to fuse, risk identification is delayed, the existing knowledge graph is static and lacks space-time correlation and the like in elevator safety management, and realizes unified data, accurate relation characterization and real-time iterative updating by constructing a dynamic risk knowledge graph fusing space-time multifactor, thereby providing reliable technical support for elevator safety risk dynamic management and early warning. In order to address the above-mentioned drawbacks or improvements of the prior art, as a first aspect of the present invention, the present invention provides a method for dynamically constructing a space-time multifactor-fused elevator security risk knowledge graph, including: S1, performing space-time reference unification on multi-source heterogeneous space-time multi-factor data, mapping various data to the same geographic space coordinate system and a time dimension scale, synchronously completing normalized reconstruction of data attribute dimensions, generating a standardized data set associated with space-time attributes, and completing dimension alignment and association integration of the space-time multi-factor data; S2, based on a standardized data set, identifying and extracting full-type entities related to space-time multifactorial in the elevator safety field, synchronously extracting space-time attribute features and field attribute features of each entity, defining and constructing space-time association relations and field association relations among the entities according to association rules and field management logic of the space-time multifactorial and the elevator safety, and forming an entity-relation system fused with the space-time multifactorial features; S3, building an elevator safety field heterogeneous information network based on an entity-relation system fused with space-time multi-factor characteristics, embedding a space-time multi-factor association reasoning rule base and building a knowledge graph body framework, defining concept lev