基于机器学习算法与SCADA系统的风电机组变桨系统变频器的故障预警方法研究

变桨系统是风电机组的关键设备,但由于风电机组长期处于复杂的工作环境,导致变桨系统故障成为风电机组故障中最常见的故障之一,而变桨系统变频器故障在变桨系统故障中的占比很高.基于此,提出了一种变桨系统变频器的故障预警方案,分析SCADA系统数据,将机器学习算法应用于故障预警,并将模型温度残差作为故障预警的指标;然后,针对随机...

RESEARCH ON FAULT EARLY WARNING METHOD OF WIND TURBINE PITCH SYSTEM INVERTER BASED ON MACHINE LEARNING ALGORITHM AND SCADA SYSTEM

Pitch system is the key equipment of wind turbine,Due to the complicated working environment for a long time,pitch system failure is one of the most common failure of wind turbine,and inverter fault in the pitch system fault accounts for a high proportion.For this reason,this paper proposes a inverter fault early warning scheme of pitch system,analyzes the data of SCADA system,applies machine learning algorithm to fault early warning,and takes model temperature residual as the index of fault early warning.To solve the problem of random factor interference model temperature residual,the sliding time window residual estimation method is proposed to analyze the temperature residual trend of the early warning model,and to judge the state of pith system inverter and then to evaluate the operation state of the pitch system.Take the operating data of a wind farm in Xinjiang Uygur Autonomous Region as an example to verify the proposed model,the experimental results show that the model and the residual estimation method of sliding time window can give early warning to potential faults in time,and the feasibility and effectiveness of the scheme are verified.