Economic Model Predictive Control for DC Microgrids (EMPC-DC)

Energy transition is one of the main pillars for mitigating climate change problems. DC microgrids are an excellent solution to integrate renewable energy since the nature of most renewable sources and energy storage devices is DC. However, they present significant challenges due to the variability of the generation sources and loads, besides the non-linearity introduced by power electronic converters, which allow the integration of renewables into the microgrid. A hierarchical control scheme is usually employed, consisting of primary, secondary, and tertiary control. Primary control is responsible for granting a stable operation. Typically, a droop control is used. Secondary control brings the system to a reference; while the tertiary control is related to the market, including an EMS that brings economic concerns for optimal microgrid operation.

The currently proposed strategies are composed of many layers of control, involving communications between each of these layers, which makes this type of strategy costly and complex. Likewise, a hierarchical technique can be prone to stability failures due to interactions between higher stages and lower stages. Moreover, using linear techniques instead of nonlinear models results in accuracy and stability errors when the operating point shifts. Therefore, it is desirable to reduce the number of stages in the hierarchy and implement a non-linear approach. A proposed MPC facilitates the unification of the secondary/tertiary stages while preserving system non-linearities.

In this report, we present MPC applied to an islanded DCMG. The main objective of the control is to achieve power-sharing while maintaining voltage levels within the set limits. Two approaches to power-sharing were implemented by simply changing the objective function of the optimization problem. First, a PPS objective, which distributes the power among all converters according to the equal percentage of available power capacity (same factor plant), thus ensuring a longer lifetime of all converters. Secondly, an economic dispatch based on the generation cost associated with each converter, reduces the consumption cost for the end user. For both experiments, they were implemented in Power Hardware in the Loop (P-HIL), where Opal-RT, a power amplifier, and resistive loads were used.

The results obtained show the practical feasibility of the implementation of the proposed control system, demonstrating satisfactory results in its development. A review of the state of the art related to the MPC for DCMG is included. Also, Figures of the plant factor and stresses are presented, together with connection diagrams explaining the experiment development process.