This article is an opinoin piece from a Pangea SI expert
First, a general note about how we deal with efficiencies when converting fuel chemical energy into other forms. Whereas the actual energy of a fuel is probably better represented by its higher heating value (HHV), which takes credit for the heat of condensation of product water from the combustion gases, non-heating uses of fuels lose that heat of condensation and can’t put it to use. All my comparisons are therefore biased somewhat in favour of using fossil fuels, by virtue of calculating the ultimate energy available in a fuel based on the LHV rather than the HHV.
We know that the Achilles’ heel of the ICE vehicle is the engine itself. Being a heat engine, i.e. a device which converts heat to work, it is limited by the 2nd Law of thermodynamics, though it is not a closed system heat engine like a steam engine and hence does not need to reject heat to a cold reservoir per se- the atmosphere accepts the combustion products directly. There are materials limits to the hot reservoir temperature, and NOx emission concerns typically keep us well below even the limits based on materials. While a stationary fixed speed power plant engine or ship-board large diesel may achieve an efficiency as high as 55%, achieving an efficiency of converting chemical energy in fuel to heat beyond about 40% of the lower heating value (LHV) of the source fuel in a smaller vehicle like a car is very challenging indeed to achieve. A 40% efficiency, from tank to motor shaft, can reportedly be achieved by a hybrid like the 2017 Toyota Prius running the Atkinson cycle on gasoline.
You can reportedly do even a little better than that- perhaps 44%- with a turbodiesel on the highway. That’s based on past estimates of highway performance for small diesels that I’ve heard anecdotally. The best I can find on the Canadian fuel economy website (Natural Resources Canada) based on real testing right now is a 52 mpg(US) (4.5 L/100 km) highway figure for a diesel, the Chevy Cruze, versus a 51 mpg(US) (4.6 L/100 km) highway figure for the Prius on the highway- given the ~ 10% density difference between diesel and gasoline, that’s actually LESS efficient than the Prius- and even worse in GHG terms because diesel has a higher C/H ratio. Even the tiny diesel Smart Fortwo only achieved 50 mpg (US) (4.7 L/100 km) on the highway, and it’s no longer available (they’re going 100% EV, which makes sense to me!). The Canadian test cycle is a little different than the US EPA test cycle so the results vary a bit by vehicle.
While stationary or ship-board diesels in constant speed applications can reportedly achieve as high as 55% thermodynamic efficiency, vehicle applications are much more troublesome, with small vehicles being worse still. As this reference makes it quite clear, achieving efficiencies with IC engines beyond 60% is a very challenging problem, where fixing one issue to improve efficiency frequently results in increasing another problem which saps efficiency.
The best plainly stated diesel vehicle engine efficiency I could find after a brief search was in fact 44%, and that wasn’t yet in vehicles at the time of publishing:
Regrettably, the 44% best case figure reported for the diesel drops dramatically in mixed city/highway driving while the Prius improves, making the Prius substantially more efficient overall in mixed driving than any other ICE vehicle on the road- until we start adding electricity from the grid. The toxic emissions from the diesel are much more substantial than those from the Prius, even when the emissions control systems are maintained properly and not being defeated by software… Furthermore, hybridization doesn’t seem to give the same gains to diesels as it does to gasoline engines. So, we’ll stick with 40% out of a hybrid ICE as a best case figure for now.
So, let’s run some well to wheels numbers, shall we? Well, that’s harder than you might think, even though we’ve got a pretty good handle on the well to tank portion of the trip for gasoline (round numbers 80% once we give fair value to the electricity used in the refinery). You would think we could just multiply that by the reported efficiency of the Prius or the diesel and get the well to wheels result, but regrettably most of the figures given for efficiencies aren’t to wheels, but rather to motor shaft.
Fortunately, unless you’re slugging along an inefficient automatic transmission, driveline losses from a manual transaxle and output shafts with their CV joints are small- in the 2-5% range, so we’re going to ignore them in this analysis since all vehicles- BEVs included- include a gearbox and output shafts. Then there’s the other parasitic losses due to auxiliary devices such as air conditioning, lighting, and topping up the old lead-acid starter battery with the alternator (itself a lossy but inexpensive device- one that is replaced with a far more efficient DC/DC converter in a BEV).
It gets really complex, really fast, and the only thing that ultimately matters is the more practical assessment of efficiency: how many miles you can drive on a gallon of gasoline, or if you prefer, how many L of fuel you’ll consume per 100 km? Well now, the difference in density between diesel and gasoline kicks in, so even that’s not a fair comparison…This really IS complicated if you care to be accurate about it!
So, let’s start with my own E-Fire electric Triumph Spitfire. I’m achieving a mixed highway/stop and go driving cycle on my daily commute using about 235 Wh/mile (143 Wh/km) out of the battery, which requires about 260 Wh/mile (159 Wh/km) out of the landscaping socket I recharge it from, once the cycle efficiency of the battery and charger are taken into account. Three quarters of that trip is done at highway speeds in the 50-60 mph (85-100 km/hr) range, and the balance varies a lot depending on just how bad the traffic is in the afternoon. That’s based on measurement of the AC power fed into the charger using a Kill-A-Watt meter, and an accurate measurement of the energy leaving my battery measured using my fuel gauge, a TBS E-Expert Pro intelligent amp-hour meter supplied by some clever guys in the Netherlands, averaged over ten thousand miles (16 000 km) of driving. It happens to match with many similar measurements made by other converters. That gives us an effective socket to “tank” output efficiency of about 90% for the E-Fire.
Grid losses in North America average about 6%. My own local grid loses less than that- closer than 3%- which is reported in the “distribution costs” section of my electrical bill. So, source to “tank”, we can take the distribution loss at 6% and we’re at 0.94*0.9 = about 85%. That’s impressive already- remember that the source to tank efficiency of gasoline is only 80%. The key is the Li-ion battery, which has a tremendous cycle efficiency.
Now comes the EV drivetrain. The inverter and AC induction motor in my EV when running at highway speeds are roughly 90% efficient from battery input to motor shaft output. OEMs reportedly do even better than that. As I hinted before, we’re going to ignore the unknown but small losses in my manual transmission, driveshaft UJs, the old Leyland differential and the UJs on the output shaft to each wheel. We’re left with 85% x 90% or roughly 76.5% from energy source (i.e., power plant gate) to wheels.
Let’s say we don’t like Ontario’s amazing 40 g CO2/kWh, 9% fossil (all natural gas) grid as a basis of comparison because it’s just too good to be true. Let’s throw in a modern natural gas combined cycle power plant at 60% thermodynamic efficiency, and natural gas recovery and distribution losses (from the GM/ANL GREET model estimates) of 97.5% and 97.5% respectively. That’s 76.5%x0.6*0.975*0.975, or roughly 44% from energy source to wheels.
For comparison, let’s use the Prius- I own one of those, and I routinely achieve the EPA performance in mixed driving of about 52 mpg (4.5 L/100km). Let’s assume for a moment that my Gen 3 Prius is about 38% thermodynamic efficiency and remember that we’re 80% efficient well to tank for gasoline. Again, we’ll forget about the differences in the final drive, which are small. That puts the overall Prius source to wheels efficiency at 0.8*0.38 = about 30% efficiency. And that’s the best ICE car you can buy…even fuelled by a modern gas power plant, the E-Fire beats the pants off the Prius- and does so with zero tailpipe emissions for passers-by to breathe.
Replace the 60% efficient powerplant with a 30% efficient coal plant and things become less flattering. Let’s give the coal the benefit of ignoring the energy lost in mining and distribution for a moment, because frankly I don’t care to know what it is- coal in my mind is dying a well-deserved death. The E-Fire would drop to a pathetic 23%, and energetically, we’d likely be better off with the Prius. We’d likely be better off with the Prius from an environmental perspective too, unless the coal plant’s flue scrubbers were exceptionally good- which is not likely on a plant this inefficient. (At this point, all the EV detractors would stop reading, if they got this far, safe in their pre-existing knowledge that fossil fuels are still king!) Fortunately, for those of us with open minds, we realize that there are very few grids in the world that have anything close to 100% of their power produced by 30% efficient coal-fired power plants, and grids worldwide are getting greener every day.
We do have another, more meaningful direct comparison to make in this case which is useful, and that’s the same car pre-conversion. The Spitfire, with its notoriously unreliable 1493 cc 4-cylinder engine, did about 8 L/100 km if driven conservatively (which it was admittedly hard to do). By comparison to the Prius, that’s 8/4.5 or roughly 1.8x as much fuel per mile- and the efficiency of the car pre-conversion well to wheels drops to 30/1.8 or 17% efficient. The EV wins hands down, even when re-charged from an inefficient last-generation coal power plant. And the fleet average ICE car in North America isn’t a Prius- it’s in fact worse than what I was getting from my 1970s car: it was 24.8 USgpm (9.5 L/100 km).
On a greenhouse gas emissions basis, the EV shines even brighter. I won’t bore you with my own calculations on that for now, but instead will refer you to the Union of Concerned Scientists report which does the evaluation on a lifecycle basis, so that people who are concerned about the embodied energy of the battery pack can’t just shrug off the result. They concluded that battery EVs are “Cleaner Cars from Cradle to Grave”. I’ve reviewed the report in detail and can’t find a flaw in their evaluation. Their re-release of the report’s conclusions updated in 2015 showed that indeed the US grid is getting greener, and EVs with it.