Effects of time resolution on finances and self-consumption when modeling domestic PV-battery systems

When modeling a renewable energy system, the timestep to use is an important consideration. Timestep, or time resolution, can have an impact on results, influencing the sizing of the system and whether or not to invest at all. In this work, real measured data for an entire year at 15-s resolution from a rooftop PV array and 8 household loads in the UK are used. The PV and load time series are averaged to lower resolution: 1-min, 5-min, 30-min and 1-h, and the results from using them as input to a 25-year simulation of PV-only and PV-battery systems are compared to the 15-s resolution results. Load resolution is confirmed to be more important than PV resolution for improving accuracy of self-sufficiency and cost metrics; the presence of a battery is confirmed to reduce the errors of using low resolution compared to PV-only. However, these findings only apply to the commonly tested Greedy algorithm but not the newly developed Emissions Arbitrage algorithm. A wider range of metrics are calculated here than in previous work, finding consistency in that low resolution overstates the benefits of PV-battery, but variation in percentage difference across the metrics used. Further aspects not studied before include: the diminishing returns in computation speed when time resolution is lowered, and the effect of time resolution on the tipping point when certain configurations become more attractive propositions than others. Time resolution of input data and modeling are issues not only for researchers in academia and industry, but from a consumer protection perspective too.