Decentralized Discrete-Time Adaptive Neural Network Control of Interconnected DC Distribution System

In this paper, the interconnected dc distribution system is represented as a class of interconnected, nonlinear discrete-time systems with unknown dynamics. The dc distribution system comprises several dc sources, here called subsystems, along with resistive and constant-power loads (CPLs.) Each subsystem includes a dc-dc converter (DDC) and exploits distributed energy resources (DERs) such as photovoltaic, wind, etc. Due to the power system frequent disturbances this system is prone to instability in the presence of the DDC dynamical components. On the other hand, designing a centralized controller may not be viable due to the distance between the subsystems (dc sources.) Therefore, in this paper the stability of the interconnected dc distribution system is enhanced through decentralized adaptive nonlinear controller design that employs neural networks (NNs) to mitigate voltage and power oscillations after disturbances have occurred. The adaptive NN-based controller is introduced to overcome the unknown dynamics of each subsystem's converter and stabilize the entire grid, assuming that only the local measurements are available to each converter. Simulation results are provided to show the effectiveness of the approach in damping oscillations that occur in the presence of disturbances.