Abstract:
In the view of the scheduling optimization problem of microgrid, a novel microgrid model that consists of a wind turbine generator, an energy storage system, a set of thermostatically controlled loads and a set of price-responsive loads is used as the research object. With the goal of achieving the minimum economic operating cost, and considering the impact of the volatility and randomness of wind power generation on the safe and economic operation of the microgrid, based on the framework of AC algorithm, an improved A3C algorithm is proposed. Asynchronous training is realized by adopting multi-threading method. The experience replay mechanism of DQN algorithm is improved from uniform sampling to importance sampling, and is added to the training of A3C algorithm. The DQN, AC and improved A3C algorithms are trained and simulated respectively and compared. Simulation shows that the utilization rate of samples is improved and the training time is reduced by improved A3C algorithm,whose training time is 1.46 minutes. When the wind power fluctuates, the model trained by the improved A3C algorithm controls the charging and discharging of the energy storage device according to the real-time electricity price of the main grid. The scheduling strategy given in the model improves the economic benefits and effectively reduces the impact of wind power fluctuations on the microgrid. The proposed scheme can provide reference for intelligent dispatching of microgrdid.