CN-121981241-A - Context-task-knowledge association model coding method, device, computer equipment and storage medium

CN121981241ACN 121981241 ACN121981241 ACN 121981241ACN-121981241-A

Abstract

The embodiment of the application provides a scene-task-knowledge association model coding method, a device, computer equipment and a storage medium, wherein the method comprises the steps of determining the scene, event, object, task, knowledge system, knowledge granularity and the value range and definition of an expandable semantic mark of a target gas hidden danger scene; the method comprises the steps of thinning scenes, tasks and knowledge into a plurality of minimum units, setting standardized field definition, constraint conditions and dynamic updating mechanisms for each minimum unit, constructing scene atoms, task atoms and knowledge atoms, respectively packaging each atom into uniform shell codes to form uniform representation of each atom, respectively formulating scene triggering task rules, task binding knowledge rules, knowledge back propagation task rules and cross-layer consistency check rules for each atom of the uniform representation, so as to ensure that atomic units of different levels are kept consistent logically, and dynamically adjusting task execution sequences and knowledge calling ranges according to external situations.

Inventors

ZHAO XINGHAO
WANG WAN
XU FENGJIAO
HUANG SHUAI
QIN TINGXIN
YANG YUEXIANG
TU XINYU

Assignees

中国标准化研究院
中国矿业大学(北京)

Dates

Publication Date: 20260505
Application Date: 20260104

Claims (10)

1. A scene-task-knowledge association model coding method, comprising: determining the value range and definition of a scene, an event, an object, a task, a knowledge system, knowledge granularity and an extension semantic mark of a target gas hidden danger scene; Based on the scene, event, object, task, knowledge system and the value range and definition of knowledge granularity of the target gas hidden danger scene, the scene, task and knowledge are refined into a plurality of minimum units, and standardized field definition, constraint conditions and dynamic updating mechanism are set for each minimum unit to construct scene atoms, task atoms and knowledge atoms which can be directly called; Packaging each scene atom, task atom and knowledge atom into a unified shell code respectively, and forming a consistent representation of each scene atom, task atom and knowledge atom by adopting space-time-event-object three-dimensional layering and extension semantic marks; And respectively making a scene trigger task rule, a task binding knowledge rule, a knowledge back propagation task rule and a cross-layer consistency check rule for each scene atom, task atom and knowledge atom which are consistently represented so as to ensure that atomic units of different levels are kept consistent logically, thereby dynamically adjusting the task execution sequence and the knowledge calling range according to external situations.
2. The method of claim 1, wherein determining the context, event, object, task, knowledge system, knowledge granularity, and value range and definition of the expandable semantic flag for the target gas hidden danger scenario comprises: establishing a scene element table according to the spatial characterization of the target gas hidden danger scene and the observable scene elements; Constructing an object set and an event set; Splitting a task into three dimensions of a main body, a stage and an action to generate a task ternary system; Dividing a knowledge system and knowledge granularity for the existing normalization file; the design can develop semantic marks.
3. The method of claim 1, wherein after the scenario, event, object, task, knowledge system, knowledge granularity range and definition based on the target gas hidden danger scenario, the scenario, task and knowledge are refined into a plurality of minimum units, and standardized field definitions, constraint conditions and dynamic update mechanisms are set for each minimum unit, and directly callable scenario atoms, task atoms and knowledge atoms are constructed, the method further comprises: The method comprises the steps of restraining relations among scenario atoms, task atoms and knowledge atoms, wherein each scenario atom is mapped to at least one task atom, each task atom is bound with at least one knowledge atom, the binding is forced constraint, otherwise, the task cannot be executed, and when knowledge clauses are updated, all task atoms referring to the knowledge atoms are rechecked.
4. The method of claim 1, wherein the encapsulating each of the scenario atoms, task atoms, and knowledge atoms into a unified shell code, respectively, using space-time-event-object three-dimensional layering and extensible semantic tags, forming a consistent representation of each of the scenario atoms, task atoms, and knowledge atoms, comprises: Defining a shell grammar; establishing a mapping table according to the segment bit in the shell grammar and the primary key value of the corresponding field in the scene atom, the task atom and the knowledge atom respectively; And assigning values to the space segment bit, the event segment bit, the object segment bit, the task ternary segment bit, the knowledge segment bit and the extension semantic segment bit of the external grammar based on the mapping table.
5. The method of claim 1, wherein after the formulating context trigger task rules, task binding knowledge rules, knowledge back propagation task rules, and cross-layer consistency check rules for each context atom, task atom, and knowledge atom of a consistent representation, respectively, the method further comprises: Continuously collecting scene atom data, and triggering scene grade upgrading according to the scene atom data; when the scene is updated, the task mapping rule is immediately re-executed, the inapplicable task is deleted, and a new task is inserted.
6. The method according to claim 1, wherein the respectively formulating a scenario triggering task rule, a task binding knowledge rule, a knowledge back propagation task rule for each scenario atom, task atom and knowledge atom of a consistent representation comprises: taking observed quantity and threshold value of Scene Atoms (SAT) as trigger conditions, wherein the trigger conditions are stored in a rule base in a Boolean logic expression; Each triggering condition corresponds to a group of task atoms; marking executing relation in the task group, automatically adjusting task executing sequence according to the risk level of the task, Each task atom binds at least one knowledge atom and stores the binding relationship in a task-knowledge comparison table; Before the task is executed, automatically checking whether the required knowledge clause is valid or not and whether version update exists or not; if the required knowledge item is invalid or a version update exists, the task is automatically marked as non-executable, When knowledge atoms are newly added or updated in the knowledge base, all task atoms referencing the knowledge atoms are automatically positioned; Carrying out dependency relationship tracking on the affected tasks, generating an affected task list, and sequencing according to the risk priority of the tasks; And re-checking the affected task through compliance, and re-writing the updated task and knowledge binding relation into an index library.
7. The method of claim 1, wherein formulating cross-layer consistency check rules for each scenario atom, task atom, and knowledge atom of a consistent representation comprises: checking whether each segment bit of the shell code is in fact match with scene atoms, task atoms and knowledge atoms of an atomic layer; If the current task code cannot be mapped to any task atom, determining that the current task code is an invalid code, Verifying whether a scene-task-knowledge mapping chain is complete; The round trip consistency test is carried out, namely, coding, analysis and recoding are carried out to obtain the same result; If the result of the coding and the recoding is inconsistent, generating an error report and triggering manual intervention.
8. A scene-task-knowledge association model coding apparatus, comprising: the determining module is used for determining the value range and definition of the scene, event, object, task, knowledge system, knowledge granularity and expandable semantic mark of the target gas hidden danger scene; The refinement module is used for refining the scenes, the tasks and the knowledge into a plurality of minimum units based on the values of the scenes, the events, the objects, the tasks, the knowledge system and the knowledge granularity of the target gas hidden danger scenes, setting standardized field definitions, constraint conditions and dynamic update mechanisms for each minimum unit, and constructing scene atoms, task atoms and knowledge atoms which can be directly called; the packaging module is used for respectively packaging each scene atom, task atom and knowledge atom into a unified shell code, and forming a consistent representation of each scene atom, task atom and knowledge atom by adopting space-time-event-object three-dimensional layering and expandable semantic marks; The setting module is used for setting a scene trigger task rule, a task binding knowledge rule, a knowledge back propagation task rule and a cross-layer consistency check rule for each scene atom, task atom and knowledge atom which are consistently represented respectively so as to ensure that atomic units of different levels are kept consistent logically, thereby dynamically adjusting the task execution sequence and the knowledge calling range according to external situations.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the scenario-task-knowledge correlation model coding method of any one of claims 1 to 7 when the computer program is executed.
10. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the scenario-task-knowledge-association model coding method according to any one of claims 1 to 7.

Description

Context-task-knowledge association model coding method, device, computer equipment and storage medium Technical Field The application relates to the technical field of gas hidden danger, in particular to a scene-task-knowledge association model coding method, a device, computer equipment and a storage medium. Background The existing gas hidden trouble investigation and emergency management technology mainly relies on manual inspection and portable detection equipment, a part of areas integrate pipe network asset information with a geographic information system GIS and a data acquisition and monitoring control system SCADA system to realize basic space positioning and operation monitoring, meanwhile, a safety plan and standard clauses are managed in a static text mode, a coding mode adopts a serial number, an administrative division number or a universal unique identification code UUID, and the technology can meet local management requirements, still stays in an equipment-data layer as a whole, and lacks scene-driven task linkage and knowledge calling. The existing scheme has obvious defects in application, namely fault exists between hidden danger discovery and task execution, a coding system is inextensible and unpalatable, machine calculation and manual use are difficult to achieve, knowledge updating is delayed, clauses are often invalid and cannot be known, compliance cannot be guaranteed, task granularity is too coarse, atomization task arrangement and dynamic pushing cannot be supported, and the problems severely restrict the refinement and quick response capability of gas hidden danger investigation. Under the background that the gas accident frequency, the urban high-consequence area management and control requirements are improved and the national policy promotes the gas safety and the standardization, the existing informatization system can not meet the requirements, and particularly under the digital twinning, emergency exercise and compliance audit scenes, the uniqueness, the traceability and the instantaneity of data are more required, so that a coding and modeling system capable of unifying semantics, dynamically updating and supporting machine reasoning and manual use is urgently needed. Disclosure of Invention The embodiment of the application provides a scene-task-knowledge association model coding method, a device, computer equipment and a storage medium. In a first aspect of the embodiment of the present application, a scenario-task-knowledge association model encoding method is provided, including: determining the value range and definition of a scene, an event, an object, a task, a knowledge system, knowledge granularity and an extension semantic mark of a target gas hidden danger scene; Based on the scene, event, object, task, knowledge system and the value range and definition of knowledge granularity of the target gas hidden danger scene, the scene, task and knowledge are refined into a plurality of minimum units, and standardized field definition, constraint conditions and dynamic updating mechanism are set for each minimum unit to construct scene atoms, task atoms and knowledge atoms which can be directly called; Packaging each scene atom, task atom and knowledge atom into a unified shell code respectively, and forming a consistent representation of each scene atom, task atom and knowledge atom by adopting space-time-event-object three-dimensional layering and extension semantic marks; And respectively making a scene trigger task rule, a task binding knowledge rule, a knowledge back propagation task rule and a cross-layer consistency check rule for each scene atom, task atom and knowledge atom which are consistently represented so as to ensure that atomic units of different levels are kept consistent logically, thereby dynamically adjusting the task execution sequence and the knowledge calling range according to external situations. In an optional embodiment of the present application, the determining the scenario, event, object, task, knowledge system, knowledge granularity, and value range and definition of the expandable semantic flag of the target gas hidden danger scenario includes: establishing a scene element table according to the spatial characterization of the target gas hidden danger scene and the observable scene elements; Constructing an object set and an event set; Splitting a task into three dimensions of a main body, a stage and an action to generate a task ternary system; Dividing a knowledge system and knowledge granularity for the existing normalization file; the design can develop semantic marks. In an optional embodiment of the present application, after the scenario, event, object, task, knowledge system, and knowledge granularity range and definition based on the target gas hidden danger scenario, the scenario, task, and knowledge are refined into a plurality of minimum units, and standardized field definitions, constraint conditions, and dynamic update mecha