Small Signal Stability Assessment and Preventive Control of Power System Based on Support Vector Regression

A small signal stability assessment method based on support vector regression (SVR) and a small signal stability preventive control method based on SVR sensitivity analysis are presented in the paper, to improve the analysis speed of small signal stability preventive control. First, damping ratio prediction SVR models of low frequency oscillation mode for small signal stability are trained offline and the optimum parameters are obtained by grid searching and 5-fold cross validation. During the process of online application, the trained SVR models are used to predict the damping ratios of the concerned low frequency oscillation modes. When one of the predicted damping ratios is less than a certain threshold, it is determined as a poor or negative damping mode, and then small signal stability preventive control is started. In the process of preventive control, for poor or negative damping low frequency oscillation modes, the sensitivities of the damping ratios with respect to the control variables (the active power of controllable generators) based on SVR model are calculated and then the control amounts of generator active power are obtained by solving an optimization model. Multiple iterations are needed to make the damping ratios meet specific requirements. Case analysis results show that the effective control measures can be obtained by the presented method and the computing speed is faster than the traditional eigenvalue analysis method.