电工技术学报  2021, Vol. 36 Issue (10): 2127-2139    DOI: 10.19595/j.cnki.1000-6753.tces.200278
电力系统与综合能源 |
风电机组健康状态预测中异常数据在线清洗
马然1,2, 栗文义1,2, 齐咏生2
1.内蒙古工业大学能源与动力工程学院 呼和浩特 010050;
2.内蒙古工业大学电力学院 呼和浩特 010080
Online Cleaning of Abnormal Data for the Prediction of Wind Turbine Health Condition
Ma Ran1,2, Li Wenyi1,2, Qi Yongsheng2
1. College of Energy and Power Engineering Inner Mongolia University of Technology Hohhot 010050 China;
2. College of Electrical Engineering Inner Mongolia University of Technology Hohhot 010080 China
全文: PDF (2192 KB)   HTML
输出: BibTeX | EndNote (RIS)      
摘要 风电机组数据采集与监视控制系统(SCADA)运行数据中含有大量异常数据,对风电机组健康状态预测影响严重,为此针对实测风速-功率、转速-功率数据,提出一种异常数据在线清洗方法。由于机组性能退化过程中数据特征趋于复杂,基于经验Copula-互信息(ECMI)选择关键特征参量作为数据清洗对象,并基于Copula建立置信等效功率区间描述其非线性与不确定性。针对置信边界外的堆积点和离群点,结合其时序特征与密度分布建立Copula数据清洗模型(Copula-TFDD),依次进行在线清洗。最后,基于实际数据与人工模拟数据分析模型的精度、运算效率以及对机组健康状态预测的影响表明,Copula-TFDD能准确并实时地识别各类异常数据,有效提升风电机组健康状态预测的性能。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
马然
栗文义
齐咏生
关键词 风电机组健康状态预测数据清洗特征参量互信息Copula理论    
Abstract:Wind turbine (WT) supervisory control and data acquisition (SCADA) data contains a large number of abnormal data, which has a serious impact on the prediction of WT health condition. Therefore, an online cleaning method for abnormal data is proposed according to the measured wind-power and rotate speed-power data. Due to the complexity of data features in the process of WT performance degradation, key characteristic parameters are selected as data cleaning objects based on empirical Copula-based mutual information (ECMI), and the nonlinearity and uncertainty are described by establishing confidence equivalent power interval calculated with Copula. Accordingly, the Copula-based data cleaning model combining the time-series features and density distribution (Copula-TFDD) of abnormal points is established, and online cleaning for the stacking points and outliers outside the confidence boundary is performed in turn. Finally, through the actual data and the simulation data, the accuracy and efficiency of Copula-TFDD are analyzed, and the influence on the prediction of WT health condition is also analyzed. The results show that Copula-TFDD can accurately and real-time identify various abnormal data, effectively improving the prediction performance of WT health condition.
Key wordsPrediction of wind turbine health condition    data cleaning    characteristic parameters    mutual information    Copula theory   
收稿日期: 2020-03-18     
PACS: TK83  
基金资助:国家自然科学基金项目(61763037)、内蒙古自治区高等学校科学研究项目(NJZY21305)和内蒙古自治区科技计划项目(2019,2020GG028)资助
通讯作者: 栗文义,男,1963年生,教授,博士生导师,研究方向为新能源发电技术。E-mail:lwyyyll@vip.sina.com   
作者简介: 马然,女,1982年生,讲师,博士研究生,研究方向为风电机组故障诊断与健康管理。E-mail:maran007@imut.edu.cn
引用本文:   
马然, 栗文义, 齐咏生. 风电机组健康状态预测中异常数据在线清洗[J]. 电工技术学报, 2021, 36(10): 2127-2139. Ma Ran, Li Wenyi, Qi Yongsheng. Online Cleaning of Abnormal Data for the Prediction of Wind Turbine Health Condition. Transactions of China Electrotechnical Society, 2021, 36(10): 2127-2139.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.200278          https://dgjsxb.ces-transaction.com/CN/Y2021/V36/I10/2127