Research On Intelligent Retrospective Technology For Quality Defects Of Flight Skin Riveting

With the development of social economy,airplanes have become a common means of transportation.Airplane skin riveting is an important link in the process of aircraft assembly.The quality of aircraft riveting determines the life of aircraft and people’s safety.When riveting defects occur,it is very important to find and handle them in time.Therefore,it is very important to trace back the quality defects of aircraft skin riveting.Nowadays,in the actual detection,most of the aircraft skin riveting quality defect detection uses manual detection method,which makes it difficult to trace the defect and greatly increases the probability of the aircraft skin riveting quality abnormality.In order to solve the above problems,this paper presents a quality tracing method based on rough set theory and Bayesian network(RST-BN)to trace the quality defects of aircraft skin riveting.The main research work of this paper includes the following parts:Firstly,aiming at the problem of how to reasonably select multiple important conditional attributes in the process of rough set algorithm reduction,an information granularity method is introduced to optimize the selection process of conditional attributes.Aiming at the problem that the influence of conditional attributes on decision attributes is not considered in the process of rough set importance calculation,the relationship between conditional attributes is added in the importance calculation process to make the obtained attributes more reasonable.On this basis,an attribute reduction algorithm based on information granularity and improved attribute importance is proposed.Comparative experiments are carried out on the UCI dataset,and through experimental verification,the proposed method can obtain higher reduction rate and classification accuracy.Secondly,in order to solve the complex and computational problem of Bayesian network,the attribute reduction function of rough set theory is used to reduce the complexity of Bayesian network,and Bayesian network is used to build a method model for retrospective analysis of aircraft skin riveting quality defects.Due to the possibility of uncertainty in event recognition,triangular fuzzy numbers are introduced to describe the defect probabilities in Bayesian nodes.By comparing the accuracy of retrospective analysis with other artificial intelligence methods,it is proved that this model is better than other methods in the case of incomplete data information of aircraft skin riveting quality characteristics.Finally,summarize the method of this article,and designs and implements a set of aircraft skin riveting quality defect retrospective support system based on the actual status and requirements of aircraft skin riveting quality defect retrospective.