摘要
提出一种基于Pin-SVM的电力系统暂态稳定评估方法。首先,采用系统指标(如平均机械功率、初始加速度和系统冲击等)和投影能量函数指标(如投影角速度、投影角加速度和投影动能PKE)构建暂态稳定指标的原始特征集,通过最大相关最小冗余特征选择方法对暂态指标集进行特征压缩,寻找对电网暂态变化敏感度高的特征子集;然后,基于Pin-SVM思想将特征子集映射到高维空间,实现非线性暂态稳定评估问题的线性转换,进而引入分位数改变系统稳定类与不稳定类之间的最近点位置,将暂态稳定分类问题转换为在Pin-SVM中寻找最优分位数距离问题,以减小边界干扰样本的影响,提高电力系统暂态评估方法的评估准确率和稳定性。最后,以IEEE-39节点系统、IEEE-145节点系统和某实际算例进行仿真计算,计算结果验证了该方法的有效性和准确性。
This paper proposes a novel transient stability assessment （ TSA） method using support vector machine with pinball loss （ Pin-SVM）. A group of system - level and projection energy function classification features are first extracted from the power system operation parameters to build the original feature set of the transient stability index, such as average mechanical power, initial rotor acceleration, system shock, projection speed, projection acceleration and projection kinetic energy （PKE） and so on. Furthermore, a feature selection approach based on maximal relevance and minimal redundancy （ mRMR） criterion is employed to evaluate the feature subset which has high sensitivity to transient changes. Then, feature sets are mapped to a higher dimensional space and the TSA problem is then transformed into a linear classification problem and use the quantile to change the closest point position between system stability class and unstable class, converts transient stability classification to the problem of finding the optimal quantile distance in Pin-SVM, which can reduce the influence of boundary interference samples. Finally, the simulated results of classic IEEE 39-bus system, IEEE 145-bus system and practical power grid demonstrate the feasibility and the validity of this proposed method.
作者
陈厚合
王长江
姜涛
李雪
李国庆
Chen Houhe Wang Changjiang Jiang Tao Li Xue Li Guoqing(School of Electrical Engineering Northeast Electric Power University Jilin 132012 China)
出处
《电工技术学报》
EI
CSCD
北大核心
2017年第11期67-76,共10页
Transactions of China Electrotechnical Society
基金
国家重点研发计划(2016YFB0900903)
国家自然科学基金(51477027
51677022
51607033
51607034)
东北电力大学博士科研启动基金(BSJXM-201501)资助项目
关键词
投影能量函数
Pin-SVM
广域测量系统
暂态稳定评估
Projection energy function, support vector machine with pinball loss （ Pin-SVM） , wide area measurement system, transient stability assessment
