用神经网络对风力发电中电力电子故障分析

Fault analysis for power electronic equipment of wind power systems by using neural networks

随着风能资源的利用水平不断提高,风力发电系统中的电力电子装置使用也越来越多,其工作可靠性要求也越来越高。风力发电装置多位于野外,为减轻检修人员的工作负担,对风机运行中远方检测到的大量数据进行快速而有效的可视化分类及故障状态粗略判断,引入了一种用单层前向神经网络来对数据进行快速分类绘制故障非故障分界线的方法。通过该方法能够很好地根据实时数据判断风机电力电子装置的故障。

With the increasing of the utilization level of wind energy, more and more power electronic devices are applied to wind power generation systems, while the reliability requirements for them become higher and higher. To alleviate the labor force of maintenance workers for wind power generating units which usually locate in open field, a method for quick classification of faults and drawing the boundary line to distinguish faults from normal state by using the single-layer feedforward neural network is proposed, which can be used to effectively determine the faults of power electronic equipment of wind power systems according to real-time data.