风电机组故障的宽度学习诊断模型  

Fault classification in wind turbine using broad learning system

作　　者：郭宏宇 霍志红[1] 许昌[1] 吾买尔·吐尔逊[2,3] 周华建 程志明 Guo Hongyu;Huo Zhihong;Xu Chang;Wumaier Tuerxun;Zhou Huajian;Cheng Zhiming(College of Energy and Electrical Engineering,HoHai University,Nanjing 211100,China;College of Hydraulic and Civil Engineering,Xinjiang Agricultural University,Urumqi 830052,China;College of Energy and Electrical Engineering,HoHai University,Nanjing 210098,China)

机构地区：[1]河海大学能源与电气学院,江苏南京211100 [2]新疆农业大学水利与土木工程学院,新疆乌鲁木齐830052 [3]河海大学水利水电学院,江苏南京210098

出　　处：《可再生能源》2022年第5期634-638,共5页Renewable Energy Resources

基　　金：政府间国际科技创新合作重点专项(2019YFE0104800);新疆自然科学基金计划特培项目(2020D03004)。

摘　　要：有效的故障诊断方法不仅能快速、准确地分辨风电机组的故障类型,还能降低风电场运行维护成本。故障诊断方法所需要的相关数据均来自于风电场监控与数据采集系统(SCADA),当数据规模庞大时,浅层神经网络和深度神经网络模型会遇到权重调整耗时间、容易陷入局部最优解的问题。文章提出了一种基于宽度学习系统的风电机组故障诊断模型。首先,对风电场SCADA数据进行预处理,特征选择后构成故障样本集;然后,利用BLS诊断模型对这些故障样本进行分类。实验结果表明,与BP,SVM,ELM,DBN诊断模型相比,BLS诊断模型有效提高了风电机组故障诊断的准确率。An effective fault diagnostic method can not only distinguish the type of wind turbine faults quickly and accurately,but also reduce the operational maintenance costs of wind farms.At present,the relevant data required for fault diagnosis methods come from the SCADA system,this information can provide a rich source of data for wind turbine fault diagnosis.In order to solve the problems of long operation time and easy to fall into local optimal solution in shallow neural networks and deep neural networks when the scale of data is large,a wind turbine fault diagnostic model was proposed based on Broad Learning System(BLS).In this paper the wind farm SCADA data are constituted the fault samples set after the SCADA data was pre-processed and feature selected,then was classified by BLS model.The test results show that the BLS diagnostic model effectively improves the accuracy of wind turbine fault diagnosis compared with BP,SVM,ELM and DBNs diagnostic models.

关 键 词：宽度学习 故障诊断 风电机组 安全运行 

分 类 号：TK81[动力工程及工程热物理—流体机械及工程]