Buildings' Transition to Active Nodes: Assessing the Viability of DC Distribution, PV and Battery Storage

Doktorsavhandling, 2024

Historically, buildings have been passive nodes in the electric grid system with one-way power flows. However, with the recent market development of solar photovoltaic (PV) and stationary behind-the-meter battery storage systems, buildings are now transitioning to active nodes, offering bi-directional power flows. Various system topologies and modelling aspects are of interest for these active nodes and their viability. This thesis compares internal building direct current (DC) distribution with the conventional alternating current (AC) distribution for single-family and office buildings. For both building types, the geographical location is altered to examine the effect of PV and load correlation on the DC performance. The energy loss over a year and the loss distribution across various components are examined for three DC topologies, including one with constant power electronic converter (PEC) efficiencies, to quantify the loss discrepancy to experimentally derived PEC efficiency characteristics. Using constant efficiencies for a single-family building underestimates the annual losses by 34% (63 kWh/a). With load-dependant PEC efficiencies and without battery storage, DC operation shows no performance enhancements compared to AC. Depending on the studied case, DC operation results in loss savings, -16.3 to -43.6% with PV and battery.
 
Two methods are proposed to reduce the grid-tied converter (GC) losses from partial load operation. One method–a modular GC design consisting of a smaller and a larger GC–is modelled for two cases: a single-family building and an office building, and presents an optimal GC size configuration of 15/85%. The loss savings relative to AC operations for a 15/85% configuration are 26% for the single-family building and 15–40% for the office. The savings depend on the office's location and system design (PV and battery sizing). For the offices, the effect on DC loss savings is examined via a parametric sweep by varying PV and battery sizes, with resulting savings up to 40% (-12.8 MWh/a) compared to AC operation. The results highlight the effect of GC sizing on the DC performance, the effect of battery storage, and how the PV and load correlation affects the DC performance.
 
Furthermore, a battery model is derived from experimental measurements of the cell's current–resistance and open-circuit voltage (OCV)–state-of-charge (SOC) dependencies. The battery model is verified against the measured voltage with good compliance (RMSE<7 mV). Three representations–including the round trip efficiency approximation–are compared for annual battery system losses. The results indicate that the cell's losses–making up 22–45% of losses for the examined case–and that the internal resistance's current dependency is essential for an accurate representation. The loss discrepancy for the round trip approximation varies between -5% to 29%, relative to the experimentally derived representation, depending on the modelled battery size.
 
The role of PV and battery storage for an airport micro grid is examined in a forward-looking case with electric aviation (EA) and electric vehicles (EVs). Seven scenarios are studied, including four with battery storage and different operation algorithms. One of the algorithms is a novel operation combining self-consumption (SC) and peak power shaving. Compared to the current situation, the techno-economic evaluation shows a significant increase in energy (89.4%) and power (+1 MW) demands from EA and EV. For the nominal battery price and peak power tariff (Ct), the novel operation shows the shortest Payback Period (PBP) of 4.8 years for the battery scenarios. With varying battery prices and peak power tariffs, the sensitivity analysis shows that Ct can significantly affect the PBP.
 
Lastly, the effect of PV module operating temperature on performance is empirically evaluated and quantified for seven arrays from annual operation. For the Building–Applied PV (BAPV) c-Si modules, the elevated operating temperature adds 1% to the total losses and 2% for the c-Si Building–Integrated PV (BIPV). Examining the results of SC and self-sufficiency (SS) verifies the correlation between SC and power rating and introduces the correlation between SS and annual yield, considering the effect of system design, level of roof integration and PV cell type. For this case study, comparing two systems with and without battery storage shows the weekly variation in SS and SC and highlights the drawback of single-objective dispatch.