An online learning algorithm approach for low voltage grid management

The smart grid concept introduces improved possibilities for coordinated distribution grid management in order to increase the receptivity for Renewable Energy Sources while simultaneously guaranteeing a safe and reliable grid operation. This paper presents a smart grid control strategy for real-time low voltage (LV) grid management applications based on an online-learning algorithm. It enables for the derivation of a schedule-forecast for installed assets. Next to coordinated voltage and line utilization control the approach optimally exploits the potential benefits of innovative grid assets for grid operation. The performance of the algorithm is demonstrated by a simulation study using a typical LV grid.