Off-policy deep reinforcement learning with automatic entropy adjustment for adaptive online grid emergency control

•A novel DRL-based voltage stability control algorithm with automatic entropy adjustment (AEA) for grid emergency control is proposed.•An advanced actor-critic structure based on a maximum entropy DRL framework is developed to encourage a broader exploration of action-state space for adaptivity to high uncertainties in dynamic operating environments.•Comparative studies with the existing DRL-based control methods in IEEE benchmarks indicate salient performance improvement of the proposed method for dynamic system emergency control.