风电外送断面极限输电能力的非参数回归估计

Nonparametric Regression Estimation for Total Transfer Capability of Transmission Interface Considering Centralized Wind Power Integration

大规模风电的随机性和间歇性导致基于典型方式计算的通道极限输电能力(total transfer capability,TTC)有效性降低。提出一种TTC非参数回归估计技术,通过风电与负荷场景聚类形成代表性中心,采用二分法重复潮流计算各场景下含稳定约束的断面TTC值,提取各场景与所属中心场景间的属性偏差及TTC偏差作为特征数据样本,经过相关性检验与非参独立筛选后,利用基于三次B样条函数展开的Group Lasso算法对TTC偏差进行非参数回归估计。算例验证表明,该方法具备较强的非线性泛化能力,能以较高精度提取输电断面TTC运行规则的显性表达式,与传统方法相比具有更丰富的信息输出与更良好的解释性,可用于含风电外送断面电力系统TTC的在线快速估计。

Total transfer capability（TTC） of power interface changes as operation state varies,and is difficult to obtain based on conventional worst scenario-based method because of stochastic wind power generation.To enable online TTC estimation for wind power-related corridors,a novel nonparametric regression technique is presented.Time-series powers of wind farm and load are clustered to determine representative scenario points in order to generate data samples including TTC and relevant attributes with stability-constrained repeated bisection power flow.Correlation verification and nonparametric independence screening are then executed to determine the most effective attributes.Group Lasso algorithm based on cubic B-spline is applied thereafter to establish regressed fitting function for TTC.According to case study,the presented technique,enabling fast online TTC estimation of wind-power exporting power systems,can provide explicit solution for TTC operation numerical rules whilst maintain high accuracy with more useful information output.