Home Energy Management System Optimization Strategy Based on Reinforcement Learning

At present, the household load base number is large and the intelligent level is low. Artificial intelligence technology can provide novel ideas for improving the intelligent degree of home energy management. Household load has great demand response potential, which can provide support for the consumption of renewable energy, participate in the peak regulation and frequency regulation, reduce the peak valley difference, as well as stabilize the fluctuation of power grid. The home energy management optimization strategy based on reinforcement learning is presented in this paper. Firstly, long short term memory is used to predict the output of photovoltaic power and electricity price. And then they are transmitted to the decision-making scheduling model as state variables. On this basis, combining with the load characteristics of household electrical equipment, the Markov decision process model based on reinforcement learning is established, and the optimal scheduling process of home energy management system is related. Finally, simulation examples are designed to verify the effectiveness of the method proposed in this paper. The results show that the proposed methodology can meet user's comfort demand while reducing the power consumption cost.