概率逻辑
可再生能源
计算机科学
分布式发电
风力发电
特征向量
矩阵分解
因式分解
电力系统
理论(学习稳定性)
分位数
数学优化
可靠性工程
算法
工程类
功率(物理)
数学
机器学习
统计
人工智能
电气工程
物理
量子力学
作者
Haosen Yang,Xin Chun Shi,Robert C. Qiu,Xing He,Qian Ai,Ziqiang Wang
标识
DOI:10.1109/tsg.2021.3128503
摘要
The increasing integration of large number of renewable energy sources (RESs) generates multiple kinds of uncertainties in power grids, which heavily challenges traditional deterministic approaches for voltage stability assessment (VSA). Under this circumstance, this paper proposes a high-dimensional spatial-temporal matrix factorization (STMF) approach for monitoring data to separate the deterministic part and uncertain elements caused by RESs, and presents a probabilistic index to online evaluate voltage stability. Specifically, this method explores the eigenvalue distribution of measurement data when penetrating multiple distributed wind and solar generations, as well as a learning-based STMF model is accordingly used to iteratively extract the random elements due to RESs. Unlike traditional matrix factorization approaches, this STMF method uses an adaptive and analytical model to depict the eigenvalue distribution of random components. Subsequently, a probabilistic index is designed to measure the voltage stability degree as well as provide quantiles to indicate its potential fluctuations. Numerous cases verify the effectiveness and advantages of the proposed method using different load changing models and various RES penetration levels, as well as other probabilistic and deterministic indicators are compared.
科研通智能强力驱动
Strongly Powered by AbleSci AI