摘要
为了准确掌握鼠笼型感应风力发电机组的运行状态,在分析其功率特性的基础上,针对风电机组建模过程中容易忽略偏航系统误差的问题提出了一种充分考虑风电机组输入参量的建模方法,并结合现场数据分析发现依据该方法建立的模型是有偏差的;在深入分析后再次提出两种引入机组惯性的模型改进方案。用LSSVM方法建立各个方案模型,测试结果表明,引入机组惯性可以有效地提高机组功率特性模型的精度,模型的相对平均误差(MRE)由6.7%下降到5.6%;多组测试数据的测试结果表现出模型有较好泛化能力,可以准确跟踪机组功率变化,为风电机组功率特性建模提供一个新思路。
In order to grasp running state of squirrel-cage induction wind power generator unit accurately, a modeling method is put forward on the base of power characteristics analysis, which fully considers input parameters to avoid the situation that yaw system error is easily ignored, however the established model has obvious error compared with operating data. After in depth analysis of the error,two more improved methods are proposed with introduction of the inertia of unit. The least square support vector machine（LSSVM） is used to establish models respectively by those methods. The results show that the introduction of unit inertia factors into the power characteristic model can effectively improve its accuracy, relative average error of the model reduces from 6.7% to 5.6%; many sets of test data show that the model has good ability of generalization and can track the unit power output accurately, which provides a new way of modeling for power feature of wind power generator unit.
出处
《可再生能源》
CAS
北大核心
2015年第6期865-870,共6页
Renewable Energy Resources
关键词
风电机组
现场数据
功率特性
惯性
建模
wind power generator unit
operating data
power features
inertia
modeling
