Mechanism-Guided Data-Driven assessment method for Multi-Scenario Short-Term voltage stability

• A systematic classification framework for voltage instability scenarios is established. • Identify key features indicating voltage stability in renewable energy power system. • A mechanism-guided data-driven voltage stability assessment method is proposed. • Addresses model-driven modeling dependence and data-driven interpretability limitations. With the increasing integration of renewable energy sources and the expansion of direct current transmission systems, power system modeling has become increasingly complex, and the causes of voltage instability have grown more diverse. This complexity challenges the accuracy and timeliness of traditional voltage stability assessment methods. Therefore, this paper proposes a mechanism-guided and integrated data-driven for short-term voltage stability multi-scenarios assessment method. Initially, traditional voltage instability inducers are classified into two categories: those dominated by active power and those dominated by reactive power. Subsequently, the mechanisms through which renewable energy and direct current transmission influence voltage instability are analyzed. And the characteristics of voltage instability under active dominance as well as reactive dominance in the new scenario are sorted out. Building on this foundation, a multi-scenarios voltage stability assessment method is proposed using the transformer algorithm. The proposed method is able to accurately classify the voltage instability caused by different reasons, so as to provide the correct action instructions for the subsequent voltage regulation. The efficacy of the proposed method is validated through simulations conducted on an IEEE 10-machine 39-node system and a 100-node system from a regional power grid in China. The simulation results show that the method proposed in this paper has good accuracy and generalization.