LAUSR

In this brief, an online learning-based adaptive secondary controller with extended state observer (ESO) is proposed for regulating the voltage and allocating the reactive power in an islanded microgrid. With the active disturbance rejection properties of the ESO technique, two delicately designed neural networks via online learning approaches are proposed to compensate for the unknown/uncertain dynamics (UD) in the microgrid and attenuate the impact of external disturbance. Moreover, a vector of switching factors is introduced to unify the voltage recovery of the critical bus and accurate reactive power-sharing into a single framework. It is also worth noting that the devised approach is deployed in a distributed fashion, which enables the microgrid to acquire the plug-and-play capability. Extensive simulations are conducted to verify the effectiveness of the proposed controller for several cases including islanding and sudden load changing.