A predictive engine for on-line optimal microgrid control

This research presents a predictive engine that integrates into an on-line optimal control planner for electrical microgrids. This controller models the behavior of the underlying system over a specified time horizon and then solves for a control over this period. In an electrical microgrid, such predictions are challenging to obtain in the presence of errors in the sensor information. The likelihood of instrumentation errors increases as microgrids become more complex and cyber threats more common. In order to overcome these difficulties, details are provided about a predictive engine robust to errors.