Research On Fault Diagnosis System For Aircraft Harness Detection Based On Knowledge Graph

In order to ensure the reliability of the laying of aircraft harness,functional tests such as continuity test,leakage test are required for the entire harness network during the overall assembly phase of the aircraft.Each aircraft harness detection will produce tens of thousands of test results.At present,testers need to check the failed results of the continuity test one by one which is inefficient.Therefore,this paper takes an automatic detection system for aircraft harness as the research object,and uses its equipment data and test results to establish a Fault Diagnosis System for aircraft harness detection based on knowledge graph.The main research contents are as follows:Firstly,clarify the target of the Fault Diagnosis System for aircraft harness detection based on knowledge graph,and design the system initially from three aspects of the system’s functional architecture,software architecture,and technical architecture.Secondly,analyze the system structure,test principle,fault type,etc.of the automatic detection system for harness.Then organize related data of aircraft harness detection,such as equipment information,connectivity and test results,and store them in the database which provides data support for the construction of knowledge graph and fault inference model of aircraft harness detection.Thirdly,design knowledge graph of aircraft harness detection and determine the entity,relationship,and attributes of the knowledge graph according to actual situation of the automatic detection system for aircraft harness and store the knowledge graph in the Neo4 j database.Thus,the construction of knowledge graph is completed preliminarily.What’s more,this paper does a simple research on the visualization of knowledge graph.Then,establish fault inference model of Aircraft harness detection based on probabilistic soft logic with the knowledge graph.And writing inference rules according to the types and characteristics of Aircraft harness detection faults.Then,use the Most Probable Explanation to predicate probability value.By comparing with the existing continuity test results,the inference model used in this article can effectively resolve five types of faults: FAULT1(F1: 0.898),FAULT12(F1: 0.957),FAULT3(F1: 0.809),FAULT4(F1: 0.827)and FAULT5(F1: 0.790).Finally,the Fault Diagnosis System for Aircraft Harness detection Based on Knowledge Graph is completed,deploy and run on the server which is in the workplace.