基于回声状态网络的磨煤机故障预警

Early Fault Warning of Coal Mills Based on Echo State Network

为实现对火电机组磨煤机断煤故障的早期预警,提出一种基于回声状态网络(ESN)的故障预警方法,结合电厂实际运行数据,与传统静态阈值方法进行比较。仿真结果表明:基于ESN的磨煤机故障预警方法输出的故障预警信号能够提前于实际断煤信号,实现磨煤机故障的早期预警。

To achieve early fault warning of interrupted coal supply from coal mills in a thermal power unit,a fault warning method was proposed based on echo state network(ESN),which was compared with traditional static threshold method,combined with actual operation data of the unit.Simulation results show that the fault warning signals based on ESN can be given in ahead of actual interrupted signals,thus achieving early fault warning of coal mills.