The Research On Reasoning Mechanism Of Early-warning Expert System For Maintenance Errors In Civil Aviation

With the development of aviation technology, aviation safety and reliability have been continuously improved, but the problems of maintenance errors in the aviation maintenance are appeared constantly, and the number of unsafe events relative to maintenance errors shows a rising trend. The use of technology reasoning in Expert System has important practical significance, which warns and controls early maintenance errors.This thesis is based on Expert System, uncertainty reasoning and case-based reasoning, and other related theoretical researches, and combines the NSFC~the study on Early-warning Mechanism and Methods of the Civil Aviation Calamity. With the maintenance errors in aviation for the background, the thesis analyses the characteristic of maintenance errors in civil aviation, proposes the union reasoning mechanism of both rule-based reasoning and case-based reasoning, which is applied to Early-warning Expert System for Maintenance Errors in Civil Aviation. The reasoning mechanism makes use of their respective reasoning advantages and forms a complementary advantage, and then enhances the flexibility of the system and its integrated reasoning abilities, so that the system could quickly and accurately predict and control the occurrence of maintenance errors, At the same time, by controlling early maintenance errors, further prevents and detects signs of insecurity incidents, and takes promptly appropriate Countermeasures, in order to nip in the bud, and eliminates hidden invalidation of the aviation maintenance errors system, and cuts off the chain of insecurity incidents, and then reduces maintenance errors rate, so ensures flight safety.The major work of this thesis is constructing the part of reasoning mechanism of early-warning expert system for maintenance errors in civil aviation (EESMECA), and it includes the following:1) Analyzing the current situation and existing problems about the filed of maintenance errors by investigation and collecting a large number of materials about aviation maintenance errors at home and abroad, and proposing the necessity and significance about the application of reasoning technology of expert system in maintenance error field.2) Researching the reasoning technology of Expert System, and analyzing the characteristic of aviation maintenance errors, and proposing a reasoning model of rule-based uncertainty reasoning and case-based reasoning in EESMECA, and then respectively designing and realizing these two kinds of reasoning technology in the application of EESMECA.3) Collecting a large number of material about aviation maintenance errors at home and abroad, and analyzing and sorting, in order to summarize the rules of maintenance errors and causations; Based on this, applying the reasoning technology of rule-based uncertainty reasoning to the EESMECA, and extending rule engine—jess, increasing the measurement methods of credibility, developing and designing knowledge base and rule-based uncertainty inference engine based on Jess.4) Exploring the application of case-based reasoning technology in the field of maintenance errors. Analyzing and summarizing these unsafe events which are caused by maintenance errors in the past 20 years at home and abroad, and then proposing the case description model, so focusing on the researches of case representation and case retrieval algorithms.5) According to the reasoning model of EESMECA, Realizing both of the rule-based uncertainty reasoning and case-based reasoning in this system, and establishing the reasoning platform about early-warning for maintenance errors in civil aviation, and then effectively improving the early-warning intelligence and accuracy.The characteristic of this thesis is that case-based reasoning technology is applied to the filed of maintenance errors, in order to predict and control early the unsafe events, and proposed the case description model. At present, case-based reasoning technology has been applied to many domestic filed, but there are not related researches in the civil aviation maintenance errors filed. The thesis has achieved some fruits in the research on the filed, but a lot of work needs to study further. The degree thesis is supported by the project -the study on Early-warning Mechanism and Methods of the Civil Aviation Calamity supported by National Natural Science Foundation of China (70371034).