中文摘要：

提出了一种基于v-支持向量回归（v-SVR）的快速预测扰动后电力系统最低频率的方法。该方法考虑了发电机的最大出力限制、旋转备用的水平及分配方式、原动机-调速器系统和负荷等对电力系统频率动态的影响。通过与PSS/E仿真结果进行比较可知,使用v-SVR方法可以快速准确地预测扰动后电力系统频率动态及最低频率,具有良好的泛化能力和推广性。进一步,可将使用v-SVR方法所训练出的模型应用于电力系统频率的在线安全稳定评估和根据评估情况制定相应的紧急控制措施,防止系统频率崩溃。

英文摘要：

A method of rapid minimum frequency prediction based on v-SVR（v-Support Vector Regression） is presented for the post-disturbance power system,which considers several influencing factors on the frequency dynamic of power systems,such as the maximum output limit of generator,the level and distribution of spinning reserve,the turbine-governor system and load,etc. Compared with the PSS/E simulation,the proposed method can predict the frequency dynamic and the minimum frequency of power system after disturbance more quickly and accurately,with better generalization ability and practicability. Furthermore,the model trained by the v-SVR method can be applied to the online assessment of power system frequency security and stability,based on which,the appropriate emergency control measures are set to prevent the collapse of power system frequency.