电器与能效管理技术 ›› 2024, Vol. 0 ›› Issue (4): 65-73.doi: 10.16628/j.cnki.2095-8188.2024.04.009

• 分布式电源及并网技术 • 上一篇    下一篇

基于SSA-GPR模型的风电机组运行状态监测

张杰1, 任康1, 马天2, 王伟璐2, 邢作霞3, 韩广明2   

  1. 1.沈阳工业大学 电气工程学院, 辽宁 沈阳 110870
    2.中国大唐集团新能源股份有限公司, 北京 100000
    3.辽宁省风力发电技术重点实验室, 辽宁 沈阳 110870
  • 收稿日期:2023-10-25 出版日期:2024-04-30 发布日期:2024-06-25
  • 作者简介:张 杰(1998—),男,硕士研究生,研究方向新能源发电技术。|任 康(1999—),男,硕士研究生,研究方向为新能源发电技术。|马 天(1988—),男,主要从事风力发电机组控制系统优化研究。
  • 基金资助:
    辽宁省兴辽英才计划项目(XLYC2008005)

Wind Turbine Operation Status Monitoring Based on SSA-GPR Model

ZHANG Jie1, REN Kang1, MA Tian2, WANG Weilu2, XING Zuoxia3, HAN Guangming2   

  1. 1. School of Electrical Engineering, Shenyang University of Technology, Shenyang 110870, China
    2. China Datang Group New Energy Co., Ltd., Beijing 100000, China
    3. Liaoning Key Laboratory of Wind Power Technology, Shenyang 110870, China
  • Received:2023-10-25 Online:2024-04-30 Published:2024-06-25

摘要:

为提高风电机组发电效率,增加经济收益,实现风电机组运行状态的在线监测,提出一种基于麻雀搜索算法优化高斯过程回归(SSA-GPR)模型的风电机组状态监测新方法。首先对数据采集与监视控制(SCADA)系统采集到的数据进行预处理分析,利用相关性分析完成模型的输入量选择;然后利用机组正常运行状态下的参数建立常态回归模型,实时计算重构误差,通过实时监测功率残差值是否超过动态故障阈值来判断机组状态。实例结果表明,所提方法的预测误差更小,并可以提前120 min实现机组异常运行状态预警。

关键词: SCADA数据, 麻雀搜索算法, 高斯过程回归, 状态监测, 风电机组

Abstract:

In order to improve the power generation efficiency and economic benefits of wind turbines, the online monitoring of the operating status of wind turbines is particularly important. A new method for monitoring the status of wind turbines based on sparrow search algorithm optimized Gaussian process(SSA-GPR) model is proposed. Firstly, the data collected from data collection and monitoring is preprocessed and analyzed. The correlation analysis is used to select the input of the model. A normal regression model using the parameters of the unit under normal operating conditions is established to calculate the reconstruction error in real-time. The unit status is determined by monitoring whether the predicted power residual exceeds the dynamic fault threshold in real-time. Through examples, it is shown that the proposed SSA-GPR model smaller prediction error and can achieve abnormal operation status warning of the unit 120 minutes in advance.

Key words: SCADA data, sparrow search algorithm(SSA), Gaussian process regression(GPR), status monitoring, wind turbine

中图分类号: