基于MTF-ResNet-ViT的风电机组精细级联故障预警

The Cascaded Precise Faults Early Warning of Wind Turbine Based on MTF-ResNet-ViT

提出一种基于MTF-ResNet-ViT的风电机组(WT)精细级联故障预警方法。第1级将SCADA数据转换为马尔可夫转移场图像,利用残差网络提取故障特征,实现WT大部件状态监测和故障预警,并对故障代码数据进行标签与扩充。第2级将标签后数据灰度图像化后,利用视觉变换器建立故障代码预警模型,实现精细故障代码预警。实验结果表明,该方法可以有效标签和扩充故障代码数据,实现精细故障代码早期预警。

This paper proposes a cascaded precise WT fault early warning method based on MTF-ResNet-ViT.In the first stage,the SCADA data is converted into Markov Transition Field images,and the fault characteristics is extracted by residual network to realize condition monitoring and fault early warning for multiple main components simultaneously,then labeled and enhanced the fault code data.In the second stage,the grayscale labeled SCADA data images are processed by the vision transformer,to establish the fault code type early warning model and realized precise fault codes early warning.The results show that the proposed method can effectively label and enhance the fault code data and realize precise WT fault codes early warning.