Semantic Query For Engineering Knowledge Graph Based On Rule Inference

In the field of engineering technology,there are a large amount of unstructured knowledge represented by interrelated facts and experiences.This kind of knowledge contains a large amount of semantics,and it is very difficult to represent it using traditional knowledge representation methods.The knowledge graph is a graph data structure in which triples formed by fact nodes and directed edges are connected to each other,and the ontology provides semantic information support for these fact nodes and directed edges.The knowledge graph provides a new solution to formalizing representation of knowledge that is difficult to structure in the engineering field.One of the most basic requirements of the knowledge graph application is the query of semantic information.The semantic information query involves the reasoning of knowledge.Research and development on the semantic query for the constructed knowledge graph can be of great significance.Ontology based on description logic has strong knowledge description abilities,but lacks representation of causal relations.Causal relation rule reasoning is an important part of knowledge logic reasoning.Based on the existing semantic Web ontology technology and semantic rules research,this paper proposes a semantic query method for engineering knowledge graphs combining rule-based reasoning and ontological representation.A modeling and semantic query system for engineering knowledge graphs are developed by programming with Visual C++.The query system has made new explorations for the knowledge representation of engineering information.The mainly complete works are as follows:Based on object-oriented technology,an object-oriented model to represent ontology-based knowledge graphs is designed.Based on the presented model,the OWL ontology classes,object properties and data properties,instances and rules constructed by the current mainstream ontology modeling tool Protégé are pharsed into object-oriented models,which lays a foundation and provides an interface for the subsequent efficient engineering-oriented knowledge graph construction and semantic query.At the same time,it is compatible with other mainstream ontology knowledge modeling systems.Aiming at the characteristics of engineering information and knowledge representation,this paper proposes a hybrid open world assumption and closed world assumption,extends the widely used SWRL and SQWRL rule definitions,and proposes and implements a semantic query method similar to SQWRL.As a result,the presented semantic query can meet the requirements of engineering knowledge representation.According to the characteristics and requirements of semantic query,a reverse chain inference algorithm for SQWRL semantic query conditions is proposed.Through reverse satisfiability reasoning for knowledge graph instances,the semantic query of knowledge graphs is realized.To establish a unified form of inference rules,all implicit semantics in the ontology are transformed into explicit rules,which are combined with the rules explicitly defined by SWRL to form a unified reasoning rule set.Finally,the developed semantic query method is applied to the automatic machining feature recognition.The presented method is verified by an example to prove the feasibility of the method.