摘要
为了评估风电机组技术改造的效果,需利用运行数据对其发电功率进行建模。采用偏最小二乘(PLS)回归的方法为风电机组的发电功率进行分段线性建模。其输入包含与本机相关性较高的相邻机组的运行数据,但不包含风速测量值,以避免因测量精度低和输入变量维数少而导致的模型精度低的问题。将得到的模型应用在改造后的机组上。对真实输出和模型输出进行比较,以评估改造对发电功率提升的有效性。该方法完全基于数据,而无需机理模型或知识,为评估系统改造效果提供了一种切实可行的新思路,可用于各种复杂系统。金紫山风场的建模和应用结果证明,目标风机机组的改造提升效果明显。
In order to evaluate the effect of technical improvement of wind turbine,it is necessary to build a model of power generation by using operational data.Piecewise linear modeling is conducted for power generation of wind turbine by using the partial least squares(PLS)regression approach.The inputs include operational data of adjacent unit that are highly correlated with the unit;but does not include wind speed measurement;to avoid the issue of low model accuracy caused by low measurement accuracy and small input variable dimension.By applying the obtained model on the generation turbine after the improvement,and comparing the true output and the outputs of the model,the effectiveness of the improvement of power generation can be assessed.The proposed method is fully based on data,which provides a practical new idea for evaluating the performance of system improvement and can be used for various complex systems,without requirement of mechanism model or knowledge.The modeling and application results on the Jinzishan Wind Farm show significant improvement of the objective generation set.
作者
杨帆
郭岑
柯国勇
王德政
叶昊
YANG Fan;GUO Cen;KE Guoyong;WANG Dezheng;YE Hao(Department of Automation,Tsinghua University,Beijing 100084,China;Xinjiang Goldwind Science&Technology Co.,Ltd.,Urumqi 830026,China;College of Urban Rail Transit and Logistics,Beijing Union University,Beijing 100101,China)
出处
《自动化仪表》
CAS
2019年第10期29-34,共6页
Process Automation Instrumentation
基金
国家自然科学基金资助项目(61873142)
关键词
偏最小二乘
风电机组
回归模型
黑箱
相关性分析
技术改造
分段线性建模
主元
Partial least squares(PLS)
Wind generation turbine
Regression model
Black box
Correlation analysis
Technical improvement
Piecewise linear modeling
Principal components
