With Tesla’s big stationary storage announcement last week, much attention was paid to how battery prices have fallen faster than expected. The $350 per kWh price for the 10 kWh battery was ahead of most analysts’ forecasts. Not surprisingly, the emphasis has been on using those batteries—at least initially—for backup power and demand charge reductions, but as SolarCity CTO Peter Rive recently noted, they’ll also soon be used for grid-connected solar customers. That grid-connected solar-plus-battery combo is exactly what Rocky Mountain Institute and HOMER Energy analyzed in The Economics of Load Defection released last month.
But just as important as how cheap batteries are getting is another crucial question: how much battery do you need?
Understanding the role of storage and how to size a battery bank is not simple. It is one of the reasons we developed the HOMER software for modeling hybrid renewable systems, including integration of battery banks. Are you using batteries for critical loads or all loads? Do you need battery backup power for hours, or days, or weeks? Is the system grid-connected or off-grid? These merely scratch the surface of the many considerations that go into modeling batteries. And how you answer them can result in vastly different battery estimates. For example, the oft-cited Konterra microgrid in Maryland has enough battery backup to power critical loads for about four hours if the broader grid goes down. By contrast, a recent report from Moody’s estimates that off-grid solar-plus-battery systems would need a whopping (and absurdly unrealistic) 60 days of storage.
BATTERIES FOR SELF-CONSUMING ROOFTOP SOLAR
Consider how batteries might interface with a grid-connected, self-consuming rooftop solar PV system, such as you might find in places without net metering. Without net metering or batteries, the solar system would be sized based on the mid-day loads when solar generation is greatest, because any solar energy in excess of those loads would go to waste. Since most people aren’t home during the day, their mid-day loads would be smaller and thus the optimal size of the solar array would be quite small.
Cost-effective batteries—like those modeled in The Economics of Load Defection—would allow much larger solar arrays. Still, it is not an obvious sizing decision because energy consumption varies from day to day and month to month and the solar production varies even more. The scenario requires sophisticated modeling to optimize based on chronological simulations.
BATTERIES FOR BACKUP POWER
Backup power for reliability and resilience is another rationale for batteries. Solar PV systems without batteries do not protect consumers from utility outages. This also presents a challenging sizing problem depending on how much reliability is desired. For this application the size of the battery bank depends on how long an outage you want to protect yourself from. For short outages, the sizing problem is not too difficult, but the same system that could provide power indefinitely in the summer (short of running the air conditioner) may only be good for a couple of days in the winter (thanks to longer stretches of cloudy, stormy days that inhibit solar production).
BATTERIES FOR OFF-GRID SYSTEMS
Most off-grid systems larger than 1 kW have a backup generator. HOMER analyses have consistently shown a great reduction in battery sizing with a backup generator, even if it is only used very occasionally. Backup generators have numerous drawbacks, especially if they are used more than occasionally, but consumers can successfully avoid the use of a backup generator if they are able to carefully manage their loads.
Understanding what electricity is used for and how much electricity different appliances require is crucial for anyone trying to live off-grid, even if it is just for an extended utility outage. Most high-tech appliances use very little power—LED lighting, portable electronics, and efficient televisions use less than a total of 2 kWh per day. Refrigerators, furnace fans, and pumps are larger loads that can use up to 2 kWh per day each. Heating, hot water, clothes drying, and cooking use a lot energy, but they can be served with natural gas or propane (though they do run on electricity in many homes and thus can be a significant consideration for some). Air conditioning is the one appliance that requires electricity and in large amounts. In dry regions of the West, evaporative coolers can be used that don’t require a lot of power, but they are useless in humid climates.
THE BATTERY SIZE CONUNDRUM
So, how big do battery banks need to be? It depends. Can you use the utility as a “battery?” Is your goal to reduce your consumption to zero or to find the best return on your investment? Are you worried about outages? If so, are they likely to be short or long? Can you be flexible about how much energy you use during an outage, especially air conditioning? Are you willing to occasionally use a backup generator? The answers to those questions affect your sizing decision in a big way.
If you only use gas appliances for the large loads during an outage and have a typical-sized PV array, then 7–10 kWh of batteries—like that of the Tesla Powerwall—is sufficient to get through most outages of most any length. If you are worried about blizzards, you might want a backup generator even though you might only use it a couple of hours per year. If you can’t imagine living without air conditioning, you will need a much larger solar array, and the trade-off between sizing the solar array, the battery bank, and your actual cooling needs becomes critical.
Although storage prices are dropping rapidly, correctly sizing the storage for the location, usage pattern, and other design constraints, such as grid reliability and rate structure, is essential to creating systems that make economic sense. We strongly caution against “back of the envelope” or “rule of thumb” estimates for such an important decision.
Peter Lilienthal, Ph.D., is CEO and founder of HOMER Energy LLC and original developer of the HOMER modeling software. HOMER Energy provides the HOMER software, training, analytical services, and community market access tools to professionals in the energy industry who desire to analyze and optimize distributed power systems, microgrids, and systems that incorporate high penetrations of renewable energy sources.