Research On Maintenance Decision-making For Civil Aircraft System Based On Deep Learning

As a typical complex system,civil aircraft system has multiple sensor data,complex structure of components,and high security requirements.Aircraft condition monitoring system can collect a lot of information in the process of operation.How to excavate the features of these state monitoring information,enhance the accuracy of state characterization of civil aircraft system,and make effective maintenance decision,are hot and core issues in the field now.Deep learning has achieved good results in large data processing,complexity analysis,and high precision prediction.According to the characteristics of civil aircraft system in the process of operation,this paper establishes framework of maintenance decision for civil aircraft system based on deep learning.The following contents are included:Firstly,this paper discusses the basic knowledge of maintenance decision and related background for civil aircraft system,summarizes the related research background and application of deep learning algorithm.By summarizing the characteristics of related models,a framework of maintenance decision-making based on deep learning is constructed,and the maintenance decisions are divided into three parts: maintenance level decision-making,fault prediction and maintenance perception.Secondly,this paper applies deep learning to maintenance level decision-making for civil aircraft system,introduces the composition of the model and optimization procedure explicitly.Through data collection,model training,result analysis and other aspects,this paper build maintenance level decision-making model for civil aircraft system based on deep learning.Again,this paper builds fault prediction model for the civil aircraft system based on deep learning.Through the combination of deep belief network and the structural equation model,we can get fault prediction results,and improve the anti-interference and accuracy of the fault prediction.Finally,the application process of the model is introduced through an example analysis.Finally,this paper studies the maintenance perception for civil aircraft based on deep learning.We use deep belief network to analysis the performance degradation for civil aircraft system parameters,and then apply the time series prediction to get performance degradation trend.Therefore,we can perceive the state of the system and adjust the maintenance time in time.Based on deep learning algorithm model,this paper analysis the maintenance decision-making comprehensively for civil aircraft system from three aspects such as: maintenance level decision-making system,fault prediction and maintenance perception,and provides a theoretical and practical model with maintenance decision-making of civil aircraft system.