Comprehensive modeling study finds electric drive vehicle deployment has little observed effect on US system-wide emissions
15 January 2014
The results of a new, comprehensive modeling study characterizing light-duty electric drive vehicle (EDV) deployment in the US over 108 discrete scenarios do not demonstrate a clear and consistent trend toward lower system-wide emissions of CO2, SO2, and NOx as EDV deployment increases.
As explained in their paper published in the ACS journal Environmental Science & Technology, the researchers from North Carolina State Univesity and the University of Minnesota found that, while the scenario parameters can influence EDV deployment—even to a most extreme scenario of adoption—this EDV deployment does not in turn produce a discernible effect on total system-wide emissions. There are three reasons for this lack of observed effect, they concluded: (1) at present the overall share of emissions from the LDV sector is only 20% of US CO2 emissions; (2) EDV charging can still produce comparable emissions to conventional vehicles depending on the grid mix; and (3) the effect of other sectors on emissions is significant.
EDVs offer three key benefits over competing vehicle technologies: (1) reduced consumption of petroleum-based fuels, (2) lower refueling infrastructure costs compared to alternatives such as H2 and compressed natural gas, and (3) a shift in energy production from vehicles to the electricity grid, where emissions from large, centralized facilities are cheaper and easier to control. While previous work has applied different methodologies and models to quantify the environmental benefits of EDVs, several consistent insights have emerged.
First, HEVs produce less emissions than conventional vehicles. Second, PHEVs with smaller battery packs are more likely to deliver emissions benefits and reduced gasoline consumption at lower lifetime cost compared to those with large battery packs in the short term. Third, significant emissions benefits, particularly from vehicles with large battery packs, only begin to accrue with clean electricity. Fourth, CO2 prices as high as 100 $/t do not provide sufficient incentive for vehicle electrification.
While these studies (along with others) have made significant contributions to the literature, they only consider a single point in time or employ sector-specific models or calculations that ignore the interaction of EDVs with the rest of the energy system over time. Recent analyses based on energy system models mainly focus on CO2 emissions and have been run with a limited set of scenarios, which make it difficult to draw insight specific to EDVs.
This paper employs an energy system model to meet the following objectives: (1) identify the conditions under which EDVs achieve high market penetration in the U.S. light duty vehicle (LDV) sector through 2050 and (2) to quantify the system-wide changes in CO2, SO2, and NOx emissions at the national level.
—Babaee et al.
The researchers used a model consisting of two components: The Integrated MARKAL-EFOM System (TIMES), which serves as a generic energy optimization framework and operates on the National US TIMES Data set (NUSTD), a TIMES-compatible data set constructed specifically for this analysis. TIMES is a bottom-up, technology-rich energy system model, which represents an energy system as a network of technologies linked together via flows of energy commodities. TIMES performs linear optimization to identify the least-cost way to satisfy end-use demands, subject to user-imposed constraints such as emissions limits and maximum growth rates on technology capacity.
In their analysis, the authors examined the effect of 5 factors on EDV deployment: crude oil and natural gas prices; a federal CO2 policy; a federal renewable portfolio standard (RPS); and EDV battery cost.
Assumed values associated with each factor were blended to create the large set of 108 scenarios that capture a wide range of potential outcomes. Given the highly uncertain role of consumer choice in future vehicle adoption, they noted, their analysis focused on the economic and environmental performance of EDVs assuming minimal behavioral barriers to vehicle adoption. “Strong and persistent reluctance on the part of consumers to adopt EDVs will dampen or eliminate the EDV-related effects presented here,” they cautioned.
Across all the scenarios, the total EDV deployment ranges from 0−42% of the LDV market with an average value of 24%—a figure broadly consistent with other projections of EDV market development.
No EDV deployment occurs with high battery costs, low oil prices, and no CO2 policy. At least 1 of these 3 scenario assumptions must change in order for EDVs to achieve some level of market penetration in 2050.
As scenario parameters shift to values more favorable to EDVs—i.e., higher oil prices, a CO2 policy, lower battery cost—the median market shares increase. The maximum EDV market penetration is 16% with the low oil price assumption versus 42% with reference or high oil prices. Similarly, high and reference battery costs limit EDV penetration to a maximum of 34% and 37%, respectively, whereas low battery costs enable the maximum market penetration of 42%. The maximum EDV market share is 42% because EDV deployment is largely limited to the compact and full-size vehicle classes, due to the higher cost of electrification of larger vehicles.
The CO2 cap results in marginal CO2 prices of 37−125 $/tCO2, which with other conditions held equal, only increase EDV deployment by approximately 3%. This result is also consistent with other studies demonstrating that CO2 prices less than 100 $/tCO2 have little effect on EDV adoption.
Finding that oil price and battery cost had the largest effect on EDV deployment, they varied these scenario parameters while holding the others constant the better to isolate the effect of EDV deployment on emissions. The high EDV deployment scenario assumes high oil prices and low battery cost, while the low deployment scenario assumes low oil prices and high battery cost. All four scenarios assume reference case natural gas prices and no RPS. They found that, without the CO2 cap, there is no change in electric sector SO2 and NOx emissions because the air pollution constraints remain binding.
Further, the system-wide net decrease in SO2 and NOx (approximately 3% for each) is largely unrelated to EDV deployment: higher oil prices lead to fuel switching in the fuel supply, heavy duty vehicle (HDV), and end-use sectors. Also without the CO2 cap, high EDV deployment creates a 21% reduction in LDV CO2 emissions but a 13% increase in electric sector CO2 emissions.
Accounting for additional changes across other sectors, the system-wide effect of high EDV deployment is a slight 0.9% decrease in total CO2 in 2050.
…it is not enough to simply incentivize the purchase of EDVs and wait for emissions benefits to accrue. The emissions benefits—if any—will depend on a broad set of future conditions. Therefore, public policies that target EDV deployment should be formulated, reviewed, and revised with careful attention paid to evolving changes to the broader energy system over time. If the primary objective is to reduce emissions, policy makers should focus on implementing targeted emissions policy rather than the promotion of specific technologies or fuels. Among the scenario variables tested, the CO2 cap produced the largest and most consistent drop in CO2, SO2, and NOx emissions. Although the observed marginal CO2 prices do not drive significant EDV deployment, the results indicate that EDVs can help lower the marginal price of CO2, particularly if scenario variables favorable to EDVs (high oil prices, low battery cost) prevail.
In the absence of a CO2 policy, the promotion of clean electricity can provide direct emissions reductions and also lower the emissions footprint from vehicle charging. The new EPA proposed carbon pollution standard and the forth-coming proposed rule on existing coal-fired power (due out in 2014) could have a significant impact on national emissions and eliminate some of the potential emissions increases associated with vehicle charging. Finally, other alternative vehicles are worth a mention. First, compressed natural gas (CNG) vehicles are not cost-effective in any scenario, including those with low natural gas prices, because low CNG prices are not enough to overcome the higher investment costs. Second, the model deploys diesel and diesel hybrids in many scenarios, which may be a cost-effective way to reduce CO2 emissions given their higher efficiency compared to conventional gasoline vehicles.
—Babaee et al.
Resources
Samaneh Babaee, Ajay S. Nagpure, and Joseph F. DeCarolis (2014) “How Much Do Electric Drive Vehicles Matter to Future U.S. Emissions?” Environmental Science & Technology doi: 10.1021/es4045677
No news here, most US power is from fossil fuels, it is well known that fossil fuel electricity powering EVs has similar CO2 emissions as ICE vehicles.
But that's not really the point. Low carbon energy sources are almost exclusively electricity sources - wind, solar, nuclear. They are not liquid fuel producing sources except for a few fairly marginal ones like biofuels.
So the point is EVs are more future proof than conventional internal combustion vehicles. It makes sense to start the transition now since EVs offer opportunities to integrate more wind, solar and nuclear into the grids.
Posted by: A.C. R. | 15 January 2014 at 12:11 PM
"Further, the system-wide net decrease in SO2 and NOx (approximately 3% for each) is largely unrelated to EDV deployment"
This is almost certainly wrongly modelled. NOx emissions from ICEs depend on catalysts warming up, so on short trips (better with PHEVs/EVs) the NOx emissions per km/mile are far higher. Cars are also throttled a lot which increases NOx emissions (catalysts work best at constant flow such as that seen from baseload power stations).
Posted by: A.C. R. | 15 January 2014 at 12:17 PM
"compressed natural gas (CNG) vehicles are not cost-effective in any scenario, including those with low natural gas prices, because low CNG prices are not enough to overcome the higher investment costs."
If that's correct, then it spells bad news for hydrogen, which is far worse in every infrastructure and vehicle cost item than CNG vehicles.
Posted by: A.C. R. | 15 January 2014 at 12:19 PM
They're downplaying that stationary power plants get a far higher percentage of available energy out of fossil fuels than ICE engines, that transporting power through a wire is generally more efficient than trucking fuel to a local gas station, and that electric vehicles go farther per gas equivalent because of stop-start, regenerative braking, greater efficiency at all RPMs, and such.
That said, better, cheaper batteries would make the virtues of electric vehicles more apparent.
Posted by: HealthyBreeze | 15 January 2014 at 01:13 PM
Reasons 1 and 3 in the first paragraph are the same thing. Denmark got 57% of its electricity from wind in December. Current generation mix is not the future and assumptions that is is are misleading at best.
Posted by: JMartin | 15 January 2014 at 01:22 PM
Well they do have a point that pure EVs aren't the best value/performance proposition at the moment. That is clearly given to hybrids, followed by plugin hybrids, and only after that pure EVs (except for some niche uses).
With lower battery cost the calculus shifts in that order, pluging hybrids get more attractive, then with ludicrously cheap batteries further into the future, the EV will have the best value proposition.
Posted by: A.C. R. | 15 January 2014 at 01:23 PM
Without going through the report to see the assumptions and so on, this is fairly meaningless.
However, in response to ACR it seems that they are looking at NG vehicles in respect of cost effectiveness in reducing CO2 emissions:
'The new EPA proposed carbon pollution standard and the forth-coming proposed rule on existing coal-fired power (due out in 2014) could have a significant impact on national emissions and eliminate some of the potential emissions increases associated with vehicle charging. Finally, other alternative vehicles are worth a mention. First, compressed natural gas (CNG) vehicles are not cost-effective in any scenario, including those with low natural gas prices, because low CNG prices are not enough to overcome the higher investment costs. Second, the model deploys diesel and diesel hybrids in many scenarios, which may be a cost-effective way to reduce CO2 emissions given their higher efficiency compared to conventional gasoline vehicles.'
the high efficiency of fuel cells means that even after reforming losses where NG is the source of the hydrogen they get double the mileage per gallon equivalent, with comparably reduced CO2 emissions.
So it is not safe to extrapolate from the judgements on NG they arrive at to assess fuel cell vehicles.
Personally without having a very thorough look at their modelling assumptions I have not got a high degree of confidence even in the conclusions they are specific about, let alone extrapolations! ;-)
Posted by: Davemart | 15 January 2014 at 01:27 PM
Davemart, the report states that low CNG prices are not enough to offset the higher investment cost.
That's bad news for fuel cell cars, because even with less fuel consumption through a hydrogen fuel cell running on reformed natural gas, that's not enough to offset the higher cost of natural gas. Natural gas is cheaper in every way than running on hydrogen, the engine is cheaper, the infrastructure is cheaper, there is no need for a reformer and 700 atmosphere pressure tanks (in stead you need a 2x smaller 150-250 atm natural gas tank).
The mileage of natural gas on natural gas ICE with battery hybrid is similar to the mileage of natural gas reformed to hydrogen in a HFCV battery hybrid car. With FCVs you start at 68 MPG but lose 30% of the natural gas in the reformer and 20% in the compressor leaving only 38 MPG on natural gas.
Posted by: A.C. R. | 15 January 2014 at 02:08 PM
One can wonder who paid for this certainly biased study.
The same study, with our 95% Hydro and 5% Wind e-energy would certainly come to very different conclusions.
Posted by: HarveyD | 15 January 2014 at 02:12 PM
ACR:
We disagree on what they were looking at.
They seem, as I stated, to be looking at the cost efficiency of saving CO2, not cost effectiveness per se.
I don't see anything there to bolster your established position of opposition to fuel cell vehicles without an unjustified stretch on what they are actually saying
As for your calculation, the 68mpge of the fuel cell car for the Toyota Highlander FCEV comes out to around 51mpge allowing for reforming losses, comfortably twice the petrol versions 20/25 mpg.
Your figure of 20% for compression losses is way high - what is your source?
It looks to me like liquifaction losses, not compression for fuel cell cars.
That runs at in the area of 10%, which is similar to the compression costs of NG for NG vehicles, although of course not identical, but ball park near enough.
For petrol cars, if you are going to count compression losses for hydrogen, then the losses in processing oil to petrol should be taken into account.
They amount to around 12% or so.
So however you cut it, fuel cell cars and hydrogen use energy around twice as efficiently as NG cars and petrol cars.
Posted by: Davemart | 15 January 2014 at 02:24 PM
Studies here on hydrogen compression efficiencies:
http://www.hydrogen.energy.gov/pdfs/review12/pd048_lipp_2012_o.pdf
http://www.hydrogen.energy.gov/pdfs/9013_energy_requirements_for_hydrogen_gas_compression.pdf
You are talking about 3kwh/kg, or near enough 10% to 12,000psi to create the pressure differential to charge at 10,000psi/
This assumes no use of waste process heat.
Posted by: Davemart | 15 January 2014 at 02:52 PM
A couple of points Davemart.
3-4 kWh for 10000 psi is electrical energy. In hydrogen terms that's 7 kWh in a fuel cell. So you invest 7 kWh of hydrogen to make 3-4 kWh of electricity to compress 33.3 kWh, that means 17% of your hydrogen is lost to compression. That's what I meant earlier with apples to apples. (liquefaction is far worse, for the same reason - liquefaction needs electricity).
Second, the FCV you're talking about is a battery hybrid.
http://en.wikipedia.org/wiki/Toyota_FCHV
So you should rightfully be comparing this to a natural gas ICE-battery hybrid, not a natural gas ICE only. That'd be more apples to oranges.
Comparing petrol to natgas reformed hydrogen is apples to oranges as well.
An apples to apples comparison would be the fuel cell battery hybrid mileage on natural gas versus a natural gas ICE-battery hybrid.
A converted petrol hybrid would easily beat 40 MPG on natural gas. The fact that that isn't done much should be very alarming for the reformed natural gas HFCV case.
Posted by: A.C. R. | 15 January 2014 at 03:11 PM
So let's run the number again, 68x0.7x0.83= 40 MPG.
Basically identical natural gas mileage as a natural gas battery hybrid.
Posted by: A.C. R. | 15 January 2014 at 03:13 PM
Lots of *opinions* on here but...meh.
Davemart nailed it: Without a thorough understanding of the assumptions they made in their model, this whole discussion is useless. If they had an agenda, one way or the other, their assumptions could make this whole thing toilet paper.
Posted by: DaveD | 15 January 2014 at 03:35 PM
With 1B$ you buy approx 300000 small electric cars or 300 2MW wind mills.
The cars will save over their lifetime around 100000 miles /car x 300000 cars / 50 MPG = 600 M gallon = 24 M tons of CO2 if the electricity is 100% green
With a capacity factor of 40%, the wind mills will save approx 600MW X 2000 MWh/MW/year x 20 year x 1 ton CO2/MWh = 24 M tons of CO2 if it replaces coal power.
BUT if you want the cars to charge green, you need cars AND windpower, to save the 24M tons. If you start replacing coalplants with wind, you only need new windmills to save 24M tons of CO2.
If you only have 1B$ to spend, choose the mills.
If you have plenty of money, do both.
Posted by: Alain | 15 January 2014 at 03:43 PM
I did not include the price of the gasoline or coal you save.
The profit of replacing coal by free wind, is most probably much higher than the profit of replacing gasoline by not-free electricity
Posted by: Alain | 15 January 2014 at 03:46 PM
Coal plants are not efficient in a wind grid, so you're not getting anywhere near 1 ton CO2/MWh savings. Coal plants are like charcoal BBQs, you can't throttle them much without wasting fuel. Plus 40% CF for wind you only get in the best locations, lots of locations are getting 20%. In my country the average of all land based wind turbines is 20%, and only 20% of the power comes from coal. CO2 savings/$ with wind are very poor here. Finally, $1B for 600 MWp wind is $1600/kWp, that is on the low side. Recent US projects are well above that.
Coal is cheap, so replacing it with wind is not that good business. Gasoline is very expensive, so it quickly is cheaper to drive electric even if that electricity isn't "free".
With lots of assumptions its always possible to push the results in your preferred route. That's certainly a good point. MIT is usually pretty decent here though.
Posted by: A.C. R. | 15 January 2014 at 03:57 PM
All this CO2 talk and comparisons ignores one thing: The EVs help us avoid more oil usage. Which for us Americans means less need to invade other countries and/or spend money occupying the middle east and p*ssing off all the muslims that decide to come try to kill us.
For the rest of the world, it means you don't have to watch Americans invade everything and push everyone around LOL
I think ignoring all of that and focusing ONLY on CO2 is a bit shortsighted. Both considerations weigh into the equation.
Posted by: DaveD | 15 January 2014 at 05:29 PM
I wonder if in comparing car emissions to power plant emissions they included upstream emissions. The US gets more oil from Canada than anywhere else and that means Tar Sands oil: Not only does bitumen production have higher CO2 emissions, it is dirty in other ways; http://livinggreenmag.com/2013/03/13/video/tar-sands-oil-extraction-the-dirty-truth/
Posted by: ai_vin | 15 January 2014 at 05:31 PM
The problem with replacing coal with "free" wind is that backing up the wind isn't free. Germany has decided that the cheapest way to do that is to burn... coal!
Truly, this appraisal relies on the assumptions. Russia's Rosatom has an order book of 20 reactors and is in talks on another 40. If the USA dropped its ridiculous regulatory regime for a sane one, nuclear electricity would be far cheaper and also zero-emission.
Other game-changers are possible, like the metal-molten-air battery. 10 kWh/liter gives you 1900 kW in the volume of a 50-gallon fuel tank, enough to drive the most aerodynamic semi rigs well over 1000 miles. Such a battery in suitably rugged packaging would spell the end of the diesel engine in OTR trucks; nobody could compete if they were paying 40¢/mile for fuel compared to 20¢/mile for electricity. The difference is $12/hour at 60 MPH, an insurmountable advantage.
Posted by: Engineer-Poet | 15 January 2014 at 05:39 PM
To the commenters stating that natural gas hybrids will match fuel cell hybrids, that does not on observation appear correct. After reforming losses of 30%, a fuel cell Highlander for example gets about 50 mpg. The gasoline hybrid Highlander using similar Hybrid technology as found in other Toyota hybrids gets around 30 mpg in round numbers. 50mpg is much better than 30mpg. Think about it for a second, A Highlander is a tall, relatively heavy CUV. That it could get 50 mpg, territory reserved for truly much smaller vehicles with better aerodynamic profiles is quite astounding.
We forget that in gasoline guise, the ICEV maxes out at around 40% BTE and is not always at that level, even in a hybrid. A fuel cell (up to 60%) is actually most efficient at part load, conditions that perfectly describe the average driving environment.
Regarding compression losses, the only fair way to compare H2 and CH4 is to use the actual amount required for the task. Converting to a fuel cell H2 equivalent consumption number is not a fair comparison. The same would of course be needed for any required natural gas compression which would skew things in favour of H2 in a PEM fuel cell.
I agree with the posters that H2 has more than its share of problems but we must be honest with our comparisons.
Posted by: otec_sheldon | 15 January 2014 at 07:06 PM
otec - yes we must be honest here. Compressive losses must be taken into account, which are large for 10000 psi compression. About 4 kWh is needed, using hydrogen to generate that with 42% efficient fuel cell system (not stack efficiency, system efficiency to net electric) gets you a 9.5 kWh hydrogen consumption to compress 33.3 kWh (1 kg) of hydrogen. This is guzzling 22% of the hydrogen (9.5 out of 9.5+33.3). Yes you could use natural gas electricity but the numbers are only marginally better, and then only if you assume the most efficient CCGT. That's probably not practical; hydrogen compressors would become a major load on the grid which onsite hydrogen fuel cell compressors avoid. Compressive losses for natural gas are very much lower than for hydrogen.
So I'll take this as 68x0.7x0.78= 37 MPG. Yes this is very marginally better than 30 MPG.
It's surprising how marginally better the natural gas powered hydrogen fuel cell is over an optimized natural gas hybrid. It is not the factor 2-3 that davemart is talking about.
Posted by: A.C. R. | 15 January 2014 at 10:34 PM
To state 3 reasons when 1 and 3 are the same is a big flag.
If we are to start comparing e power or emissions it is best to factor in large uptake of intermittent renewable and grid stabilization services that V2G enables.
As (already respectable) portable e storage improvements are poised to expand exponentially.
There are too many positives to ignore.
I would optomistically expect the shift to be showing promising results for early adopters within five years.
Posted by: Arnold | 15 January 2014 at 10:42 PM
It shouldn't be too difficult for electricity generation in the US to be completely carbon neutral by 2030 by simply increasing the number of nuclear reactors --at existing sites -- up to 8GWe each. Combined with hydroelectricity, biowaste fueled power plants, and rooftop solar, electricity in America could be completely carbon neutral.
Marcel
Posted by: Marcel Williams | 15 January 2014 at 10:59 PM
ACR:
We are talking about compression to 12,000 psi in the station to fill a car at 10,000 psi.
My first link makes clear that to do this the losses in hydrogen to do the compression are currently 5%, not the 17% you claim (page 18):
http://www.hydrogen.energy.gov/pdfs/review12/pd048_lipp_2012_o.pdf
The theoretical energy cost is only:
' The theoretical energy to compress hydrogen isothermally from 20 bar to 350 bar (5,000 psi or ~35MPa) is 1.05kWh/kg H2 and only 1.36 kWh/kg H2 for 700 bar (10,000 psi or ~ 70 MPa). Greater compression energies are required to fill vehicles in practice due to compressor inefficiencies and heating during fast fills.
http://www.hydrogen.energy.gov/pdfs/9013_energy_requirements_for_hydrogen_gas_compression.pdf
You are double counting the losses.
I did not exclusively compare efficiencies of PHEV FCEV to natural gas and petrol, in fact I gave no specifics on this, as I don't have any.
I clearly used the figures I have, which are for petrol against hydrogen in a fuel cell, and my figures for natural gas after that is also compressed are based on:
'A CNG-powered vehicle gets about the same fuel economy as a conventional gasoline vehicle on a gasoline gallon equivalent basis.'
http://www.afdc.energy.gov/fuels/natural_gas_basics.html
My reference to PHEV FCEV was made in the event that hydrogen costs prove to be at the upper end of expectation, ie around the same as petrol, as other estimates put is as as low as around a third of petrol costs:
http://www.hydrogen.energy.gov/pdfs/progress12/ii_0_dillich_2012.pdf
The aim there would be not so much to save energy as to save money by using cheaper electric for running around.
So I used no apples to oranges comparisons, and the energy costs I have quoted for compression are in the right area.
Hydrogen used in fuel cells is around twice as energy efficient after all losses as petrol or natural gas.
Posted by: Davemart | 16 January 2014 at 01:26 AM