MIT responds: batteries are critical to the future electric grid, but they have their limits


I enjoyed pv magazine’s article (“MIT sees wind+solar going big, dreams of nuclear, forgets batteries exist”) on our paper about the role of nuclear power in deeply reducing carbon emissions from the electricity sector. But I assure you, we didn’t forget that batteries exist.

Energy storage is a complex combination of total capacity, duration, flow and cost. The article alludes to cost-effective batteries that might store power from the sun during the day to meet power needs at night. This may be close to the ideal system for battery storage, especially in a sunny location near the equator.

With likely lower power demands at night, one would have a predictable power supply during the day, not quite matching the tendency for usage to skew into the evening. Moreover, supply would be even throughout the year: you would only need enough storage from midday to meet significant evening demand, with a bit for low nighttime demand. In fact, our modeling of storage accounts for cost, capacity, duration and flow, the critical dimensions of any storage technology.

In terms of the results we get, the technology could be battery storage, pumped hydro or any other storage with those specifications. We used pumped hydro to cost our storage technology out, because it was the least costly option today. Whatever the storage option, it will have some cost that must be taken into account. And that cost will tend to rise the more capacity and longer storage required.

The problem , however, is that virtually no place in the U.S. is as ideal as the above example — certainly California and the desert Southwest come close, but even there the winter comes with shorter days, a lower sun angle and some cloudy days.

So here are some real numbers: I have a solar PV installation on my roof in Maine, along with a Tesla Powerwall able to store 12 kWh. My annual production has been 10 to 15% below what the system was supposed to produce.

More problematic for storage, I’ve produced 1.2 to 1.3 MWh in the late spring/summer months, but less than .1 MWh last Jan and Feb.— since the roof was covered with snow a good chunk of the time (see figure).

Even if it were sunny 31 days out of the month, the best I could generate through the winter months would be about .5 MWh per month, just due to the shortened day length and low sun angle. Unfortunately, the winter months are when I am using 1.3 MWh per month. I can set my Powerwall on self-power mode and hardly use power from the grid during the summer—the solar/battery example cited in the pv magazine article—except I’d be producing (and spilling) a huge amount of power that I couldn’t use.

Fortunately, I am connected to the grid and can carry credits from summer to winter. My maximum banked generation this year was 1.9 MWh. If I were to install enough Powerwalls to store the extra generation I produced over the summer, I would need about 160 of them. At $12,000 each that would be almost $2 million of investment in storage just for my single residence. Battery costs will have to come down a lot to make that calculation work in favor of storage.

The Powerwall makes sense for me to protect against outages—the $12,000 less the federal tax credit makes the battery competitive with a hard-wired generator. A photovoltaic solar system connected to the grid is still a good deal for me—I estimate an internal rate of return of 4% to 6% depending on whether the modules only last for the 25-year warranty or last another 10 beyond that.

But this investment calculation would fall apart if I was not hooked into the grid, where I take advantage of a power supplier with dispatchable power.

Admittedly, Maine is probably the other extreme from the desert Southwest, and grid scale operations are going to gather economies of scale I can never hope to get with a small residential installation. That’s why we constructed a model to carefully take into account the real variation in patterns of demand and supply over the course of hours in a day and over seasons—and also to exploit, through grid connection, favorable geographical patterns of different renewable supplies and patterns of demand, using what we believe are reasonable utility-scale costs.

Obviously, this provides a great advantage over the fully-off-the-grid calculation for my house, but it is hard to erase all of that supply and demand mismatch.

Finally, please note that our study finds that deep decarbonization without low-cost nuclear is entirely possible but is less expensive if we have a reliable, low-carbon, dispatchable generation option at a reasonable cost.

The views and opinions expressed in this article are the author’s own, and do not necessarily reflect those held by pv magazine.

This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: