Model Predictive Control for Current Sharing and Voltage Balancing in DC Microgrids

In this paper, a Model Predictive Control (MPC) technique is presented for the control of energy sources in a dc microgrid. Proportional current sharing is achieved, where the current of each source is shared appropriately by the weight terms. Similarly, a cost function is presented to achieve average voltage of all the source nodes in the network. In addition, a Luenberger observer is designed to estimate the unknown states and disturbances. The centralized MPC optimization problem is then placed in quadratic programming form and solved using the mpcqp solver, which allows for code generation and real time implementation using an embedded system. A four node dc microgrid model is used as a case study. Real-time simulation using Opal RT is shown demonstrating the feasibility of online implementation of the controller.