基于变精度粗集的神经网络故障诊断模型研究

Study on neural network based on variable precision rough set for fault diagnosis

为处理航空发动机故障信息系统,提出了一种基于变精度粗集和神经网络相结合的故障诊断模型。利用上(下)分布约简理论,剔除目标信息系统中的冗余属性。通过选取主要影响发动机状态的因素,应用BP神经网络建立诊断模型并对网络进行训练和测试。实验结果表明,该模型具有一定的实用性,为航空发动机的故障诊断提供了一种有效的判断方法。

For dealing with the fault information system ofaero-engine, a combination model of variable precision rough set and neural network for fault diagnosis is given, The upper （lower） distribution reduction theory is used to reduce the unnecessary factors in target information system. BP network is used to build diagnosis model by choosing the mainly influential factors of aero-engine state, then it is trained and tested, The experiment shows that the model has certain practicability, so an effective judgment method is provided for the aero-engine fault diagnosis,