基于SDAE特征提取的含风电电网可用输电能力计算

风力发电的不确定性显著增加了电力系统可用输电能力(ATC)计算的难度。基于点估计的Gram-Charlier级数展开理论和深度学习技术,提出了一种计及越限概率要求的ATC快速计算方法,考虑的约束类型包括静态安全、静态电压稳定和暂态稳定约束。假定风电出力概率分布已知,结合两点估计法和Gram-Charlier级数展开,通过两个确定性场景的最大输电能力(TTC)计算结果逼近TTC的累积分布函数。为了快速、准确地获得确定性场景的TTC,利用堆叠降噪自动编码器(SDAE)建立了TTC计算的深度学习模型。获得TTC的累积分布函数后,将断面功率超过TTC的概率定义为越限概率,推导了给定越限概率要求下ATC计算的表达式。实际电网仿真结果表明,所提方法能够有效计及多类安全稳定约束,快速、准确计算不同越限概率要求下的ATC。

Available Transfer Capability Calculation in Power System With Wind Power Based on SDAE Feature Extraction

The uncertainty of wind power makes available transfer capability(ATC)calculation more difficult. A fast ATC calculation method under certain limit violation probability request is proposed based on the point estimate method, Gram-Charlier series expansion theory and deep learning. The types of constraints considered include static security, static voltage stability and transient stability constraints. Assuming the probability distribution function of wind power is known, the cumulative distribution function of total transfer capability(TTC)is approximated through the TTC results of two deterministic operation scenes according to two-point estimate method and Gram-Charlier series expansion. In order to calculate TTC of the deterministic operation scene fast and accurately, a fast TTC calculation model based on stacked denoising autoencoder(SDAE)is developed. After the cumulative distribution function of TTC is known, the limit violation probability is defined as the probability that the power of flowgate is more than TTC and the calculation formula of ATC under certain limit violation probability is derived. Experiment results of a real power system demonstrate that the proposed method is able to consider multiple security and stability constraints effectively and calculate ATC under different limit violation probability requests fast and accurately.