US 12,736,935 B1 · Hierarchical temporal ontology for industrial facilities

Abstract

The present disclosure provides a computer-implemented method for constructing a hierarchical industrial knowledge ontology, comprising generating a multi-tier semantic graph model representing an industrial facility. The graph comprises an equipment components tier defining components that compose individual pieces of equipment, an equipment tier defining physical asset instances, an equipment-to-equipment connectivity tier defining physical connectivity relationships among the equipment, a flow tier defining directional transfer relationships for process or utility media, an equipment-to-function tier defining functional characteristics at the equipment or component level, a functional group tier defining nested functional subgraphs composed of multiple equipment and flow relationships, and an operating group tier defining outcome-driven subgraphs associated with specific operational objectives or modes. The ontology enables bidirectional semantic traversal across vertical, horizontal and semantic abstraction levels and relationship types to integrate context across tiers. The system maintains temporal layers preserving historical versions of each semantic tier, including transient state-conditioned groups, and supports reconstruction of prior plant states at specific timestamps for causal analysis of operational events.

Claims28

Dependent claims

A computer-implemented method for constructing and maintaining a hierarchical industrial knowledge ontology that unifies physical structure and operational behavior of an industrial facility, the method comprising:

generating a multi-tier graph data model representing the industrial facility, the model comprising:
a physical structure layer defining nodes and relationships corresponding to physical assets and their fixed connectivity;
a functional abstraction layer defining nodes and relationships that decouple physical assets from their functional characteristics and process-driven groupings;
and
an operational intent layer defining nodes and relationships corresponding to outcome-driven subgraphs associated with specific operational objectives or modes;
maintaining a temporal layer on the graph data model, the temporal layer being configured to preserve a historical version of the semantic relationships and connectivity within each layer, thereby enabling the reconstruction of past plant configurations at a specified timestamp for causal analysis;
and
implementing a hybrid ontology sustainment mechanism to govern modifications to the graph data model, the mechanism comprising the steps of:
generating candidate updates to the graph using a machine learning model analyzing real-time plant data;
submitting the candidate updates for review and validation by one or more subject matter experts via a review dashboard that presents the candidate updates in an organized format within the context of existing ontological structures across the multiple tiers of the graph data model, along with supporting evidence from the plant data;
and
merging validated updates into the graph data model only after approval and automatically logging the change event in the temporal layer.

9 more elements

A computer-implemented system for constructing and maintaining a hierarchical industrial knowledge ontology that unifies physical structure and operational behavior of an industrial facility, the system comprising:

one or more processors configured to generate a multi-tier graph data model representing the industrial facility, the model comprising:
a physical structure layer defining nodes and relationships corresponding to physical assets and their fixed connectivity;
a functional abstraction layer defining nodes and relationships that decouple physical assets from their functional characteristics and process-driven groupings;
and
an operational intent layer defining nodes and relationships corresponding to outcome-driven subgraphs associated with specific operational objectives or modes;
a temporal layer configured to preserve a historical version of the semantic relationships and connectivity within each layer, thereby enabling the reconstruction of past plant configurations at a specified timestamp for causal analysis;
and
a hybrid ontology sustainment mechanism configured to govern modifications to the graph data model, the mechanism comprising:
a machine learning subsystem configured to analyze plant data and generate candidate modifications to the graph data model;
a validation interface comprising a review dashboard configured to present the candidate modifications in an organized format within the context of existing ontological structures across the multiple tiers of the graph data model, along with supporting evidence from the plant data, to one or more subject matter experts for review;
an update management subsystem configured to integrate validated modifications into the graph data model while maintaining version control.

8 more elements

Dependencies are inferred from the claim text.

Hierarchical temporal ontology for industrial facilities

Life of this patent

Hierarchical temporal ontology for industrial facilities

Life of this patent

Abstract

Claims28