基于数据驱动的电网日内临界惯量预测方法OA

Critical inertia is a key index to evaluate frequency safety and stability of the power system,which can only be obtained by analytical solution,but this method involves a nonlinear process and can't adapt to the calculation needs of complex power grids with a large number of units.In view of this,this paper proposes a data-driven method for predicting critical inertia of the power grid.A prediction model is established based on the deep belief network,and by analyzing the multiple factors affecting the critical inertia,multiple types of variables such as random power shortage,inertia time constant affected by the change of power generation plan and system load level are set as input eigenvalues,and the critical inertia is used as the output variable.Multiple types of power shortage prediction faults are traversed and set,corresponding to 96 periods of intraday power generation plan of the power grid,as well as.Meanwhile,combined with PSS/E and Python time-domain simulation,initial sample data is generated and the critical inertia sample dataset under multiple operating scenarios is constructed.Afterwards,the critical inertia of the power grid through joint training and testing sample sets is predicted.Finally,the improved IEEE 39 node example system is used to verify the effectiveness of the prediction method and three statistic indicators are used to evaluate the prediction accuracy.