The first question people usually ask after learning that I have solar panels on my roof is whether I bought a battery backup system as well. This makes sense from a basic off grid viewpoint, but not so much in the real world. Most electrical grids have such low levels of renewable generation right now that they can easily act like huge batteries for the small numbers of people who get solar roofs. For grids like this installing individual battery systems is a waste of resources. In this post I'll try delve into why most solar homes don't need their own energy storage yet and how this will change in the future.
2021 Update: Strong demand for Powerwall’s has made Tesla increase their price yet again. They’re now $7500, and there’s another $1,000 for a “Gateway” that controls it and a crazy $3500 installation cost (I have to imagine you could cut this by DIYing, or just hiring a good local electrician). This makes the Powerwall ROI worse, but hopefully eventually supply will catch up and prices will fall the way the raw price of lithium ion batteries has been falling. https://electrek.co/2021/01/17/tesla-increases-price-powerwall/
Why do You Need a Battery?
A battery can absorb excess solar power when the sun is shining brightly and give it back when it isn't. The most common use case for this happens when the sun sets every night. The size of a battery needed to handle each evening depends on the season and location. A long, cold winter night at my house in PA can draw over 40 kWh of power while a short temperate summer night here can draw less than 5 kWh. If I was in Arizona running the AC on summer nights and very little heat in winter this would be quite different. In Pennsylvania my winter draw is amplified because I use a high efficiency heat pump that runs on electricity. If I heated with another fuel, or had a better insulated house these numbers could be much lower, if I heated with old school resistance electric heat they could be much higher.
Another source of intermittency is clouds that pass by and reduce generation during the day, or sometimes for days on end. For example, there was a cloudy stretch of nearly 3 days last October where my solar generation was consistently below my home's consumption. My house drew nearly 60 more kWh than my panels generated during this period. Smaller passing clouds require a far smaller battery system to handle, often under 5 kWh. The big question is what percentage of solar generation a house can actually use over the course of an entire year. Here's a graph that shows this for my house at different battery sizes.
As you can see with no battery only 27% of the power my panels generate is actually used by my house. The rest goes back to the grid for my neighbors to use. If I add a tiny 6.75 kWh battery this jumps to 42% and a 13.5 kWh pack takes it past 50%. As the pack size keeps increasing though the percentage of self use starts tapering off and the big red grid draw bar stays pretty flat. This is because of the biggest intermittency gap cause, Winter.
With the shorter days from November through February my panels generated about 1,800 kWh while my home consumed about 5,000 kWh. This 3,200 kWh difference is about equal to the excess solar production my home produces during the rest of the year. To go off grid I would need a way to store this huge power difference. As you can see in the above graph the solar use percentage starts leveling off at higher battery sizes. The reason for this is that the extra kWhs added would only charge and discharge a couple times a year. Battery capacity from 70 kWh to 3200 kWh for my home would only charged and discharged once per year. This is not an efficient way to use a battery that costs a lot and is capable of thousands of charge cycles so expanding to that size would be foolish.
Does a Powerwall Make Sense?
The most common form of home energy storage today comes from Lithium-ion (Li-ion) batteries. For just $7050 you can buy a 13.5 kWh Tesla Powerwall 2 made with this tech. As you can see in the above graph that would increase my self usage to 51%, but the weird thing is that it would save me $0. This is because in PA I have net metering which means that I get to use the kWhs I generate to offset the kWhs I consume no matter when I generate them. If my state didn't have net metering then I would have more incentive to add a battery.
Another way to accomplish this is through time of use rates (also called TOU rates) that charge different amounts for a kWh based on the time of day it is consumed. Electricity production costs change throughout the day based on supply and demand, so a TOU rate can charge customers more fairly based on their usage patterns. But they also confuse a lot of customers and there are still lots of utilities that just charge a single flat rate per kWh no matter when it is consumed.
California utilities have truly embraced TOU rates and have some with huge peak price changes. The TOU-D-PRIME rate from SCE adds 26 cents for each kWh used between 4 and 9 pm in summer. A Powerwall 2 is warrantied to provide around 37,800 kWh of total discharge over its life or 18.6 cents/kWh if sold for $7,050. On this rate plan every kWh the Powerwall lets you shift out of this peak is profit. Of course to get an exact ROI you have to factor in prices during weekends and winter, and how many kWh of use the battery will really shift each year. In the end though if you're on this rate it's likely that a battery will have a higher ROI than bonds.
The battery backup provides even more value by keeping a home powered during blackouts. These are now happening more often in CA due to wild fire issues. If a battery backup avoids the need to install and fuel a backup gas generator, along with this peak price avoidance then its ROI could be really impressive. With battery prices continually trending downwards home based batteries may become one of the best possible investments in California in the near future.
To truly optimize ROI on the above rate the Powerwall would stop feeding energy back into the grid after 9 pm. This still means that your home would be drawing power from the grid at night. Looking into why a utility would incentivize people to keep drawing power late at night shows why most home owners still don't need batteries right now.
When Does A Grid's Energy Mix Requires Batteries
To understand when batteries are needed for a grid you need to look into how much power different energy sources provide to that grid throughout the day. Luckily the EIA provides a handy web tool that does just this for pretty much anywhere in America. This tool lets you drill down to your particular grid and see daily and even hourly electricity generation by source. Here's hourly generation by source for my grid in Pennsylvania (PJM) on April 27.
As you can see on the above graph my grid has a pitifully low amount of solar generation. Our wind gen is low too, but its high enough to at least be visible. When the wind picks up you see the green area thickening and the brown Coal and Natural Gas areas contracting. These plants are being fed less fuel similar to how utilities have reduced generation during times of lower demand for decades. This shows how initial deployments of renewable generation displace fossil generation without any battery storage. This does require the utility to predict of how much renewable generation there will be at a given time, but they're are actually pretty dang good at this. As they gain more experience and have renewable generation sites across a larger area their predictions will only improve.
There are also grids that have installed such a high percentage of renewable generation that battery storage could help them reduce costs and emissions even further. California's grid (CISO) is an example of this, with a far higher percentage of renewable generation than PJM. In this graph you can see how gas usage only peaks up at the end of the day when the sun goes down but demand is still high. This is also the time of day the CA TOU rate I discussed above charges extra for, which makes sense because gas generation has fuel and personnel costs that renewables simply don't.
This graph also shows how gas generation in CISO currently has a bit of a floor. There's a period around noon when CISOs total energy generation can be above their total demand, but their gas plants consumption doesn't go to zero, it's just as high as it was at 5 am before the sun came up. This is because a lot of gas generators can take hours, or even days to start back up after being shut off. Operators who expect their plants to be needed sooner than that leave them on at their lowest consumption mode instead of turning them off entirely so they can quickly ramp back up to meet the demand that they know is coming. When there's enough battery storage or fast starting gas plants to meet the demand that these gas plants are starting back up for then the gas floor can drop as the plants that can't start up fast get truly shut down for longer periods of time, and eventually forever. This is part of the reason why CA utilities are starting to contract for some truly huge battery storage.
More Solutions than Just Batteries
Of course batteries aren't the only way to solve the intermittency problem for renewable generation. Excess renewable energy can be stored by pumping a bunch of extra water up behind dams, or creating green hydrogen. Vertically mounted bifacial solar panels can generate more at dawn and dusk. Wind can be mixed with solar with the understanding that there are many times when the wind is blowing, but the sun is down and vice versa. This is particularly true if you build efficient long distance transmission so localized weather conditions matter even less. Another option is to simply massively overbuild renewables so that they produce what you need when conditions are bad, and you simply turn some of them off when conditions are good. Given the falling prices of batteries though, they just may continue to be the best solution, particularly if we ensure that the coming EV transition includes vehicle to grid (V2G) discharge capability and EV battery life further improves.
The reality is that most of these options will be implemented by utilities at the grid level. If we want homeowners with solar panels to buy their own batteries then we'll need to change how we charge them for power. Time of use rates with aggressive peak pricing like in California are a great way to do this, but to be fair they need to be applied to all utility customers, not just those with solar panels. In most of the U.S. right now the most profitable choice is to simply install solar without a battery. Given the extremely low percentage of renewable generation in many U.S. grids, like PJM, this is probably the right solution for the time being. Hopefully someday in the near future all of our grids will have enough renewable generation that storing excess green electricity an important consideration everywhere.