BP神经网络在雷达故障诊断中的应用研究

对雷达混合电路故障诊断问题进行了研究,针对传统雷达故障诊断中的“自适应差”,“过分依赖故障数学模型”的缺陷。采用了一种故障诊断表示模型,提出将专家系统与神经网络相结合的诊断方法,避免了在实际中雷达电路难以建立精确的数学模型的问题。利用人工神经网络理论为其建立了BP神经故障诊断网络,增加了雷达电路故障诊断系统的“自适应性”。进而给出了网络训练算法和诊断步骤,最后给出了在某型雷达设备上仿真调试诊断的实例。

Application Research of BP Neural Network in Fault Diagnosis of Radar

Aimed at "self adaptation limitation" and "dependency of fault math model" of traditional fault diagnosis of radar,a method based on integration of expert system and neural network is proposed.A BP neural network is founded,which bases on theory of ANN,the training algorithm and fault diagnosis steps are described.An imitated diagnosis examples in some part of a radar system is also given.