Contributions on DC microgrid supervision and control strategies for efficiency optimization through battery modeling, management, and balancing techniques

Identifier:  TDX:3181

Handle: https://hdl.handle.net/20.500.11797/TDX3181

Authors:  García Elvira, David

Abstract:
This dissertation presents a set of equipment, models and control strategies, that have been developed with the final goal of improving the operation of a DC microgrid. Two control strategies are proposed to improve the efficiency of the DC-DC converters that interface the microgrid’s power units with the DC bus. The first strategy is centralized Bus Voltage Optimization Control, which manages the power of the microgrid’s Battery Energy Storage System to make the bus voltage follow the optimum voltage dynamic reference that minimizes the converters’ losses. The second control strategy is Online Optimization of Switching Frequency, which consists in locally operating each converter at its optimum switching frequency, again minimizing power losses. The two proposed optimization strategies have been validated in simulations. Moreover, a new converter-based active balancing topology has been proposed and its control strategy has been designed. This equalizer topology consists of a single DC-DC converter that performs cell-to-cell charge transfer employing power routing via controlled switches. The control strategy of the equalizer has been designed to achieve rapid SOC balancing while avoiding imbalance overcompensation. Its performance has been validated in simulation. Finally, a simple degradation model of an NMC battery cell with graphite negative electrode is proposed. The model combines the simplicity of an equivalent circuit model, which explains the fast dynamics of the cell, with a physical model of the Solid-Electrolyte Interphase (SEI) layer growth process, which predicts the capacity loss and the internal resistance rise in the long term. The proposed model fine-tunes the capacity loss prediction by accounting for the incorporation of unused lithium reserves of both electrodes into the cyclable lithium range to reach the OCV limits after the side reaction has consumed cyclable lithium. The SEI degradation model can be used to perform predictive control of batteries oriented toward extending their lifetime.