A Fully Distributed Robust MPC Approach for Frequency and Voltage Regulation in Smart Grids with Active and Reactive Power Constraints

Shortly, power distribution grids will incorporate large amounts of distributed energy resources and flexible loads, allowing the operation of a portion of the network in islanded mode to increase the reliability and resilience of the whole power system. A fully distributed robust model predictive control (MPC) strategy for voltage and frequency regulation in interconnected distribution grids is stated. Each grid node represents a collection of prosumers with a large active and reactive power regulation capacity. The advantages of this approach rely on the capability to afford any type of uncertainties, without making any assumption on the probability density function, on distributed generation and load nowcasting. We propose a two-stage architecture: at the first stage, an MPC approach, based on the distributed alternating direction method of multipliers (dADMM), is performed, considering the data nowcasting; instead, the second stage (based on robust distributed team decision theory) takes as input the trajectory of the first stage to compensate the noise that affects the system. The developed architecture has been tested on a modified IEEE5 bus system, considering multiple loads and renewable generation.