Batteries are key to clean energy — and they just got much cheaper
Clean energy future might be closer than we previously thought
Eric Holthaus April 10, 2019 8:00AM (UTC) salon.com
This post originally appeared on Grist.
Batteries are critical for our clean energy future. Luckily, their cost has dropped so low, we might be much closer to this future than we previously thought.
In a little less than a year, the cost of lithium-ion batteries has fallen by 35 percent, according to a new Bloomberg New Energy Finance report. Cheaper batteries mean we can store more solar and wind power even when the sun isn’t shining or wind isn’t blowing. This is a major boost to renewables, helping them compete with fossil fuel-generated power, even without subsidies in some places, according to the report. Massive solar-plus-storage projects are already being built in places like Florida and California to replace natural gas, and many more are on the way.
The new battery prices are “staggering improvements,” according to Elena Giannakopoulou, who leads the energy economics group at Bloomberg NEF. Previous estimates anticipated this breakthrough moment for batteries to arrive in late 2020, not early 2019.
According to the report, the cost of wind and solar generation is also down sharply — by between 10 to 24 percent since just last year, depending on the technology. These numbers are based on real projects under construction in 46 countries around the world.
The lower battery prices have big implications for electric cars, too. There’s a key cost threshold of about $100 per kilowatt hour, the point at which electric vehicles would be cheap enough to quickly supplant gasoline. At this rate, we’ll reach that in less than five years.
Now that cheap batteries are finally here, we’re well on our way to electric modes of transportation and always-on renewable energy — and not a moment too soon.
What’s driving the plunge? Giannakopoulou cites “technology innovation, economies of scale, stiff price competition and manufacturing experience.” Other storage methods, like pumped hydro, still account for the vast majority of energy storage capacity, but lithium-ion batteries are much more flexible and don’t require specific locations or environmental conditions to work. Like everything in the built environment, lithium-ion batteries also require mining and manufacturing. There’s still a chance that some new exotic battery technology will quickly supplant lithium-ion, but its ubiquity and — now — cheapness will be hard to beat.
Electric vehicles will become cheaper to own and operate than gas ones. In places like California, Texas, and Germany, electricity prices have occasionally dropped below zero — a sign that the grid wasn’t yet ready to handle the glut of renewable energy produced there. Now, more of that cheap power will be stored and passed on to consumers. This could be the moment when renewable energy starts to shut down fossil fuel for good.
Here is some green energy for you: Sen. Elizabeth Warren said Monday that she would immediately move to place a “total moratorium” on new federal fossil fuel leases if elected president, blocking energy companies from drilling offshore and producing oil, gas and coal from U.S. government-owned land.
What Senator Warren actually said was that she would issue "an executive order that would prohibit new leases for fossil fuel drilling offshore and on public lands". Exploration and production would continue on current leases, on-shore and offshore. Considering the current glut of fossil fuels compared to demand, I'm unsure any reserves covered in such an order would be needed in the future. There are a lot existing leases that have not as yet been developed some of which have been held for many years. This would hardly put a crimp in supply for a couple of decades.
Yeah, I know what she said and that is insane. The green energy people would love it until they realize how much money they will have to cough up to replace federal bonus monies, rentals, and royalties. It is not just offshore, the feds own a ton on onshore land too. Just another in the long line of “I want everything for free I just don’t want to pay for it”. That strategy has worked so well in other countries.
Income from rents/royalties/bonus is about one tenth of one percent of the national budget.
All told, the gross income (before payments to states) from onshore oil and gas resources averaged $3.0 billion annually from 2005 to 2014, comprising the following amounts:
Revenue streams and rates
Oil and gas
$1.50 annual rent per acre for the first 5 years, $2 annual rent per acre thereafter
12.5% of production value in royalties
Water depth 0–200m: Years 1–5 rent is $7/acre, year 6 rent is $14/acre, year 7 rent is $21/acre, year 8+ rent is $28/acre
12.5% for leases located in water depths less than 200 meters
While royalty rates can reach as high as 18.75%, and the federal corporate income tax rate can reach as high as 21% depending on company income, companies may pay less. Revenue policy provisions, including royalty relief and tax expenditures, can result in smaller revenue and tax payments to the federal government to promote other policy goals.
To create incentives for companies to produce additional oil and gas on certain leases on the Outer Continental Shelf where extraction is anticipated to be unprofitable, the federal government may grant some lease holders royalty relief. Royalty relief means that these lease holders do not have to pay royalties on some amount of production, or they pay a smaller percentage of royalties, for the oil and gas they extract. There are four situations in which a lease holder may gain royalty relief:
In some situations, if oil and gas prices rise above certain thresholds, lease holders that previously gained royalty relief must start paying royalties at the regular rate again.
Tax expenditures are defined in the law as “revenue losses attributable to provisions of the federal tax laws which allow a special exclusion, exemption, or deduction from gross income or which provide a special credit, a preferential rate of tax, or a deferral of tax liability.” These exceptions may be viewed as alternatives to other policy instruments, such as spending or regulatory programs.
The Treasury estimates the total dollar amount of each tax expenditure in a given year and publishes a report of these estimates.
I'm sort of interested in batteries. It seems energy storage is and will be a growing need in the future. And I'm interested in the process of keeping those new generation batteries powered up. Post more on this when available.
Wind and solar are low density energy sources, so it takes a lot of capital investment to capture the energy. Both of those sources are intermittent so backup energy generation system or a storage system is required to provide electricity on demand. As long as the wind or solar system is "tacked on" to an existing grid that can provide electricity on demand the price for the renewable system is not onerous. But, if the renewable system is designed to "stand alone" then the price with current technology is very high. It is highly unlikely that lithium batteries will ever be cheap enough to provide the storage for an entirely renewable electrical grid. Here is some analysis using real cost data:
So-called “renewable energy” sources are inherently intermittent with wind turbines producing when wind velocity is high enough and solar producing when the sun is high in the sky. These favorable conditions do not perfectly match electricity demand so to have an entirely renewable electrical system there is need for a storage system. In Texas in 2017 wind turbines produced about 30% of rated capacity. Output from utility scale solar photovoltaic (PV) systems in India report annual production at 24% of rated capacity. Thus, an entirely renewable system requires 3 to 4 times the installed capacity of a conventional fossil fuel system, as well as an energy storage system.
Wind farms cost about $2000/kW, recently constructed utility scale PV solar systems cost about $1000/kW (residential rooftop systems cost about $3000/kW), and a recently completed Australian lithium battery storage system cost about $700/kWh. By comparison, a modern combined cycle gas turbine (CCGT) system costs about $700/kW, and it doesn’t need the electricity storage system. Thus, in addition to the storage system an entirely renewable electricity generating system takes about 6 to 8 times as much invested capital as a CCGT system. And the storage system is very expensive as will be shown below.
Tesla completed a 129kWh electricity lithium-ion nattery storage system in the Hornsdale Power Reserve near Jamestown South Australia for a reported capital cost of $90 million or about $700/kWh. The battery life was estimated as 15 years with annual maintenance estimated at $4 million to $5 million, or about 5% of installation cost.
TCU Mechanical Engineering Professor E. E. Michaelides did a thorough detailed study to determine what it would take to replace the 69% of electricity generated in the Texas Grid with fossil fuel with renewable wind and solar. (See “Making Texas Green,” Mechanical Engineering, March 2019, pp. 36-41.) The study concluded that the fossil fuel produced energy (based on 2017 data) could be replaced with 30,800 MW capacity of wind turbines plus 52,200 MW of solar PV cells and 16,020 GWh of storage. (It was assumed that the existing nuclear systems that provide about 11% of electricity generated would remain in service.) Based on the capital costs given above the combined capital costs of the wind and solar generating systems would be $113.8 billion. (Because wind and solar are “low density” energy sources they require a lot of land, in this case an estimated 4600 square miles, acquisition of which would be a formidable task.) If we assume that substantial improvements are made to the lithium-ion battery system (batteries, enclosures, heating, air-conditioning, etc.) getting the cost down to $300/kWh, the cost of the storage system would be an astronomical $4.8 trillion. To put that in context, the annual GDP of Texas is a bit less than $2 trillion. This suggests that a lithium-ion battery system is a non-starter for long-term storage. (Note: the Hornsdale system does not provide long term storage to back up the grid, rather it smooths out the short-term fluctuations inherent in wind generation systems.)
Michaelides did not provide an estimate of cost for the storage system but suggested an alternative to batteries involving electrolysis of water to produce hydrogen which would be stored underground until needed when it would be used as fuel in gas turbine electricity generating systems. I don’t have a cost estimate for that approach, or for utility scale “flow” batteries being developed by Lockheed Martin and others.
Making Texas entirely “green” would be even more expensive than noted here because it would entail eliminating natural gas for heating homes and gasoline and diesel for transportation. Those changes would increase the amount of electricity needed substantially. There is a lot of competition for allocation of capital so it is unlikely that Texas would be willing to devote more than three years of GDP to making Texas energy entirely “green.”
“Over the last five years, the world began to experiment with storage; in the next five, storage will become a key grid asset,” says Ravi Manghani, Wood Mackenzie Power & Renewables director of energy storage.
Wood Mackenzie, 4/19/2019
As I stated, the primary source that I used was the article “Making Texas Green” by TCU Engineering Professor E. E. Michaelides in the March 2019 issue of the ASME publication Mechanical Engineering, pp. 36-41. This article is the first attempt I have seen to determine how much storage would be required to make viable an electrical grid that uses only solar and wind (with some nuclear in this case) with no fossil fuel backup. Michaelides did a detailed analysis of the Texas grid electricity consumption throughout the year and generation of electricity with wind and solar throughout the year. There is a lot of seasonal variation in both supply and demand, and these do not perfectly match. The result is that a lot of energy must be stored for long periods, not just storage while the sun is out for use overnight. The amount of storage required was determined to be about 16 times the daily average use in Texas. That seems like a lot, but when seasonal variation is considered, and experience in Germany and England where there have been periods for as much as a week in which wind and solar produced only 1 or 2% of nameplate capacity is considered, it may be reasonable. I let the “Making Texas Green” article speak for itself, I did not attempt to critique it.
All I did was take results presented in the ‘Making Texas Green” article and applied typical installation costs as discussed below.
The Tesla Hornsdale battery capacity is initially 129 MWh and the cost was $90 million, yielding an initial $697/kWh cost. (The battery capacity declines over time.)
The 2019 Tesla Powerwall has capacity of 13.5 kWh and costs around $10,000 installed for a total cost of about $740/kWh.
The cost of lithium-ion batteries at the time Hornsdale was done was about $240/kWh and is now about $200/kWh and is expected to decline to $62/kWh by 2030. That is speculative, of course. A utility scale battery will require housing, ancillary sensors and controls and a thermal conditioning system. If I assume that the Hornsdale battery cost $240/kWh, then the housing and ancillary equipment cost about $460/kWh. If I assume that battery cost falls to $60/kWh and the housing and ancillary cost (which will have a less steep “learning curve” than the battery) is cut in half, I get a total cost of $300/kWh. (Even if I just use the anticipated future battery cost of $60/kWh then 16,020 GWh of storage would cost a significant $960 billion)
Solar farms cost about $1000/kW to build.
Wind farm cost varies from around $1300/kW to $2200/kW:
The need for energy storage to make large scale renewable (wind and solar) electricity generation systems viable causes interest in developing alternative utility level storage systems that are less expensive than lithium-ion batteries. Lockheed Martin is one such company:
Note that Lockheed Martin’s flow batteries are intended for short-term and medium-term utility-level storage, not the long-term storage indicated as being required by Michaelides.
I have a Master’s Degree in Mechanical Engineering with over 50-years experience. I worked for over 40 years for Lockheed Martin (retiring at the end of 2003) and worked for a consulting company and a construction company before that. I designed missiles, spacecraft, and rockets and know nothing about Lockheed Martin’s energy storage business except for what is publicly available.
Thanks for the additional information, James. I should have been more clear in my question regarding attribution. Maybe I missed somewhere in the article what data Professor Micaelides was using in his analysis. When I have a little more time I will try to look through some of your hyperlinks to see if I can find that. Although the article is March 2019, I was wondering if it was based on 2018 data or earlier.
James, I have one observation for the article by Professor Michaelides. It is not his analysis which I think is likely correct. The problem inherent with trying to model a fast evolving technology is that by the time sufficient data is available, the technology has changed. The professor is using 2017 data. The articles that I post, and my comments based upon them, are current (2019). Part of my interest and a reason for the number of articles on the subject is the stunning strides in improved efficiency of latest solar technology and how much less expensive PV modules have become. The same evolutionary arc is ongoing for batteries. The entities deploying utility scale solar globally are reporting surprising costs per kWh. Along with the willingness of capital markets to invest in those projects, it appears we are fast reaching the point of price equivalence with hydrocarbon generated electricity grids.
According to the article it was based on Texas Grid data from 2017. Given the publication date I would surmise that the analysis, which probably took several months, was done in 2018.