基于功率曲线的风电机组数据清洗算法

针对风电机组性能分析过程繁琐低效、数据清洗不彻底以及传统方法难以有效识别复杂多变的异常发电状态的问题,提出一种用于风电机组功率曲线分析的数据清洗算法。通过分析风电机组数据采集与监控(SCADA)系统采集的风速功率数据,优化数据处理规则与数据分析过程,提出最优组内方差清洗算法,检测机组发电性能异常的状态,降低对检测工具和数据维度的硬性要求。实例分析表明该方法实用、高效,在不增加硬件设备投资的前提下,能准确清洗风电机组功率曲线数据并识别出机组异常运行状态,显著提高了风电机组性能分析的准确性。

Wind Turbine Data-cleaning Algorithm Based on Power Curve

In view of inefficient analysis process for wind turbine performance, inaccurate data cleaning and hard to identifying of wind power generation status, a data cleaning algorithm is put forward to analyze the wind turbines power curve. Through the analysis of the wind power curve data collected by supervisory control and data acquisition(SCADA)system, the optimal interclass variance algorithm is proposed to identify the poor performance status accurately with optimizing data processing rules and analysis approach. The detection tools and data dimension are not the obstacles of power curve analysis. Example analysis shows that the method is practical and efficient, and can accurately clean wind turbines power curve data and identify abnormal status performance under the premise of no increase in the hardware equipment investment, while significantly improving the wind turbines performance analysis accuracy.This work is supported by Jilin Province Science and Technology Development Project(No. 20150204084GX).