Edge Knowledge Graph Fusion for Industrial Data Model Mismatch
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The integration of industrial data from field devices and third-party commercial software is hindered by differing data models, making it difficult to perform data fusion effectively.
Innovation Solution
An industrial data processing method and apparatus for an edge device that constructs a knowledge graph with ontology to unify data structures, allowing real-time and external data to be fused and reasoned, and matched with server ontologies to enhance data capacity and flexibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If data fusion is performed using predefined policies for relational databases, then data integration is simplified, but it cannot be applied to free databases constructed by industrial field devices and third-party commercial software historians
Solution Approach 1:
The patent introduces an intermediary layer (data fusion layer) that includes ontology construction and data reasoning components. This intermediary translates data from different sources (industrial field devices and third-party commercial software) into a unified data model, enabling compatibility between diverse data sources without requiring changes to the original systems.
Solution Approach 2:
The patent transforms the data representation parameters by converting raw data from various sources into ontology-based structured data. This parameter transformation enables different data models to be represented in a unified framework, allowing predefined fusion policies to work across diverse data sources.
2Adaptability or versatility
If different data models are used by field devices and third-party software historians, then each system maintains its own data structure flexibility, but data fusion becomes extremely difficult
Solution Approach 1:
The patent segments the data fusion process into distinct functional layers: data collection, ontology construction, data reasoning, and data fusion. This segmentation allows each layer to handle specific tasks independently, reducing overall system complexity while maintaining support for multiple data models.
Solution Approach 2:
The patent creates a universal data fusion framework that can handle multiple data models through a common ontology structure. The unified data model serves as a multi-functional interface that accommodates various source systems while providing consistent data fusion capabilities across all inputs.
3Reliability
If complex conversions are performed between different data structures, then data fusion completeness improves, but processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary action by constructing ontologies and establishing data relationships before actual data fusion occurs. This pre-processing creates a ready-made framework that accelerates the fusion process, as data can be directly mapped to the pre-defined ontology structure without requiring complex real-time conversions.
Solution Approach 2:
The patent replaces mechanical data conversion processes with knowledge-based reasoning. Instead of performing complex structural transformations, the system uses ontology-based mapping and logical reasoning to achieve data fusion, significantly reducing computational overhead while maintaining fusion completeness.
Data Source
AI summary
Various embodiments of the teachings herein include an industrial data processing method for an edge device. An example includes: collecting real-time data from a field device; constructing a first knowledge graph stored in the edge device, wherein the first knowledge graph comprises first ontology; and creating an instance by using the first ontology of the first knowledge graph and according to the real-time data, to fuse the real-time data with the first knowledge graph.