采用预测模型与模糊理论的风电机组状态参数异常辨识方法

发布时间：2018-04-27 00:03

本文选题：风电机组 + 风电场数据采集与监控系统　；参考：《电力自动化设备》2017年08期

【摘要】：为提高风电机组的停运预警能力,基于风电场数据采集与监控(SCADA)系统数据提出了一种风电机组状态参数的异常辨识方法。对参数进行划分,针对与环境因素密切相关的状态参数,采用神经网络建立了状态参数预测模型。采用本机组近期SCADA样本、本机组历史样本和其他机组近期样本分别作为预测模型的训练数据,对比分析了基于3类样本建立的模型的预测精度。采用平均绝对误差对基于本机组历史样本和其他机组近期样本建立的预测模型进行选择。定义了异常程度指标量化预测残差的异常程度。为了提高异常辨识的精度,采用模糊综合评判对筛选出的预测模型的异常辨识结果进行融合。最后,以国内某风场的1.5 MW风电机组为例进行了异常分析,并与传统的风电机组状态参数异常检测方法进行了对比,实例分析结果表明所提出的异常辨识方法具有更高的准确性。
[Abstract]:In order to improve the early warning ability of wind turbine outage, an abnormal identification method of wind turbine state parameters is proposed based on the data of wind farm data acquisition and monitoring system (SCADAA). According to the state parameters which are closely related to environmental factors, the prediction model of state parameters is established by neural network. Using the recent SCADA sample of the unit, the historical sample of the unit and the recent sample of other units as the training data of the prediction model, the prediction accuracy of the model based on the three kinds of samples is compared and analyzed. The prediction model based on the historical samples of the unit and the recent samples of other units is selected by using the mean absolute error. Anomaly degree index is defined to quantitatively predict the anomaly degree of residual error. In order to improve the accuracy of anomaly identification, fuzzy comprehensive evaluation is used to fuse the results of anomaly identification of the selected prediction model. Finally, the anomalous analysis of 1.5 MW wind turbine in a domestic wind field is carried out and compared with the traditional method of detecting abnormal state parameters of wind turbine. The analysis results show that the proposed anomaly identification method is more accurate.
【作者单位】：国网河南省电力公司电力科学研究院;重庆大学输配电装备及系统安全与新技术国家重点实验室;
【基金】：国家电网公司重大科技专项(智能变电站母线及智能组件可靠性研究)~~
【分类号】：TM315

【参考文献】