Dispatchable Solar is Here
By: Chandu Visweswariah
Key Take-Away
Energy systems of the world are facing steep challenges including urgent growth in electricity demand, limited grid resources and decarbonization imperatives. Thanks to improvements in solar panels and battery storage, pairing solar farms with batteries is now the cheapest, fastest and cleanest way of ramping up electricity generation. Best of all, this energy can be provided with fewer demands on aging and limited grid infrastructure.
Energy Challenges
Energy systems of the world face four serious challenges.
- Demand is growing. One reason is the advent of Artificial Intelligence and the power-hungry data centers that are required. Another reason is the growth of electric vehicles, electric cooking and heat pumps, which increase electricity demand but lower total energy demand. The growth of the middle class adds to energy demand. Industrial processes are being electrified, thus increasing factory electricity requirements.
- There is no quick way of satisfying additional electricity demand. Oil, gas and coal-fired plants take about 3 years to build, while nuclear power plants take even longer.
- Grid resources are limited. Even if we can build power plants quickly, the transmission of the energy from where it is produced to where it is needed does not exist and takes 10 years or more to build out. Consider this: worldwide, there are more than 3,000 GW (gigawatts or billions of watts) of renewable energy projects stalled in the queue of projects awaiting grid interconnection.
- The decarbonization imperative demands higher penetration of renewable sources like wind and solar. In addition, adopting renewables increases energy security in many parts of the world. However, their intermittency poses hurdles. For one thing, wind and solar energy are not dispatchable, i.e., they do not supply a steady amount of energy 24 hours a day. For another, they are wasteful of grid resources. An example: a 1 MW solar farm requires a 1 MW transmission line to evacuate the generated power. But most of the time (and certainly at night and during cloudy periods), the farm is producing far less than 1 MW of electricity, thereby wasting transmission capacity.
The Solution
Thanks to steady year-over-year improvements in solar panels and battery technology, we can now pair solar farms with on-site battery storage to provide dispatchable solar energy economically. The idea is simple: Between dawn and dusk, produce more energy than needed. Store the excess in batteries locally. Discharge those batteries between dusk and dawn (and during cloudy periods) to provide a uniform amount of power. These power plants can be built quickly (within a year), they produce carbon-free electricity, and, best of all, the resulting energy is less expensive than fossil fuel or nuclear plants. In the case of a local load like a data center, remote infrastructure, dedicated factory, microgrid or military installation, such power plants can be built with little to no reliance on grid capacity.
The Formula
Let’s say we want to provide a steady, dispatchable 1 kW of power, evacuated through a 1 kW transmission line. How much solar do we need? How much battery capacity? How much will it cost? Can we supply uniform power every hour of the year 24×365? The answers depend, of course, on how sunny the location is, the average length of daytime, how often back-to-back cloudy days occur, how we design our system, and so on.
Solar has a capacity factor of 20%, which is the ratio of actual energy produced to peak production under ideal sunlight conditions. Therefore, we need at least 5 kW in our exercise, so as to produce 24 hours x 20% x 5 kW = 24 kWhrs in an average day, which is the minimum required for a steady 1 kW. But how much storage do we need?
We start with the actual energy generation profile of 5 kW solar panels in a sunny place, Las Vegas in this example. For about 9 hours of the day, the panels generate at least 1 kW (yellow energy) which is directly provided to the grid or local electricity load. During those 9 hours, anything above 1 kW is routed to the battery (orange energy). The amount of energy in the battery can be approximated by the area of a triangle with width of ~9 hours and height of ~4 kW, i.e., 18 kWh. Since there are roundtrip losses in the battery, only 90% of battery capacity is usable, and cloudy conditions can occur, a battery size of 17 kWh is very reasonable to provide the 14.4 kWh needed during the rest of the day (green energy). This simple analysis tells us that we need at least 5 kW of panels and 17 kWh of batteries. As a practical matter, in places which are less sunny, 6 kW of solar would work better.
If we built such a system, how many hours in a year would we be able to provide steady energy? It turns out that in a sunny location like Las Vegas, such a combination will provide steady power in 97% of the 24×365 hours of the year. Note that even a gas or nuclear facility requires periodic maintenance and has unanticipated failures, so they rarely exceed such an availability metric. In a cloudy region such as Birmingham in the U.K., we would be able to provide steady energy in 62% of the hours in a year. While that number is low, it is much more useful than a non-paired solar farm and is still of value to a system operator as a dispatchable resource.
The figure above shows the percentage of hours during which our design can provide steady energy in different cities of the world, as well as the cost of energy in USD/MWh — more on cost analysis in the subsequent section. What would happen in the remaining “shortfall” hours of the year? There are two cases. If the solar farm is not grid connected and solely supplies a local factory, data center or military installation, then either the load has to be reduced during those shortfall hours, or a generator of some type would have to be fired up. In the case of a grid-connected solar farm, the situation is no different from a gas or nuclear plant being taken down for maintenance – during those shortfall hours (which are very predictable), there would have to be other generation sources brought into play. Nonetheless, having steady dispatchable energy between 62% and 97% of the hours of a year (depending on solar resources of the location) can be a huge contribution to both grid-connected and microgrid generation.
Cost Analysis
At the end of the day, the Levelized Cost of Energy (LCOE) represents the cost at which a kWhr of electricity can be produced by various alternatives. In just the last year, the price of solar panels as well as batteries has fallen dramatically. We see in the figure above that pairing solar with batteries produced electricity at $183/MWh (or 18.3 ¢/kWhr) in 2019. That figure has fallen dramatically to $104/MWh (or 10.4 ¢/kWh) in 2024. Just from 2023 to 2024, solar panels got 22% cheaper while batteries enjoyed a 40% cost reduction. In fact, while battery costs have fallen to $162/kWh on average (the figure used in this analysis), the recent Tabuk and Hail battery auctions in the United Arab Emirates achieved a stunning cost of $72/kWh. The cost-reduction trend is clear and dramatic!
How does this compare with coal, gas and nuclear?
Solar paired with batteries is less expensive than nuclear ($182/MWh) and coal ($118/MWh). Depending on the reliability with which we desire electricity, the solar/battery solution can be less expensive than gas-fired turbines as well. More importantly, the trend is clear: commodity-based energy generation (nuclear, coal, gas) is increasing in cost, while manufactured solutions (solar, batteries) are reducing in cost year-over-year. Very soon, the solar/battery solution will be the cleanest, cheapest and most reliable solution bar none. And if generous fossil fuel subsidies worldwide, to the tune of $7 trillion per year, start to get withdrawn, the game will be over instantly. Importantly, the solar farms paired with batteries can be built much more quickly.
The Tradeoffs
Let’s pretend we are building a solar/battery farm to provide 1 GW of steady power. To provide perspective, the load of the entire state of Vermont is about 0.9 GW, so our imaginary power plant would be sufficient for all of Vermont.
Our “formula” from the previous section would call for 6 GW of solar (a bit more than the minimum requirement of 5 GW), and 17 GWh of battery storage. We now know that depending on how sunny our location is, we could provide steady power between 62% and 99% of the hours in a year with such a configuration.
If we had less solar (less than 6 GW but more than 5 GW), then cloudy days (or multiple cloudy days in a row) would lead to more hours in which we are unable to supply the 1 GW goal. If we had more solar panels, it would increase project cost but also boost reliability, especially if paired with more battery storage.
If we had more battery storage, it would help in persistently cloudy conditions but would increase system cost. Lower battery storage would decrease reliability. Thus, a system can be optimized for each situation taking into account reliability requirements and local solar irradiation data. These tradeoffs are explored in more detail in Ember’s report.
Examples
Battery-paired solar farms and solar-powered data centers are springing up everywhere! Countries like China and India are significantly reducing their fossil fuel imports (and increasing their energy security) by relying on battery-paired solar farms. Fully 75% of new solar farms in the U.S. have batteries on-site. The United Arab Emirates is developing a gigawatt-scale 24×365 solar project which will consist of a 5.2 GW photovoltaic plant combined with 19 GWh of batteries to provide 1GW of uninterrupted supply to the grid. Sonoran in Arizona and the Mesa plant in Australia are smaller examples of the same. Dubai has a solar-powered data center. The world’s largest solar-powered data center is the 100 MW Moro Hub in the United Arab Emirates that was commissioned in 2022. A titanium melting furnace in West Virginia operates on solar electricity as does a 16-hotel resort in Saidi Arabia. As trends advance, watch for these generation plants to dominate new energy capacity construction, irrespective of national priorities and policies.
Conclusions
We are at an exciting juncture in the evolution of electricity generation. For the first time in history, solar power combined with batteries gives us the cheapest, cleanest, firm dispatchable energy. What is more, it is ideal for data centers, factories and remote locations with inadequate (or non-existent) grid resources. Now that we are at the economic inflection point, watch for massive disruption in the electricity generation space, with concomitant environmental benefits! This is the beginning of the end of coal, gas and nuclear generation.
