Model predictive control for reactive power management in transmission connected distribution grids

In this work a multi-objective model predictive control for reactive power management in transmission connected distribution grids with high share of wind power is presented. The proposed control utilizes reactive power capabilities of wind farms and tap-changer positions in order to improve distribution grid operation. Control signals namely tap-changer positions and reactive power set-points are smoothed over the forecast horizon. Further possible optimization objectives are power loss reduction, voltage profile smoothing and complying with reactive power exchange limits with the transmission grid. A mixed-integer non-linear optimal power flow problem (MINLP-OPF) is formulated incorporating grid operation limits. The performance is evaluated on a real German 110-kV distribution grid with 1.6 GW wind power for one year. With the proposed control, reactive power exchange within allowable limits is increased from 58.3% to 94.5%, compared to a reference operation where only tap-changer positions are utilized for loss reduction with a single time-step optimization.