摘要
为了最大限度地减少风机停机时间和提高风机发电量,基于风机功率曲线特性,结合多元统计Hotelling T^2控制图,提出了一种风力发电机性能及故障监测方法。首先,根据SCADA系统历史数据集,应用粒子群算法(PSO)寻优最小二乘支持向量机的模型,构造风电机组参考功率曲线。然后,计算风场各风机功率特性的多元峰度、多元偏度,将其偏离参考曲线的程度作为评估风力发电机性能的指标。最后,监测风机发生故障的时刻,引入用于监测风机的Hotelling T^2多变量质量控制图。将该方法用于某风场1.5 MW级风力发电机,实例表明,该算法可以有效地对风电机组状态及故障进行监测,为风电机组的故障识别及分析提供了一种新的方法。
In order to minimize the downtime and maximize the power generation of wind turbine,a wind turbine performance and fault monitoring method is proposed based on the wind turbine power curve characteristic,combined with the multivariate statistical Hotelling T^2 control charts. Firstly,according to the historical data set of SCADA system,the particle swarm optimization（PSO）is used to optimize the model of the least squares support vector machine,and the reference power curve of wind turbine is constructed. Then,the multivariate kurtosis and multivariate skewness of each wind turbine power curve in the wind farm are calculated,and the degree of deviation from the reference curve is used as an index to evaluate the performance of the wind turbine. Finally,to monitor the time of the turbine failure,introducing the Hotelling T^2 multivariate quality control charts. The method is applied to a 1.5 MW wind turbine in a wind farm. The result shows that this algorithm can effectively monitor the state and fault of wind turbine,which provides a new method for fault identification and analysis of wind turbine.
出处
《可再生能源》
CAS
北大核心
2018年第2期302-308,共7页
Renewable Energy Resources
基金
河北省科技计划项目(16214510D)
河北省科技计划项目(17214304D)
作者简介
梁涛（1975-），男，博士，教授。主要从事风力发电技术及其系统分析等。E—mail：ckzyj0122@163．com.
