摘要
针对目前检测风电机组偏航对风不准需添加额外设备的问题,利用风机机理和数据驱动的方法分析SCADA系统中存储的历史数据,提出基于区间划分的偏航对风不准预警方法。首先根据机组的限电、偏航、故障、维护等状态变量删除无关数据,利用随机森林和3σ原则相结合的方法剔除数据样本中存在的随机噪声,然后对功率进行等区间划分,在每一功率区间内,将风速-偏航误差二维区间内的散点图拟合一条二次曲线,求曲线的最低点对应的偏航误差即为偏航对风误差,试验结果验证了该方法的有效性,最后设计了偏航对风误差在线预警应用系统,实现机组偏航对风误差矫正的闭环落地。
In view of the problem that additional equipment is needed to detect the yaw of wind turbines,the historical data stored in SCADA system is analyzed by using the fan mechanism and data-driven method,and an interval division based yaw warning method is proposed.Firstly,irrelevant data is deleted according to the state variables such as power limiting,yaw,failure and maintenance of the unit,and random noise in the data sample is eliminated by using the method combined with random forest and 3σprinciple.Then,the power is divided into equal intervals,and the scatter plot within the two-dimensional interval of wind speed and yaw error is fitted into a quadratic curve within each power interval.The yaw error corresponding to the lowest point of the curve is the yaw to wind error.The test results verify the effectiveness of the method.Finally,an online early warning system for yaw to wind error is designed to realize the closed-loop landing of the unit yaw to wind error correction.
出处
《青海电力》
2024年第S01期81-85,共5页
Qinghai Electric Power
关键词
风电机组
偏航对风误差
区间划分
最小二乘法
wind turbine
yaw to wind error
interval division
least square method
