The operation of residential energy hubs with multiple energy carriers (electricity, heat, mobility) poses a significant challenge due to different carrier dynamics, hybrid storage coordination and high-dimensional action-spaces. Energy management systems oversee their operation, deciding the set points of the primary control layer. This paper presents a novel 2-stage economic model predictive controller for electrified buildings including physics-based models of the battery degradation and thermal systems. The hierarchical control operates in the Dutch sequential energy markets. In particular common assumptions regarding intra-day markets (auction and continuous-time) are discussed as well as the coupling of the different storage systems. The best control policy it is best to follow continuous time intra-day in the summer and the intra-day auction in the winter. This sequential operation comes at the expense of increased battery degradation. Lastly, under our controller, the realized short-term flexibility of the thermal energy storage is marginal compared to the flexibility delivered by stationary battery pack and electric vehicles with bidirectional charging.

In the context of building electrification, the operation of distributed energy resources integrating multiple energy carriers (electricity, heat, mobility) poses a significant challenge due to the nonlinear device dynamics, uncertainty, and computational issues. As such, energy management systems seek to decide the power dispatch in the best way possible. The objective is to minimize and balance operative costs (energy bills or asset degradation) with user requirements (mobility, heating, etc.). Current energy management uses empirical battery ageing models outside of their specific fitting conditions, resulting in inaccuracies and poor performance. Moreover, the link to thermal systems is also overlooked. This paper presents an ageing-aware nonlinear economic model predictive controller for electrified buildings that incorporates physics-based battery ageing models. The models distinguish between energy storage systems (chemistry, ageing state, etc.) and make explicit the trade-off between grid cost and battery degradation. The proposed algorithm can either cut down on grid costs or extend battery lifetime (electric vehicle or stationary battery packs). Additionally, substituting NMC cells with LFP chemistries optimizes grid performance during the summer, yielding a 10% grid cost reduction and a 20% decrease in degradation. Finally, the grid cost and degradation of the presented MPC when using aged batteries are improved with respect to the state of the art by 10% and 5% respectively, in periods with high solar generation and low thermal loads like summer.

Home Energy Management Systems (HEMS) play a key role in the electrification of the residential sector. In a future where different energy carriers must be converted and used seamlessly, HEMS are responsible for orchestrating all energy assets within a home. To determine the power flows, the HEMS must be aware of its impact on the aging of the system's batteries. On the other hand, since this is typically performed through optimization models, various variables are needed, such as ambient temperature, solar radiation, and initial conditions. These inputs are uncertain and require proper forecasting. This paper presents the operation of an aging-aware HEMS using probabilistic forecasts of exogenous information, directly relating them to the quality of the degradation control. The relationship is studied under deterministic and stochastic optimal control settings.