【摘要】 提出了一种基于粗糙神经网络的歼击机操纵面故障诊断方法 .给出并证明了可利用粗集方法对故障信息进行快速特征提取的方法 ,用其作为神经网络的前置系统进行信息预处理 ,减少了所需样本数目 ,从而简化了神经网络结构 ,减少了网络训练时间 ,并且充分利用了神经网络容错及抗干扰能力 ,有效地降低了故障诊断中的误报率和漏报率 .该方法可以进行组合故障的诊断 ,且具有较好的鲁棒性 .仿真实验说明了该方法的有效性和实用性更多还原

【Abstract】 A fault diagnosis method for the fighter control surfaces is presented, which is based on rough neural network. The feature extraction based on the rough set method is given and proven, and can be utilized to pre process the fault information. Therefore, the needed training samples can be reduced, the neural network structure can be simplified, and the training time of the network can be shortened. The method takes full advantage of the neural network’s capability of fault tolerance and anti disturbance, reduces the false alarming rate and omission alarming rate, can diagnose the composed faults and can retain good robustness.更多还原