Abstract: Considering the adverse effects of fluctuating output power of distributed power sources on fault recovery,a dynamic fault recovery scheme for distribution networks using photovoltaic output power prediction is proposed. After the fault occurs,the dynamic output power of photovoltaic system is predicted,and load dynamic characteristics are studied. Then,based on the objectives of maximizing the total important load for distribution network recovery,minimizing fault recovery costs,and minimizing fault-induced power outages,an optimal fault recovery model based on dynamic tree knapsack model is constructed. This model can quickly adapt to changes in the distribution network topology. An improved simulated annealing genetic algorithm is used to solve the optimal fault recovery plan,which improves the algorithm’s convergence speed and stability. Finally,the simulation is conducted in IEEE 69-node system,and the results show that the proposed algorithm can adapt to the fault recovery of complex distribution networks,improve the reliability of distribution network power supply,and reduce the recovery cost and power outage loss caused by distribution network faults.