摘要
风电机组长期在低温高湿的环境中运行会造成风机叶片的结冰,该现象会严重影响风机的发电效率。且不同风机的运行数据在不同工作情况下也存在很大的差异,首先去除掉风机数据中的无效值,接着对风机叶片正常与结冰数据之间的不平衡进行下采样处理,然后利用RFE算法挑选出与叶片结冰最有关联性的几个特征,利用处理好的数据构建XGBoost算法模型,最后通过与其它算法模型做对比,验证针对风机叶片结冰预测本算法具有更高的准确性。
The long-term operation of wind turbine in low temperature and high humidity environment will cause the icing of fan blades,which will seriously affect the power generation efficiency of wind turbine.Moreover,there are great differences in the operation data of different fans under different working conditions.Firstly,the invalid values in the fan data are removed,and then the imbalance between the normal and icing data of fan blades is down sampled.Then,several features most related to blade icing are selected by RFE algorithm,and the XGBoost algorithm model is constructed by using the processed data.Finally,by comparing with other algorithm models,it is verified that the algorithm has higher accuracy for fan blade icing prediction.
作者
高博
张亚
GAO Bo;ZHANG Ya(School of Electrical and Information Engineering,Anhui University of Science and Technology,Huainan Anhui 232001,China)
出处
《佳木斯大学学报(自然科学版)》
CAS
2022年第5期132-135,共4页
Journal of Jiamusi University:Natural Science Edition