新型雷达可更换的板级电路复杂、数量多。为了快速诊断板级电路故障, 提出了采用分类的粗糙集-神经网络-专家系统的混合系统实现雷达装备的故障诊断。简要介绍混合系统进行雷达板级电路故障诊断的工作原理;建立了故障诊断样本;并以实例对粗糙集-神经网络-专家系统的混合系统进行训练和实际诊断。结果表明,该方法诊断准确性高,比其他方法更为快速有效,较好地解决了雷达可更换板级电路的故障诊断问题。

In the new type of radar, there are many complicated replaceable board-level circuits. To diagnose the board-level circuit fault rapidly, we put forward a hybrid system of rough set, neural network, and expert system for fault diagnosis of radar equipment. The working principle of the hybrid system in board-level circuit fault diagnosis of radar was introduced. A platform of fault sample training and fault diagnosis was established. An example was trained and actually diagnosed with the hybrid system of rough set-neural network-expert system on this platform. The diagnosis result showed that the fault diagnosis method is superior to others and can accurately locate fault, which can solve the problem in replaceable board-level circuit diagnosis of radar.