摘要
针对某型战机航空火控系统故障诊断方法对维修人员、检测设备依赖性大、故障诊断时间长等弊端,选用了对非线性对象有较好的控制及扰动消除效果的Elman神经网络方法,并将其应用于火控系统的故障诊断,为了提高网络性能,对Elman网络进行了相应的改进,在结构单元增加了自反馈增益因子α,并建立了基于Elman神经网络的火控系统故障诊断模型,通过一定的故障样本进行了训练和测试,结果证明该方法能有效地识别出故障原因,故障诊断准确率较高,有较强的鲁棒性。
The fault diagnosis of an airborne Fire Control System（FCS） on a certain fighter aircraft has some disadvantages, such as high dependency on the maintainers and the detection equipment, and long time cost on fault diagnosis. To solve the problem, Elman neural network, which could preferably control the nonlinear objects and eliminate disturbances, was chosen and used in the FCS fault diagnosis. The Elman network was modified to improve its performance. A self-feedback gain factor（or） was added in the structure cell. The fault diagnosis model of FCS based on the Elman neural network was established. Training and testing with some fault samples showed that： 1 ） Reasons for failure can be effectively identified with this method; 2） The accuracy rate of the fault diagnosis is higher; and 3） The robustness of the system is better.
出处
《电光与控制》
北大核心
2009年第6期66-68,85,共4页
Electronics Optics & Control
基金
空军创新基金资助(CXJJ0521)
