Behavior Trees for Smart Grid Control

The goal of increasing the share of renewable sources in energy supply leads to a decentralization of generation and growing in-feed into lower voltage levels. The integration of renewables therefore requires a substantial transformation of the existing distribution grids. This, the greater complexity due to new actors like prosumers and ongoing adaptation to regulatory requirements calls for new and more flexible approaches to distribution grid control. Behavior Trees are a proven and well-established technique for describing control logic and are regularly used in the gaming and robotics industry to implement artificial intelligence. That makes them a promising option for the high-level control of Smart Grids. This paper describes the implementation of Behavior Trees into a newly developed control framework for Smart Grids. Benefits and weaknesses are discussed in comparison to the previous conventional approach.