摘要
针对风力发电机故障诊断精度低的问题,基于数据采集与监视控制系统,即SCADA数据集,提出了优化极端梯度提升树的风力发电机故障诊断方法。首先以均衡准确率作为评价模型的指标,并对数据特征进行选择,去除多余特征;然后对XGBoost相关超参数进行调优,提高模型故障诊断的准确性;最后构建XGBoost故障诊断模型。结果表明,所使用的方法取得了98.5%的均衡准确率,并将实验结果与其他常用算法结果相比,提高了1%-3%,表明了所提方法能够有效地提高风力发电机故障诊断性能。
Aiming at the problem of low fault diagnosis accuracy of wind turbine,a fault diagnosis method of wind turbine is pro⁃posed based on the data acquisition and monitoring control system,namely SCADA dataset,the method can optimize the extreme gradi⁃ent lifting tree.Firstly,the equilibrium accuracy is used as the index for evaluating the model,and the data features are selected to re⁃move the redundant features;then the XGBoost-related hyperparameters are tuned to improve the accuracy of the model fault diagno⁃sis;finally,the XGBoost fault diagnosis model is constructed.The results show that the used method achieves an equilibrium accura⁃cy of 98.5%,and the experimental results improve by1%-3%compared with other commonly used algorithms result,indicating that the proposed method can effectively improve the performance of wind turbine fault diagnosis.
作者
耿家豪
廖宇
马先超
王港
GENG Jiahao;LIAO Yu;MA Xianchao;WANG Gang(Hubei Minzu University,Enshi 445000,China)
出处
《通信与信息技术》
2022年第6期42-46,共5页
Communication & Information Technology
基金
教育部2021年产学合作协同育人项目(批准号:202102648020)资助的课题
博士启动基金(批准号:MD2019B006)资助的课题
湖北民族大学高水平培育项目(批准号:PY22012)资助的课题。
