一种具有故障诊断与预测功能的信息化节点测试设备研制

发布时间：2018-04-05 07:17

本文选题：节点测试设备　切入点：分布式测试系统　出处：《哈尔滨工业大学》2017年硕士论文

【摘要】：随着电子信息系统复杂化、大型化、网络化、智能化程度的不断提高,电子信息系统的测试设备朝着“分布式采集,集中化分析管理,共享数据资源”的分布式测试系统的方向发展。以往的分布式测试系统的节点测试设备只负责状态监测、往往不具备故障诊断和故障预测功能,并且通用性差,需针对测试节点的测试需求开发专用设备,测试成本高,维护困难。针对这一问题,本课题研制一种具有一定通用性、信息交互能力、故障诊断能力和故障预测能力的节点测试设备,以构建信息化分布式的测试系统,提高对复杂电子信息系统的测试能力和故障诊断、预测能力。在硬件设计上,选用高集成度、端子可复用的仪器模块满足被测节点测试信号的测试需求,利用仪器模块的端子配置功能和超宽测量输入替代TUA(Test Unit Adapter,测试接口适配器)的信号分配与调理功能,取消了专用的TUA,实现了节点测试设备的互换性、通用性和设备的小型化。另外,通过选用便携式的PXI机箱,进一步减小了节点测试设备的体积。针对电子信息系统故障原因与症状的随机性与不确定性、故障样本少、状态信息有限和先验消息多源异类等特点,本课题在改进DS(Dempster-Shafer)证据理论处理冲突证据不当问题的基础上,提出了基于灰色关联分析与改进DS证据推理的故障诊断方法。该方法结合了灰色关联分析处理“小样本、贫信息、不确定性”问题的优点与DS证据理论在信息融合中考虑不确定性的优点。针对大功率电源的电压和光纤陀螺的随机漂移误差均随时间呈趋势变化的特点,提出了基于ARMA-Elman神经网络的故障预测方法,该方法结合了ARMA模型对于线性时间序列的拟合能力与Elman神经网络对非线性时间序列的映射能力。提出的故障诊断算法与故障预测算法均以组件的形式实现,供节点测试设备调用。测试软件基于本单位的联合试验平台开发,通过开发具有基本功能的组件构建试验方案实现测试软件的功能。基于本单位联合试验平台提供的信息交互服务,实现各节点测试设备间相互访问、共享数据和相互调用。在故障诊断和故障预测组件的开发工作中,通过构建组件调用Matlab引擎的框架,方便将来扩充更多的算法以提高组件故障诊断和故障预测的能力。最后,完成了软硬件的系统集成测试工作,通过虚拟被测对象(Unit Under Test,UUT)和实物信号源对设备具有的信息化能力、远程测试能力、故障诊断能力和故障预测能力进行验证,测试结果表明本课题研制的节点测试设备能够满足研制要求。
[Abstract]:As the electronic information system becomes more and more complex, large-scale, networked and intelligent, the test equipment of the electronic information system is oriented to "distributed collection, centralized analysis and management,"The development of distributed test system based on sharing data resources.In the past, the node test equipment of the distributed test system was only responsible for state monitoring, and often did not have the function of fault diagnosis and fault prediction, and the generality was poor. Therefore, special equipment should be developed to meet the test requirements of the test nodes, and the test cost was high.Maintenance is difficult.In order to solve this problem, this paper develops a kind of node test equipment which has certain generality, ability of information exchange, ability of fault diagnosis and ability of fault prediction, in order to construct an information distributed test system.Improve the test ability, fault diagnosis and prediction ability of complex electronic information system.In the hardware design, the instrument module with high integration and reusable terminal is selected to meet the test requirements of the test signal of the node under test.By using the terminal configuration function of the instrument module and the signal assignment and conditioning function of the ultra-wide measurement input instead of the TUA(Test Unit Adapter (test interface adapter), the special TUAs are eliminated, and the interchangeability, versatility and miniaturization of the node test equipment are realized.In addition, the volume of the node test equipment is further reduced by selecting the portable PXI chassis.Aiming at the randomness and uncertainty of fault causes and symptoms of electronic information system, the small number of fault samples, the limited state information and the multi-source heterogeneity of prior messages, this subject is based on the improvement of DSN Dempster-Shafer evidence theory to deal with the problem of improper conflict evidence.A fault diagnosis method based on grey correlation analysis and improved DS evidence reasoning is proposed.This method combines the advantages of "small sample, poor information, uncertainty" problem in grey relational analysis and the advantage of considering uncertainty in information fusion based on DS evidence theory.In view of the fact that the voltage of high power supply and the random drift error of fiber optic gyroscope are changing with time, a fault prediction method based on ARMA-Elman neural network is proposed.This method combines the fitting ability of ARMA model for linear time series and the mapping ability of Elman neural network to nonlinear time series.Both the fault diagnosis algorithm and the fault prediction algorithm are implemented in the form of components, which can be called by the node test equipment.The test software is developed based on the joint test platform of our unit, and the function of the test software is realized by developing components with basic functions to build the test scheme.Based on the information interactive service provided by the joint test platform, the test equipments of each node can access each other, share data and call each other.In the development of fault diagnosis and fault prediction components, it is convenient to expand more algorithms to improve the ability of component fault diagnosis and fault prediction by constructing the framework of component calling Matlab engine.Finally, the system integration test of hardware and software is completed, and the information ability, remote test ability, fault diagnosis ability and fault prediction ability of the equipment are verified by virtual object unit Under Test UUTU and physical signal source.The test results show that the node test equipment developed in this paper can meet the requirements of the development.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP277

【参考文献】