基于特征椭球和支持向量机复合映射的暂态稳定预测模型

发布时间：2018-03-24 13:02

本文选题：暂态稳定识别　切入点：特征椭球　出处：《电网技术》2017年08期

【摘要】：针对电力系统暂态稳定预测问题,提出基于特征椭球理论(characteristic ellipsoid,CELL)和支持向量机(support vector machine,SVM)复合映射的方法。通过采集系统受扰后的同步相量测量单元(phasor measurement unit,PMU)量测数据,将其映射至多维样本空间中生成特征椭球。椭球形态的变化能反映出受扰系统在稳定与失稳场景下不同的动态行为,从而识别出系统的暂态稳定性。以包含了系统暂态过程信息的椭球几何属性,如体积、偏心率、中心点、体积变化率等,构造支持向量机评估模型的输入特征,训练所得支持向量机模型能够实现电力系统暂态稳定在线预测。该方法以椭球的动态变化作为暂态稳定判据,利用较少的量测数据就能为暂态稳定识别提供足够的信息量,同时保证较高的准确率。通过IEEE 39节点系统和我国西南某省级区域输电系统的算例测试,验证了方法的可行性和有效性。
[Abstract]:In order to predict the transient stability of power system, a method of complex mapping based on characteristic ellipsoid theory and support vector machine (SVM) is proposed. The data of synchronous phasor measurement unit (PMU) after the system is disturbed are collected. The change of ellipsoid shape can reflect the dynamic behavior of the disturbed system in stable and unstable scenarios. In order to identify the transient stability of the system, the input features of the support vector machine (SVM) evaluation model are constructed to include the ellipsoidal geometric attributes of the transient process information, such as volume, eccentricity, center point, volume change rate, etc. The trained support vector machine (SVM) model can be used to predict the transient stability of power system on line. Using the ellipsoid dynamic change as the transient stability criterion, the method can provide enough information for transient stability identification by using less measured data. At the same time, the high accuracy is guaranteed. The feasibility and effectiveness of the method are verified by the example of IEEE 39 bus system and a provincial transmission system in southwest China.
【作者单位】：四川大学电气信息学院;四川电力科学研究院;
【基金】：国家自然科学基金资助项目(51437003,51207098)~~
【分类号】：TM712

【参考文献】