The Energy Collective Group

This group brings together the best thinkers on energy and climate. Join us for smart, insightful posts and conversations about where the energy industry is and where it is going.

9,802 Members

Post

Cost of Batteries for Electric Vehicles Falling More Rapidly than Projected

Full Spectrum: Energy Analysis and Commentary with Jesse Jenkins

Summary: The cost of battery packs for electric vehicles has fallen more rapidly than projected, with market leading firms in 2014 producing batteries at ~$300 per kilowatt-hour of storage capacity, on par with market projections for 2020.

Electric vehicle (EV) battery costs have fallen more rapidly than many projections, according to a new survey of battery costs published in Nature Climate Change. Researchers from the Stockholm Environment Institute scoured peer-reviewed journals, consultancy reports, and news items to construct an original data set of EV battery pack cost estimates from 2007 to 2014. Average battery pack costs have fallen 14 percent per year across the industry, which has seen sales volumes double annually in recent years. EV battery packs now cost $410 per kilowatt-hour (kWh) of storage capacity on average (with a 95 percent confidence interval ranging from $250–670 per kWh).

 Cost of Li-ion battery packs in BEV

Figure 1: Cost of Li-ion battery packs in BEV. (Source: Nykvist & Nilsson, 2015)

The cost of batteries produced by market leading firms, such as Renault-Nissan and Tesla Motors, however, have fallen further, to an average of $300 per kWh, according to the study. These estimates are on the order of two to four times lower than many recent peer-reviewed papers have suggested and already equal to the average cost projected for 2020 in a variety of papers. Costs for market leaders have declined at an average of 8 percent per year, the study estimates.

At $300 per kWh, electric vehicles can begin to compete economically with traditional petroleum-fueled internal combustion engines when gasoline costs $3-5 per gallon (€0.73-1.22 per liter), according to separate analyses from global consulting firm McKinsey and the International Energy Agency. The U.S. Department of Energy (DOE) has set a target of $150 per kWh for battery electric vehicles to become broadly competitive and see widespread market adoption.

In the near-term, the researchers believe economies of scale, improvements in cell manufacturing and learning-by-doing in pack integration, rather than advancements in cell chemistry or other R&D breakthroughs, will help manufacturers continue to produce cheaper batteries.

EV battery sales volumes are current doubling annually and car manufacturers are partnering with battery makers to invest in larger production facilities and cut costs. Renault-Nissan is working with LG to produce enough batteries for 1.5 million electric vehicles per year by 2016 while Tesla Motors and Panasonic are building a “Gigafactory” in Nevada that will produce 500,000 packs for EVs along with additional batteries for stationary energy storage, for a total of 50 million kWh per year of battery production. Tesla and Panasonic are targeting a further 30 percent decline in battery pack costs by 2017, which would require a 7 percent annual decline in costs, consistent with a continuation of recent rates for market leading firms.

The study’s authors conclude that economies of scale are likely to drive down battery costs to $200 per kWh in the near future. Further cell chemistry improvements may be necessary to hit the $150 per kWh target envisioned by the U.S. DOE. At those prices, electric vehicles may soon break out of niche markets and achieve much wider-scale adoption. 

Publication: Rapidly falling costs of battery packs for electric vehicles,” Nature Climate Change, Vol. 5 (April, 2015): 329–332. 

Björn Nykvist is a Research Fellow and Måns Nilsson is Deputy Director and Research Director at the Stockholm Environment Institute.


Note: This is article is part of an ongoing series of concise summaries of interesting and important conclusions from new research and peer-reviewed journal articles. This series at Full Spectrum is written in partnership with Observatorio de las Ideasa Spanish-language publication which finds and summarizes important, cutting-edge ideas for policy makers, business leaders, and others on key topics like energy, health care, economics, and more.


Conversation Starters:

  • How cheap to batteries need to be before electric vehicles will be widely adopted?
  • Can the battery industry continue to achieve strong cost declines?
  • Will economies of scale and learning by doing be enough to make batteries cost competitive? Or will new battery chemistries and R&D be required?
  • What impact does growing demand for stationary batteries for grid connected uses have on costs and prices in the electric vehicle battery sector?

Content Discussion

Hops Gegangen's picture
Hops Gegangen on April 13, 2015

 

So, when the EV batteries degrade enough that they don’t serve well in vehicles, will they go on to play a role in storing intermittant renewable energy? 

I would think so. 

Then eventually the material gets recycled.

Jenny Sommer's picture
Jenny Sommer on April 13, 2015

Or the batterie banks for gridstorage and homestorage will be much cheaper in the long run.
Probably some organic flow battery that you can buy like heatingoil today.
Lasting 40k+ cycles and when the electrolyte is done you flush it out and refill it.

The study is far too conservative. I predict sub 100$/kW before 2025.

Clayton Handleman's picture
Clayton Handleman on April 13, 2015

Jesse,

Some time ago, Bloomberg put out a graphic comparing the drop in cost of EV batteries compared to the experience curve for laptop batteries.  The EV batteries were considerably more costly.  As shown below, they made the mistaken interpretation that the EV batteries would ramp down a parallel experience curve.

 

Click the graph to enlarge.  (Note:  Leaf battery info was something I added)

However what has happened instead is that the experience curve for EV batteries rapidly took a nose dive towards the laptop battery experience curve.  This makes a lot of sense.  One set of costs was based upon the cost of integrating the batteries and the other was based upon the actual cost of the cells.  The cell cost was well down its experience curve.  However the packaging was relatively low volume so there were rapid cumulative doublings of the packaging portion.  This has led to rapid cost reductions in Li-ion.  As the two cuves merge, the EV battery costs will likely follow the lower line and the rate will be less per cumulative doubling.  I wrote this up some time ago and included links to the Navigant and McKinsey reports that discussed cost reductions in Li-ion. 

Experience curves are surprisingly good predictors of cost reductions.  This link leads to a post that explains experience curves and some of the history.

 

 

Nathan Wilson's picture
Nathan Wilson on April 13, 2015

Will an exploding BEV industry drive cheap battery storage into grid applications?

Assuming batteries accelerate past $150 to bottom out at $100/kWh, for a system with 15 hours of storage, 70% depth of discharge, and $500/kW for balance of plant, storage would cost $100*15/.7+500=$2640/kW (which is $176 per usable kWh).  With 200 equivalent full 15h cycles/year and 12 year life, that’s 7.3 ¢/kWh; with 6% interest this comes to $2640/kW/15h/200*0.1193=10.5 ¢/kWh.  To install such a system at the residential level, the cost would be 50-100% higher.

With the rare except of places like Hawaii, which burn imported oil for electricity, this cost simply cannot compete with fossil fuel backup (nor can pumped-hydro, since the EIA reports a capital cost of $5.3/W, which more than offsets the increased service life).  So energy storage has poor prospects for helping variable renewables replace fossil fuel.

Clayton Handleman's picture
Clayton Handleman on April 14, 2015

Nathan,

How did you come up with 200 cycles per year?  I would think that in a high penetration scenario, considerable geographical decorrelation with wind and high decorrelation with solar and wind, we would see less duty cycle required.  Further, in a likely load shift scenario much of the storage could be covered free by auto batteries.

The expert conscensus is now moving towards $100/ kwhr by 2025.  But the conversation on TEC threads is about high penetration renewables over a multi-decade timeframe.  In addition to experience curve driven by rapid EV growth, there are also other battery technologies that appear likely to be adopted.  It would be interesting to rework your analysis for $50 / kwhr with a 100 day equivalent.  While a little more aggressive it is reasonable.  For back of the envelope I think your numbers are reasonable for a what if but far from compelling in terms taking battery storage off of the table.

 

 

Joris van Dorp's picture
Joris van Dorp on April 14, 2015

So energy storage has poor prospects for helping variable renewables replace fossil fuel.”

Rather than being a disadvantage, this is precisely why renewables+storage has been popularised so aggressively in recent years.

While society grapples with the implausibility of replacing fossil fuels with renewables+storage, fossil fuel stakeholders can spend that much more time without having to worry about the horrific implications (to their interests) of advanced nuclear power becoming popular. Anything, nomatter how ridiculous, which perpetuates the illusion that nuclear power is not essential to protect our common future is bankable for fossil fuel stakeholders.

The renewables+storage boondoggle is serving to extend the historic waste of time and resources caused by the renewables-only boondoggle of previous decades, after that boondoggle seems finally to be losing it’s appeal due to the costs caused by RE intermittency becoming too great to ignore.

Nathan Wilson's picture
Nathan Wilson on April 14, 2015

Clayton, good point about the importance of cycle rate on battery economics.  My goal was the most favorable situation for batteries, so I assumed a PV-centric world (roughly a desert-like 300 sunny days/year and less hours of sun in the winter).  I agree that wind and solar together would put much less cycles per year on the batteries, which would extend their calendar life a bit, but could make the per-cycle battery cost proportionately more!  

Updating with your speculative values of $50/kWh and 100 cycle-equivalent per year,  assuming 20 year calendar life, and as before: 15 hours of storage, 70% depth of discharge, and  $500/kW for balance of plant, storage would cost $50*15/.7+500=$1570/kW (which is $105 per usable kWh). so at 6% interest the cost per kWh is $1570/kW/15h/100*0.08718= 9.1 ¢/kWh.

(By the way, I pulled the 0.08718 Capital Recovery Factor out of a table in an old text book, but I’m sure there is a way to get it from MS Excel also; with the assumed 20 year battery life, we really can’t ignore interest).

I would not take batteries off the table entirely, but I consider them a bit of a long shot.  I think people often under-estimate how big the cost gap is between batteries and fossil fuel (even in places like Hawaii).

Jenny Sommer's picture
Jenny Sommer on April 14, 2015

I would use the batteries that have at least 40k cycle life and 100% range. 

Existing grid batteries do much better than your example.

First you offset expensive spinning reserve. You provide auxiliary services such as frequency regulation on a level unreached by fossil plants. The response is instantaneous. When prices go negative the battery takes load. The cycle is therefore 200% “depths” if you want.

Obviously batteries are for short term storage up to some days. Storage will never be a problem anyways. Germany got over 200TWh of storage in place already. 

 

 

 

Hops Gegangen's picture
Hops Gegangen on April 15, 2015

 

I don’t think any debate about the future of batteries is complete without considering capacitors, which store charge on surfaces without a chemical reaction. We currently have what are called supercapacitors. In an EV, these can be used for bursts of power or to accept a surge of current from a charger or regenerative braking. But they don’t hold a charge very long, nor very much compared to batteries.

Researchers at MIT and elsewhere are using nanotechnology to build capacitors with even higher capacity and the ability to hold a charge longer — several hours — and to last longer.

If the demand for energy storage increases, I could imagine these things being mass produced inexpensively. Then the peak solar during the day can be buffered for a few hours into the evening when air conditioning is stll a big load and other appliances are running. Or when you get home from work in the EV as the sun is going down, the capacitor would have a charge to transfer to the EV battery. Or if you are retired, maybe you just the electricity to the grid during peak demand. 

The EVs themselves might start to have more capacitors rather than batteries for short trips. At which point quick-charge outlets in parking lots would allow you keep the charge topped off. 

Clayton Handleman's picture
Clayton Handleman on April 15, 2015

Thanks Nathan,

Battery storage is an important detail in the complex discussion of transitioning to a low carbon future. You have done a nice job bringing it to the table in an easy to understand quantitative way.  I still see a few things that give me pause with your numbers but probably will not have time until July to dig in and generate a thoughtful rebuttal.  If you are game, at that time, perhaps we could see if we can come close enough to agreement to develop a collaborative post/article for TEC.

Clayton Handleman's picture
Clayton Handleman on April 15, 2015

Even sticking with Li-ion the pros have been predicting $160 / kwhr by 2025.  Now Musk and others are saying $100 / kW by 2025.  I am on record since 2013 predicting $100 / kw by 2029 and $50 / kw by 2034.  The $50 number may need revision if oil prices stay low and slow the introduction of EVs but I don’t think it will push out beyond 2040.  Auto batteries require high energy density.  Grid storage not so much.  This makes a secondary market for repurposed EV batteries very attractive.  With a secondary market to sell their used batteries to, EV’s will have a reduced lifecycle cost for their cars and this will accelerate the introduction of EVs.

Bob Meinetz's picture
Bob Meinetz on April 16, 2015

Hops, from what I understand high voltage is not a problem, but it requires a lot of capacitors to store even a fraction of the chemical energy available in gasoline, or electrochemical energy available in chemical batteries. A bit like storing energy in a flywheel, spring, or weight raised into the air – all seem feasible until you start crunching the numbers.

You’ve probably heard of EEStor, the supercapacitor manufacturer and Wall St. darling of a few years back. The company has attracted a lot of controversy for not delivering on claims:

EEStor’s claims for the EESU exceed by orders of magnitude the energy storage capacity of any capacitor currently sold. Many in the industry have expressed skepticism about the claims. Jim Miller, vice president of advanced transportation technologies at Maxwell Technologies and capacitor expert, stated he was skeptical because of current leakage typically seen at high voltages and because there should be microfractures from temperature changes.

http://en.wikipedia.org/wiki/EEStor

Bruce McFarling's picture
Bruce McFarling on April 17, 2015

“How did you come up with 200 cycles per year? I would think that in a high penetration scenario, considerable geographical decorrelation with wind and high decorrelation with solar and wind, we would see less duty cycle required.”

In a high penetration scenario where we assumed battery storage was the only storage and there was little or no dispatchable RE … then battery storage cycles would track total storage cycles. But in a high penetration scenario where battery storage is part of a storage portfolio and there is dispatchable RE … then battery storage cycles ought to be in the higher frequency, lower total storage capacity end of the firming capacity.

Indeed, in the “duck belly” scenario for projections of high PV penetration in Southern California in the winter season, there would often be two storage cycles per day … storage of surplus wind power for use in the morning, and then storage from the solar peak for use during the later part of the demand peak as we head toward the earlier winter dusk and sunset. If there is capacity constrained dispatchable generation (dammed hydro, biogas with peak generating capacity, etc.), the economics of the batteries look better if they are built up to the smaller of the two daily firming demands, and the gap between the smaller and larger of the firming demands are filled with the capacity constrained dispatchable generation.

Another form of “double cycle” is when a generating capacity is more efficient operated as a relatively stable output generator (constant output or steady ramp-up / ramp-down), while load is expected to rise and then fall more rapidly, where the stable output is (1) started in a position to generate a surplus over load, with the surplus stored, (2) the surplus declining until load passed output of the generator (3) the storage drawn to supplement generation until the load returns to the output of the generation (4) at which point the growing surplus is used to charge the storage again and (5) the generator is shut down and the balance of power drawn from the storage.

As far as total storage capacity, the higher frequency tasks require a smaller total storage capacity, so working out the cost of the battery storage for 100 annual cycles, 200 annual cycles and 300 annual cycles is working out the potential for battery storage to play a range of roles in the firming portfolio. The “natural” economic position of capacity constrained generation and fueled generation in the firming portfolio is toward the lower frequency end, and the “natural” position of battery storage would be toward the higher frequency end.

Willem Post's picture
Willem Post on April 19, 2015

Nathan,

Regarding battery system capacity sizing, here is a real life example for a stationary lead-acid battery  system for an off the grid house. Similar factors apply for other type batteries, whether stationary or mobile:

Example of determining required battery capacity: 

House electrical consumption, overcast winter day 10,000 Wh/d; excludes plug-in vehicle

PV solar production, overcast, winter day 5,760 Wh/d

Generator production 2,800 Wh/d; minimizes battery discharge.

Battery discharge 1,440 Wh/12 V/d = 120.0 Ah/d; 12 V batteries are used

Period of consecutive low PV solar 6 days

Battery discharge after 6 days 720 Ah

System overhead factor 1.2

Battery aging factor 1.2

DOD factor 2.0

Temperature factor 1.11 (if in an unheated space)

Battery capacity required = 120 x 6 x 1.2 x 1.2 x 2.0 x 1.11 = 2,302 Ah

Battery cost = (24) 100 Ah units x $300/100 Ah = $7,200

DOD after 6 days = 720/2400 x 100% = 30%, well below the 50% design limit.

Jenny Sommer's picture
Jenny Sommer on April 19, 2015

EV batteries and HomeStorage probably won’t be the same batteries.
I’d rather have organic, non burnable, nontoxic electrolytes at home than some high density, high discharge cells that catch fire now and then if penetrated, abused or improperly managed. Imagine what that would mean for insurance.

I don’t see that happening.

Log me in for under 100$/kw before 2025 (that is for gridstorage and low density solutions).