位置:成果数据库 > 期刊 > 期刊详情页
基于类神经网络的MW永磁风力发电机短路故障智能诊断
  • 期刊名称:电机与控制应用
  • 时间:0
  • 页码:24-29
  • 语言:中文
  • 分类:TM315[电气工程—电机]
  • 作者机构:[1]新疆大学电气工程学院,新疆乌鲁木齐830047
  • 相关基金:国家自然科学基金项目(50867004 50767003); 新疆大学博士科研启动基金项目(BS100123)
  • 相关项目:新疆风电场短期发电量预算方法研究
中文摘要:

针对大型永磁风力发电机故障诊断工作薄弱的现实,通过建立基于发电机结构的典型短路故障仿真模型,对多种电磁场和温度场故障联合研究,利用同一模型场路耦合求解电磁场和温度场,得出短路故障发生时的电磁场和温度场数据及分布规律;为提高诊断可靠性,先利用BP网络诊断,后利用Elman神经网络对动态系统诊断较为优越的特点诊断,最后概率神经网络(PNN)对短路故障发生时的多源数据(电流、磁场、温度和振动特征信息)融合处理诊断,以判断发生单一或者复合故障;结合MW永磁直驱样机的试验数据及风电场的运行数据,对发电机典型短路故障进行理论结合试验的诊断,分类对比了诊断结论。

英文摘要:

Aim at the feeble fact of large permanent magnet wind power generator fault diagnosis work,typical short circuit fault simulation model was built based on generator structure.Typical short circuit fault about thermal and electromagnetic field was studied united.Utilize same model,field and route coupling method were used to solve electromagnetic and thermal field,data and distribution discipline of electromagnetic and thermal field were obtained when the fault happened.To enhance diagnosis reliability,BP neural network was utilized firstly,after that,Elman neural network was employed later because it was predominant to dynamic system diagnosis,PNN network was used to estimate simplex or multiple fault that happened by syncretize multiform data(current magnetic field temperature and vibration character information) for diagnosis when short circuit fault happened.Combine experiment data of MW permanent magnet direct drive sample generator and circulation data of wind farm,paper diagnose and classify fault by theory integrate experiment,classify and contrast diagnosis conclusion.

同期刊论文项目
同项目期刊论文