• A novel model to multi-area power system with high penetration of IBRs is proposed. • A model-data-fusion-based method to linearize the FN constraint is proposed. • The linear relationship between FN and generator status is captured via DNN. • Wind power uncertainty is considered in the two-stage stochastic FCUC model. With the rapid increase of renewable energy penetration and HVDC lines, modern power system evolves into a system with low-inertia and weak frequency support. This paper proposes a novel model to multi-area power system with high penetration of IBRs and a model-data-fusion-based method to linearize the FN constraint for the frequency-constraint unit commitment (FCUC) model. The multi-area SFR model incorporates various types of IBRs, including long-distance VSC-HVDC and offshore wind farms with unified frequency regulation strategy. Then, the deep-neural-network-based method establish an accurate linear relationship between the post-disturbance frequency nadir (FN) and the generator status variable. Considering the wind power output uncertainty, an FCUC model is proposed based on the two-stage stochastic programming. Numerical studies indicate that the proposed method can approximate the FN with high accuracy and efficiency. Furthermore, the FCUC formulation effectively ensures that sufficient generators are kept online to provide adequate inertia and primary frequency regulation in each area.
Shen et al. (Wed,) studied this question.