A Machine Learning Based Optimized Energy Dispatching Scheme for Restoring a Hybrid Microgrid
Karim, MA; Currie, J; Lie, TT
Abstract
A microgrid operated in stand alone mode is highly vulnerable to instability when the integration of intermittent energy sources are considered. If a short circuit fault occurs in a microgrid while operating at its design limit, often cost effective system recovery becomes a challenging task. Under such contingencies predictive analysis can be used to strengthen the system restoration schemes. In this study, a system based on machine learning algorithm is implemented to forecast the security of a standalone microgrid and based on the forecasting, schedule multiple backup diesel generators under the contingency of loss of a major generating unit. The underlying objective is to maintain the voltage stability with an optimized economic dispatch scheme, right after clearing a critical three phase short circuit fault. Finally, a promising set of outcomes are observed and discussed.