Nonlinear Model Predictive Control for Microgrid Operation

Armin Nurkanovic

Siemens Corporate Technology and Department of Microsystems Engineering (IMTEK), University Freiburg

Tuesday, February 11, 2020, 11:00

"Room 01-012, Georges-Köhler-Allee 102, Freiburg 79110, Germany "

Microgrids (MGs) are small electrical power grids consisting usually of Distributed Energy Resources (DER), such as small distributed generators (DG) and Renewable Energy (RE) devices, storage devices and loads. Due to the low inertia of the DGs and volatile production of RE devices, ensuring voltage and frequency stability becomes a nontrivial task. State-of-the-art power system control is hierarchical, where each hierarchy operates in a different timescale. In this talk we give a brief introduction to primary and secondary control, which are part of this hierarchy.  Primary control is mainly used to ensures power system stability and secondary control for removal of steady-state off-sets, voltage control and economic considerations. 

In the second part of the talk we present how to derive detailed Differential Algebraic Equation (DAE) based models for MGs, which are letter used for Nonlinear Model Predictive Control (NMPC). Our NMPC formulation allows to consider secondary voltage and frequency control, steady-state equal load sharing, economic goals and all relevant operational constraints in a single optimization problem. The challenge is to control the fast and large dynamic system in real-time. To achieve this goal, we use the recently introduced Advanced Step Real-Time Iteration (AS-RTI) scheme. The controller responds efficiently to large disturbances and mismatches in the predictions and effectively controls the fast-transient dynamics of the MG. Our NMPC approach outperforms a state-of-the-art I-controller usually used in microgrid control and shows minor deviation to a fully converged NMPC approach. The NMPC approach is demonstrated on three different complex MG examples. The talk concludes with presenting several ideas for future research.