부산대학교 도서관

User-side energy demand forecasting is one of the essential functions of smart grids in the future. By forecasting electricity consumption from consumers, energy companies can make appropriate adjustments to the amount of electricity generated, reducing the burden on the grid and reducing energy waste. In this paper, an optimization method for energy prediction method based on neural network is proposed. Firstly, a dual-CNN structure is used to extract the state of the electrical equipment at the user and the power required for the state and predict it. Then, the neural network-based genetic algorithm was used to adjust the weight of the neural network. Both methods are more mature methods, and the combination of the two methods may be able to build on the original methods and achieve accurate prediction of energy demand.