Two-step control of multiple DERs based zero emission power grid using DNN enabled terminal sliding mode control

This research paper proposes a novel two-step control for a zero carbon power grid, which combines the use of deep neural networks (DNNs) and global terminal sliding mode control (GTSMC). The proposed power grid is composed of several distributed energy resources (DERs), including wind and photovoltaic (PV), and a hydrogen energy system. The two-step control configuration first employs DNNs for maximum power point tracking (MPPT) reference generation, followed by GTSMC based tracking controller. The control system's stability is evaluated using Lyapunov stability criteria, and its performance is assessed through simulations in MATLAB with real-world commercial and residential loads of 2-12 KW and weather profiles from a wind farm in Jiangsu, China.