【作者】 刘洋；

【导师】 许渤；

【作者基本信息】 电子科技大学 ， 电子信息（专业学位）， 2024， 硕士

【摘要】 随着航空电子系统和机载总线网络的不断发展,故障诊断技术在保障飞行安全和提升系统可靠性方面的重要性日益凸显。传统的诊断方式主要依赖于数据采集卡和监控软件逐条查看消息块,这种方法需要专业人员且效率低下。本文基于机载总线网络应用场景对通信异常状态诊断的需求,进行了技术研究和实践,旨在将更加自动化、智能化的高效故障诊断方法应用到1553B总线、1394B总线、FC网络当中。本文主要工作可分为两部分:一方面,本文详细分析了1553B总线通信状态诊断软件的具体功能需求,构建了软件的整体功能架构。结合1553B总线的消息传输特性,设计并实现了一系列功能模块,这些模块能够满足诊断软件的具体需求。尤其是在数据处理模块,从多个角度解析、处理,得到能反映总线通信状态的参数,并且对异常状态的识别方法进行了特别设计和实现。通过仿真测试和实际场景测试,验证了该诊断软件不仅能准确反映1553B总线系统的通信状态,还能正确判定所有异常通信事件。相比传统方法,新的诊断软件在进行简单配置后,分析诊断过程可实现无人值守,并且通过图形化输出方式更加直观地呈现通信状态。另一方面,本文在研究1394B、FC网络的故障诊断方法时,考虑了具体的业务需求场景,以通信日志数据为分析对象,将基于神经网络的专家系统技术应用于机载总线网络的故障智能诊断中。选择全连接神经网络进行样本训练和分析推理,并对相关功能进行了编程实现。经过特征数据集的样本训练和验证,基于混合方法(专家系统、神经网络)的智能故障诊断技术在复杂系统和多变环境中表现出色且可靠。与传统故障诊断方法相比,这种智能化诊断技术在处理复杂故障诊断问题时展现出明显优势。综上所述,本文的研究成果为提高机载总线网络通信状态诊断效率和准确度提供了有效的技术支持。更多还原

【Abstract】 With the continuous development of avionics systems and onboard bus networks,the importance of fault diagnosis technology in ensuring flight safety and enhancing system reliability has become increasingly prominent.Traditional diagnostic methods primarily rely on data acquisition cards and monitoring software to inspect message blocks one by one,which requires professional personnel and is inefficient.This thesis conducts technical research and practice based on the demand for diagnosing communication abnormal states in the application scenarios of onboard bus networks.The aim is to apply more automated,intelligent,and efficient fault diagnosis methods to the 1553 B bus,1394 B bus,and FC network.The main work of this thesis can be divided into two parts:On one hand,this thesis provides a detailed analysis of the specific functional requirements of the 1553 B bus communication status diagnostic software and constructs the overall functional architecture of the software.By integrating the message transmission characteristics of the 1553 B bus,a series of functional modules have been designed and implemented to meet the specific needs of the diagnostic software.Particularly in the data processing module,parameters reflecting the bus communication status are analyzed and processed from multiple perspectives,and special methods for identifying abnormal states have been designed and implemented.Simulation tests and real-world scenario tests have verified that this diagnostic software can accurately reflect the communication status of the 1553 B bus system and correctly identify all abnormal communication events.Compared to traditional methods,the new diagnostic software can achieve unattended analysis and diagnosis after simple configuration,and more intuitively present the communication status through graphical output.On the other hand,in the research of fault diagnosis methods for 1394 B and FC networks,the specific business requirements scenarios were considered,and communication log data was used as the object of analysis.The expert system technology based on neural networks was applied to the intelligent fault diagnosis of avionics bus networks.A fully connected neural network was selected for sample training and analysis inference,and the related functions were programmed and implemented.Through sample training and validation of the feature dataset,the intelligent fault diagnosis technology based on a hybrid method(expert system and neural network)demonstrated excellent and reliable performance in complex systems and variable environments.Compared to traditional fault diagnosis methods,this intelligent diagnostic technology shows significant advantages in handling complex fault diagnosis problems.In summary,the research results of this thesis provide effective technical support for improving the efficiency and accuracy of communication status diagnosis in onboard bus networks.更多还原