Online Estimation of Plant Participation Factors for Automatic Generation Control in Power Systems With Variable Energy Resources

Abstract

The article discusses the limitations of existing automatic generation control (AGC) systems that appear under the impact of variable energy resources. To overcome identified issues, the authors proposed an approach that advances the functional block responsible for computation of plant participation factors (PF). This approach connects an optimizer with a component for power flow calculations and allows online estimation of plant PFs to increase flexibility and selectivity of AGC. The corresponding optimization models were established to perform conventional and advanced control strategies. To meet performance requirements imposed by variable energy sources, the machine learning (ML) model, namely the densely connected neural network, was designed for power flow calculations. Besides, Lasso regression method was proposed to select relevant features for the considered control tasks and improve the performance of the machine learning based power flow model. Finally, the software tool was developed to implement the proposed approach and tested on a model of real 60 GW interconnection containing 464 nodes and 742 branches. The results of the software testing confirmed its feasibility and easy integration into existing AGC systems.