【作者】 孙慧平；

【导师】 袁赣南；

【作者基本信息】 哈尔滨工程大学 ， 导航、制导与控制， 2007， 硕士

【摘要】 综合导航显控台是导航系统的核心装置，其可靠性关系到整个导航系统的安危。如何快速准确的诊断系统故障就成为对导航保障部门的迫切要求。本文针对以往综合导航显控台故障检测工具简单、检测人员经验不足等问题，从实际应用的角度出发，深入分析和研究了综合导航显控台的智能故障检测与诊断技术。首先详细叙述了综合导航显控台故障诊断专家系统的设计过程。对综合导航显控台各种故障现象和检测经验进行搜集整理，利用故障树分析法和Access数据库知识建立综合导航系统的故障知识库；借助SQL语言强大的搜索功能在VB编程环境下设计推理机和解释机，实现了专家系统的推理和解释功能。接着分析了基于神经网络的故障诊断方法。具体分析了利用BP网络建立诊断模型的方法，建立了综合导航显控台串行通讯的BP网络故障诊断模型，利用引入动量项的BP改进算法进行网络训练学习，收敛速度非常快，并利用训练好的网络模型进行实际诊断，其输出数据和实际数据值误差很小，具有较高的识别精度。然后编写了单板故障检测与诊断软件。详细分析了综合导航显控台串行通讯故障、网络通讯故障、CAN总线通讯故障，编写了包括网卡、串口卡、CAN卡等板卡级故障检测软件。最后将专家系统诊断、神经网络诊断和单板故障检测有机组合成综导显控台智能故障诊断系统。更多还原

【Abstract】 Integrated Navigation Display Control Console (INDCC) is the core equipment of the navigation information processing. Its dependability associated with the safety of navigation system. How to detect the fault on the INDCC makes a big challenge to the navy maintenance and repair department.In allusion to the problems of the simple tools and shortage of peoples’ experience in the formerly fault detecting, considering the actual application, the article gone deep into the analysis of the fault diagnose technique on the INDCC. The works are mainly on following aspects.Firstly, the paper described the design of the fault Diagnosis Expert System for INDCC in details. It collected and coordinated the malfunction phenomenas and detecting experiences of INDCC, on the basis of Fault Tree Analysis and the ACCESS 2000 Data Base, knowledge base of the systems was built up, and reasoning machine and explaining machine were designed on the basis of the SQL language.Secondly, sample of the fault detection based on the neural network, especially on the BP neural network, was studied. The BP model of the COM was built up, after the training by the BP improved studying algorithms which momentum was introduced into, the sample was convergent successfully and had high identification precision.Thirdly, software of the singlechip of the INDCC were studied and designed detailedly, including COM, network communication and CAN bus communication.Finally, the intelligent fault diagnosis system was implemented according to combining the expert system diagnosis, neural network diagnosis and singlechip detection.更多还原