Data-driven Method of Renewable Energy Based on Generative Adversarial Networks and EnergyPLAN
Yang Liu, Huaguang Zhang, Yunfei Mu · 2021
This paper proposes for the first time to combine the concept of deep learning to analyze the data under the energyPLAN platform. In the process of data simulation, the Generative Adversarial Networks (GAN) algorithm is introduced to compensate for the missing data of individual energy indicators, so as to realize the intelligent monitoring and renewable energy. It is significant for secondary utilization of energy. Regarding innovation of Energy Management Module (EMM) and availability of renewable energy development, we have achieved phased results, but there are still challenges: (1) The feature extraction of the data is broken down or stolen, and the distortion problem occurs. The original data of the training sample is abnormal, and the integrity of the data needs to be expanded; (2) Large-scale data relies solely on simulation. There are still feasibility problems, and further verification is needed. In the article, the realtime power data and simulation collected by the international platform energyPLAN are used to verify the deep learning, and the compensation data is integrated into the verification at the same time.