基于粗糙集与BP神经网络的烟气机故障诊断

Fault diagnosis of turbine machine based on rough set and artificial neural network

结合粗糙集理论和神经网络在信息处理方面的优势,提出了一种基于粗糙集理论与BP神经网络相结合的烟气机故障诊断方法。首先对故障诊断数据中的连续属性进行离散化,然后根据粗糙集理论进行故障诊断决策系统约简,获得最优决策系统。最后在最优决策系统的基础上,设计BP神经网络对烟气机故障进行诊断。试验结果表明,该方法可以有效提高烟气机故障诊断的精度和效率。

A new hybrid system of rough set and neural network for intelligent fault diagnosis was presented in this paper. Firstly, the continuous properties in diagnostic decision system were made discretely. Secondly, based on rough set theory, the diagnostic system was simplified to obtain optimum decision system, Finally, with this decision system, BP neural net work was designed to diagnose the faults occurred in turbine machine. The test results show that this method can efficiently enhance the precision of fault diagnosis for turbine machine.