基于轨迹输入特征支持向量机的湖南电网暂态稳定在线识别

Hunan power grid transient stability online detection based on support vector machine with trajectory input features

针对目前基于人工智能算法的电力系统暂态稳定评估输入特征选择困难的问题,采用发电机单一状态变量故障清除后一段时间内的轨迹作为支持向量机(Support Vector Machine,SVM)的输入特征,并采用网格法寻找支持向量机在交叉验证意义下的最优参数。通过对这类支持向量机的性能进行详细的分析和对比,给出了3个分类准确度很高的发电机状态变量。在湖南电网的测试系统上仿真实现了该模型,仿真结果证明所提基于轨迹输入特征的支持向量机具有很高的精度,为支持向量机输入特征的选取提供了新思路。

Regarding to the difficulty of selecting input features in power system transient stability assessment based on intelligent algorithm, this paper adopts post fault trajectory of a single generator state variable as the input features of support vector machine （SVM）, and uses grid search method to find the optimal parameters for SVM under the meaning of cross validation. By analyzing and comparing the performance of these SVMs, three generator state variables with high classification accuracy are proposed. The simulations on Hunan power grid show that the accuracy of the proposed SVM based on trajectory input features is very high and provides a new way to select the input features of SVM.