基于多级信息融合的转台故障诊断系统研究  

Study of fault-diagnosing system of turntable based on hierarchical information fusion

作　　者：付强[1] 苏宝库[1] 

机构地区：[1]哈尔滨工业大学航天学院,哈尔滨150001

出　　处：《哈尔滨工业大学学报》2006年第11期1906-1909,共4页Journal of Harbin Institute of Technology

摘　　要：根据转台系统复杂的结构特点,提出了一种多级信息融合的故障诊断方法.首先针对来自各子系统的多传感器信息,应用神经网络进行局部故障模式分类;然后,将多个子系统诊断结果应用D-S证据理论加以综合,给出系统级的故障诊断决策.实验表明,该方法在复杂系统故障模式分类和不确定性推理具有较强的处理能力,能有效地提高故障诊断的可靠性和准确度.A fault-diagnosing method based on hierarchical information fusion is presented according to the complex structure of turntable. In this method, Firstly neural network is applied to field fault mode classification on the basis of multi-sensor information from subsystems, then D - S evidence theory is used to synthesis the diagnosis results from subsystems and give the system-level fault-diagnosing decision. Experiments show that this method has powerful processing ability of fault mode classification and uncertain reasoning in complex systems, and also it can effectively improve the fault-diagnosing reliability and accuracy.

关 键 词：信息融合 故障诊断 神经网络 证据理论 转台 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]