军事知识服务平台
基于时序降维的电力电量平衡主导影响因素辨识方法
Identification Method of Dominant Influencing Factors for Power and Energy Balance Based on Time-Series Dimensionality Reduction
摘要 [目的]随着波动性新能源渗透率持续提升,新型电力系统平衡方式呈现概率化特征,传统基于固定场景的电力电量平衡分析已无法适应.亟需精准辨识影响平衡结论的关键源荷特征,提升平衡场景对真实源荷波动的代表性.为此,本文提出一种基于时序降维的电力电量平衡主导影响因素辨识方法.[方法]首先,定义源荷时序特征实现平衡场景的降维分析,解决源荷时序曲线分析维度高、时序关联强的问题.其次,结合Nataf变换与数据驱动模型,表征源荷时序特征对电力电量平衡指标的多元耦合映射关系.然后,提出基于全局灵敏度分析的电力电量平衡主导影响因素辨识方法.最后,基于省级电网源荷数据开展算例分析.[结果]相较于主成分分析场景聚合,所提方法在Kolmogorov-Smirnov距离度量下的计算误差降低了27.33%,验证了所提方法的有效性.[结论]所提方法量化了源荷时序特征对电力电量平衡指标的影响程度,提升了典型平衡场景对源荷波动的代表性,为新能源电力系统的平衡分析提供了理论与技术支撑.
Abstract:
[Objective]With the continuously increasing penetration of volatile renewable energy,the balance mode of the new-type power system exhibits probabilistic characteristics,rendering the traditional fixed-scenario-based power and energy balance analysis inadequate.To improve the representativeness of balance scenarios for actual source-load fluctuations,it is essential to accurately identify the key source-load features that dominate balance conclusions.To this end,this paper proposes a method for identifying the dominant influencing factors of power and energy balance based on time-series dimensionality reduction.[Methods]First,source-load time-series features are defined to implement dimensionality reduction of balance scenarios,addressing the high dimensionality and strong temporal correlation of source-load time-series curves.Then,the multivariate coupling mapping relationship between source-load time-series features and power and energy balance indices is characterized using a data-driven model combined with the Nataf transformation.Next,a global-sensitivity-analysis-based method is proposed to identify the dominant influencing factors of power and energy balance.Finally,a case study is conducted using source-load data of a provincial power grid.[Results]Compared with principal component analysis-based scenario aggregation,the proposed method reduces the calculation error by 27.33%under the Kolmogorov-Smirnov distance metric,which verifies the effectiveness of the proposed method.[Conclusions]The proposed method quantifies the influence of source-load time-series features on power and energy balance indices and enhances the representativeness of typical balance scenarios for source-load fluctuations,thereby providing theoretical and technical support for the balance analysis of renewable energy systems.
刊 名 电力建设 ISTICPKU
年,卷(期): 2026, 47(10)
分类号: TM73
机标分类号: TM73 TP391.41 F832.51
在线出版日期: 2026年10月19日