Retracted: A Modified Balancing Approach for Renewable Based Microgrids Using Deep Adversarial Learning

This paper basically investigates the enhancement of a modified non-square direct matrix converters (MC) for renewable microgrid applications using generative adversarial networks (GAN). Such a deep learning technique which benefits from two multi-layer adversary perceptions is used to enhance the security of the data in the microgrid. This paper also provides a new application of the non-square direct MC in the microgrid system which is able to provide balanced output with any desired amplitude and frequency under unbalanced condition specifically in the case of using renewable energy sources such as photovoltaics (PVs) and wind turbines. In addition, a new modified social spider optimization (MSSO) algorithm is introduced to help improving the training process of GAN. Simulation results show that matrix converter based on GAN makes it possible to convert any input voltage to the desired output voltage which leads to the elimination of the back to back converter of wind turbine.