POWER SMOOTHING OF PARALLEL GENERATORS USING MODEL PREDICTIVE CONTROL

Abstract

Improving energy efficiency is an ongoing process throughout the energy sector, and one method to accomplish this is by smoothing the power output of a generator. The goal of smoothing the power output is to reduce the amount of microtransients present during a load shift, which reduces instantaneous losses and allows for efficiency gains. This thesis investigates the use of a model predictive controller (MPC) on a two-generator, parallel-connected microgrid to accomplish power smoothing using an energy storage device. A steady-state state space model employed within the MPC recursively determines an optimal control input of source current between the generator and the battery for load changes. This model proved to be successful in smoothing the power transient from the generator and was expected to decrease the fuel consumption when compared to conventional droop control. Ultimately though, an improvement in fuel efficiency proved to be inconclusive with the steady-state model used and is likely due to an incomplete transient model of the generator.