粗糙集-神经网络在民航飞机故障数据处理中的应用  

Application of Rough Set-Neural Network in Civil Aviation Aircraft Fault Data Processing

作　　者：张鹏[1] 郝一川[2] 

机构地区：[1]中国民航大学工程技术训练中心,天津300300 [2]中国民航大学航空自动化学院,天津300300

出　　处：《控制工程》2016年第1期87-90,共4页Control Engineering of China

基　　金：民航联合研究基金(U1533201)

摘　　要：飞机系统的复杂性使得故障诊断的难度越来越大,多种诊断方法相互融合将会成为飞机故障诊断研究的发展趋势。由于粗糙集理论能够有效简化知识,将粗糙集理论和神经网络结合起来,利用基于信道容量的知识相对约简算法对输入信息进行简化。提高网络的收敛性,达到提高整个融合系统效率的目的。通过对飞机的故障诊断实验证明了该方法的有效性。Because of the complexity of aircraft systems, fault diagnosis is getting more and more difficult. The combination of different methods achieves improvement and becomes a tendency of research. Since rough set theory can effectively simplify information, it is combined with neural networks, and the knowledge relative reduction algorithm based on channel capacity is used to simplify the input information, which improves the convergence of the network and simplifies the process of neural network training. The effectiveness of this method is verified by aircraft fault diagnosis tests.

关 键 词：故障诊断 飞机系统 模糊集 神经网络 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]