Soft Actor-Critic-Based Early Warning Method for Quench Risk in High-Temperature Superconducting Cables

To address the challenge of predicting quench risks in high-temperature superconducting (HTS) cables under multiple-source disturbances, this paper proposes an intelligent early-warning method based on the Soft Actor-Critic (SAC) algorithm. At first, a multi-index evaluation system covering electrical and cooling characteristics is constructed, and the entropy weight method is adopted to determine the weights for comprehensive risk quantification. Secondly, a deep reinforcement learning model based on SAC is established to achieve dynamic risk prediction under continuous operating conditions. Finally, multi-scenario experiments and comparative analysis with other mainstream early-warning methods demonstrate that the proposed approach achieves effective early quench warning, with prediction errors consistently maintained within acceptable ranges.