C18 -- Performance Evaluation of 75 Residential Rooftop Solar Photovoltaics (PV) and Battery Systems: A Cold-Climate Comparison for Time-Based Control and Backup Reserve Modes of Operation

Residential energy generation with lithium-ion battery energy storage offers homeowners the potential for energy independence, lower electricity bills, increased resilience, and reduction in carbon emissions. The aim of this study is twofold: (1) in-situ performance evaluation and (2) data-driven simulation (using historical data for training in order to predict year-round scenarios) of 75 grid-connected single-family homes in Atlantic Canada. Each home is equipped with a 5 kW (17.1 kBtu/h) rooftop PV system and 13.5 kWh (46.1 kBtu) battery energy storage. Thirty-eight homes have been programmed for time-of-day electricity rates using a time-based control strategy and the other 37 for resilience (backup reserve strategy) using flat rates. The study provides a rich and diverse sample of single-family homes, including vintage: 1960-2010; primary heating system: electric baseboard, mini-split heat pump, furnace, or wood stove; and battery install location: basement or garage. The analysis uses smart meter, inverter, and battery energy management telemetry data. The first part of the study includes quantification of demand reduction, bill savings, and aggregated energy savings. Winter round-trip efficiency was determined to be 92% for basement and 91% for garage installs, based on weekly round-trip calculations. The second part of the study describes a data-driven Discrete-Event Simulation (DES) framework for residential grid-connected PV and battery energy storage (PV/battery) via a simple energy balance model. Here, we focused on modelling the time-based control battery mode of operation. The model was validated by comparison with in-situ PV/battery data from the PV/battery installations. The purpose of developing the DES model is to provide a highly-accurate in-situ-data-driven tool to estimate year-round PV/battery performance for existing as well as future systems. Results indicate that the DES model is promising, especially for the discharge cycle, while the charge cycle may require calibration.