小世界回声状态网络的风机变桨系统预警模型研究

Research on Early Warning Model of Fan Propeller System Based on Small-world Echo State Network

提出一种基于自编码器的小世界回声状态网络的故障预测方法,以此来应对风电机组变桨系统可能产生的故障,提高故障响应速度。通过分析系统中的监测参数和数据,建立基于自编码器的小世界回声状态网络系统模型。实例表明该方法可在故障发生前提前作出预警,验证了方法的可行性,有效降低了机组的运维成本。

A fault prediction method of small-world echo state network based on Autoencoder is presented to deal with the possible faults of wind turbine pitch control system and improve the fault response speed.By analyzing the monitoring parameters and data in the system,a small-world echo state network system model based on Autoencoder is established.An example shows that the method can give early warning before the failure occurs,verify the feasibility of the method,and effectively reduce the operating cost of the unit.