Well, let's do the calculation. In the UK there is approximately 100GWh spare capacity during the night low. If you use all unused capacity (max is about 55GW, say we use 50GW 24/7), then there is an additional 250GWh per day. This is all existing capacity, no extra power stations, links, storage or anything will be needed - we just use what we already have installed more effectively.
Assume we all have a Model S with a 84kWh battery. Charging it from totally empty to full requires about 100kWh (charger and distribution losses of 16% total). That means you could charge a million cars on the spare night capacity or 3.5 million with the total spare capacity.
There are 28.7 million cars on the road in the UK, so all cars could recharge once every 29 days if only charged at night, or once every 8 days if charged using unused day capacity. Assuming you get 250 miles per charge (instead of the 300 advertised), this implies 3100 and 11000 miles per year respectively. Estimated annual mileage for all 4-wheeled cars in the UK was 8430 in 2010 (this includes business vans so mileage for cars will be lower).
So despite using very conservative estimates, the current UK grid as is can easily handle all cars being converted into EVs. Perhaps a surprising answer - but a result of the much higher efficiency of electric power!
Even if there is supposed "capacity" in the grid, that doesn't mean the extra electricity needed comes from nowhere. So what I was asking is, how many times the the current load on the electric grid, would be created if all automotive machinery fed off it?
You claim 5X times as much demand, and even then, this only holds if vehicles recharge once a month!
Oh, I don't buy this 300 mile range propaganda. That might be true, for a handful of types, under absolute ideal conditions maybe. The real-world tests end up showing anywhere from 35 to 60 miles or so. Not quite enough to credibly run your car for a month, at the cost of quintupling the amount of electricity has to be generated.
Anyway, that's the computation people need to see. Not so much these claims about "if everyone plugs in at night only," as if that can be the complete solution.
Nowhere did I say 5 times! The extra electricity you need to generate is just 33% for 11100 miles per year for all cars in the UK. However my point is that you don't need to build any new generation capacity for this - it already exists, but as yet is unutilised. This is where a smart grid doing load balancing using EVs becomes interesting.
Any claim that real-world tests show you get only 10-20% of the claimed range requires some serious evidence to back that up - I presume you do have credible links that prove this rather than just making that up?
This is exactly the sort of FUD that one often gets when discussing wind/solar/EV. If you have more accurate numbers (which you can back up with links) then please go ahead and improve upon my calculations.
"In the UK there is approximately 100GWh spare capacity during the night low. If you use all unused capacity (max is about 55GW, say we use 50GW 24/7), then there is an additional 250GWh per day. This is all existing capacity, no extra power stations, links, storage or anything will be needed - we just use what we already have installed more effectively."
And then, you went on to compute how to distribute that 250 GWh to all cars on the road. And concluded that it could be done by "refueling" each electric car just once a month. That sounds like 5X the amount of electricity to be generated, and limiting everyone to one battery charge per month to achieve that.
I understand that all the proponents of battery-powered electrics like to obfuscate the picture, by harping on this "no extra distribution systems needed." That's what I was objecting to. That extra energy comes for something. It's either more coal or natural gas being burned, or more nuclear spent fuel, or more windmills where people complain about them, and so forth.
On the other hand, get rid of the battery altogether, as the energy storage medium, and then we can be rid of all the gimmicks and apologies. EVs don't have to be just battery powered.
Oh, as to range, yes, take a look at most claims made by EV manufacturers, and actual tests, and it ends up being in the range I stated. For instance, the relatively new Ford Focus, a claimed range of 76 miles.
And these are all small cars, let's not forget. Compacts, in US terminology. So yeah, I don't buy calculations that depend on your average-sized road hog running 250 or 300 miles on one battery charge.
Bert, the Ford Focus has just a 23kWh battery so no surprise it's range is much smaller too!
Tesla claims 300 miles, the EPA achieved 265 miles and the European test achieved 310 miles. People have done drives over 420 miles on a single charge (lots of links if you Google it), so it is certainly feasible to go well beyond the advertised range. Alternatively if you use it on a track day your range will be a lot less. It all depends on how you drive it, but my 250 mile estimate for a real-world range is reasonable.
Bert, you are confusing GW (power) and GWh (energy). You can't do 250GWh / 50GW = 5x - those are completely different things! Lets restate my figures using GWh only for clarity:
Actual daily use is about 1050GWh. I assumed 350GWh (33%) extra generation for EV based on existing spare capacity. That allows you to recharge every 8.2 days and drive 11100 miles per year assuming 250 miles per 100kWh charge.
So yes that extra 33% electricity needs to be generated and will of course require extra gas/coal depending on the power generation mix. However the interesting thing is that even if the extra electricity was generated from coal, that still generates less overall pollution and CO2 than using petrol.
When you say "get rid of the battery altogether, as the energy storage medium", what exactly do you propose we use instead? The energy needs to come from somewhere, be sustainable and low carbon, what options are there besides electricity from nuclear, wind and solar? If you use a different storage medium, electricity will need to be converted to that and back. For example generating hydrogen from electricity and then burning it in an ICE is one of the most inefficient things you can do.
the average US home uses 940 KWh per month. And if you assume 28 mpg cars, and 1250 miles per month average driving distance per car (all fairly optimistic values for the US), that means that the average car needs 44 gallons of gasoline per month. Okay, that's realistic, and btw, a whole lot of what people call "car" these days won't come close to that economical, and that's just one car in the household.
Gasoline contains 10 KWh per liter of energy, or 37.85 KWh per gallon. So that means that this car is using 1665 KWh per month, compared with 960 KWh for the average home. Which is more than 1.7 times as much energy. One car in the household. Households have 2 or more cars, perhaps not all driven as much as 1250 mi/mo, so to say that we need to double the amount of energy delivered by the grid, to individual homes, seems hardly an overstatement.
Electric cars are more efficient, no doubt, but if there's a major shift to electric cars, we're not talking about everyone in a Nissan Leaf, either. That figure of 44 gal/month is pretty conservative for most households in the US, I'll wager. I'm sure the readers can decide on their own whether that's reasonable or not. So my contention is, the extra amount to electricity to be generated by the grid, if we all start driving battery-powered electrics, is substantial. Not negligible.
I think a good bet for the not-too-distant future is H2 reformers on board, creating H2 from some sort of hydrocarbon fuel (could be biofuel), feeding the H2 directly to a fuel cell. Avoid the large storage battery (perhaps a hybrid-size battery is good enough), avoid distributing H2 or storing H2 in a tank, avoid all the energy starvation batteries create. No apoligies, and a real electric drivetrain.
Bert, those are very interesting calculations from the other perspective. However note the US residential electricity consumption is only 36% of the total, so doubling it means 36% extra generation. Obviously if your electrify a big gas guzzler it wouldn't be as efficient as a purpose built 100mpge EV, but we can safely assume doubling efficiency to 56mpge will be feasible. Assuming that, the total US electricity generation would go up by just 31%. Which seems quite feasible, right?
Yes I like fuel cells, but it is a shame they cannot handle anything but hydrogen effectively - the input needs to to excruciatingly pure. Reforming is pretty inefficient even on a large scale. And while batteries are expensive, what about the cost of a reformer and fuel cell? Will it last 100000 miles? EV's are on the market today and moving towards mainstream. I don't want to be pessimistic, but it seems fuel cell cars are still 10 years away.
My feeling is that the best solution to the range anxiety of EVs is a small efficient generator module using LPG or biofuel.
Wilco, yes, I looked into the overall efficiency of a fuel cell car and on-board reformer. For moderate levels of power output, the combined efficiency of the entire drivetrain, including also the electric motors, came out to around 48-50 percent.
So you might say, not that great. But it is that great. Because this compares to the actual efficiency of an internal combustion engine, in real-world driving, which is only about 18 to 20something percent. (The 30 percent figure often cited is, like so many such, ONLY under absolute ideal conditions. The inescapable limit being, the difference in Kelvin between the combustion peaks and the exhaust temperature.) So this fuel cell approach easily doubles the efficiency, and consequently creates a lot less CO2.
At high output levels, fuel cells lose some efficiency. But a hybrid-sized battery should take care of that. High output is used rarely in cars, e.g. in passing or moving away from a stoplight. The battery would provide that spurt of energy.
Car companies are showing renewed interest in fuel cells, as also reported by EE Times recently. My bet is, because battery storage by itself is just too compromised of a solution. Fuel cells capable of using less than pristine H2 are also being developed, with there again some loss of efficiency, but overall remains somewhere above 48 percent.
I would agree that your LPG or similar generator solution is the best in the near term, but my hope is for an uncompromised "true electric." With very few moving parts, and not depending on the Carnot cycle in any way.
Modern engines are far more efficient than you think. The petrol engine in the Prius is 37% efficient, and a modern diesel can be 42.5% efficient. These are actual engines on the market, not laboratory setups: http://en.wikipedia.org/wiki/Brake_specific_fuel_consumption (also check out the amazing 51.7% efficiency of the enormous diesel engine used in Emma Maersk).
I agree 100% electric is the best solution. But I believe the roadmap for automotive fuel cells will be bumpy: at one side they are being squeezed by ever more efficient engines (coupled with thermogenerators to harvest electricity from the waste heat). On the other hand batteries continue their slow but steady improvement in capacity/cost and soon provide a range and recharging time that is good enough for 99% of people.
I'm sure fuel cells will find a niche - my guess is CHP for homes: cost/size/weight are no big issues, they can run on relatively clean methane, need no start-stop cycling, and the waste heat is used for hot water and central heating, so you can get 80+% efficiency.
But once again, I am talking about reality, not propaganda.
Here is a quote from a Wikipedia article, to show you what I'm getting at:
"The most efficient type, direct injection Diesels, are able to reach an efficiency of about 40% in the engine speed range of idle to about 1,800 rpm. Beyond this speed, efficiency begins to decline due to air pumping losses within the engine. Modern turbo-diesel engines are using electronically controlled, common-rail fuel injection, that increases the efficiency up to 50% with the help of geometrically variable turbo-charging system; this also increases the engines' torque at low engine speeds (1200-1800RPM)."
"The gas turbine is most efficient at maximum power output in the same way reciprocating engines are most efficient at maximum load. The difference is that at lower rotational speed the pressure of the compressed air drops and thus thermal and fuel efficiency drop dramatically. Efficiency declines steadily with reduced power output and is very poor in the low power range - the same is true in reciprocating engines, the friction losses at 3000 RPM are almost the same whether the engine is under 10% load or not having any useful output on the driveshaft."
So, the efficiency of modern internal combustion engines in actual operation in cars, *especially* in non-hybrid cars, is way lower than the efficiency at ideal operating modes. If instead you use something with an electric drivetrain and fuel cells, that efficiency does not drop dramatically at lighter loads. Quite the contrary.
So, the 48 percent efficiency of the fuel cell care with H2 reformer in no way can be compared to the peak-optimistic-never-actually-operates-there 40 or 50 percent figure cited for diesels. In diesel hybrids, during city driving only, MAYBE.
You're right that engine efficiency varies with rpm and torque. However due to 6-speed transmissions (and CVT on hybrids) one can keep an engine very close to or at the optimum.
If you look at the graph in the same article, the 225 g/kWh line (ie. 36% efficiency) spans from 1200rpm to 3300rpm and a factor of 2 in torque (y-axis). That is more than the range one would need for typical driving. So engines are far more efficient than you think - the only case where an engine goes down to 20% efficiency is if you rev it to the maximum rpm without any load (the idling case is handled with start/stop). Hardly a realistic scenario.
New engines like HCCI will come to the market in the next few years, further increasing efficiency. So unless that 48% for a fuel cell with reformer can be improved as well, it seems unlikely it is going to be able to compete with future engines. And yes you have to compare fuel cells with engines running at their highest efficiency as that is how most hybrids work.
A related question: could I use solar panels on my roof to recharge my car?
The standard size is a 4Kwp panel which in the UK generates about 3000kWh per year. That's 30 recharges of a Model S battery, so you can travel at least 7500 miles using solar power alone. That's very close but not quite enough to cover the average UK annual mileage of 8430. You'd need 4.5Kwp in the UK or live in a country with 10% more sun!
Rick, the way it works is that when your panels produce power, and you are not charging your EV, the electricity will be used by your neighbourhood. As a result power stations do not need to generate what your panels generated. If you recharge at night, extra power needs to be generated. Overall power stations use slightly less gas/coal then they would if you recharged during the day peak. So you can use the grid as "storage" by changing the timing of when electricity is generated. This doesn't cost anything extra, in fact it is cheaper.
The fact is that solar power comes just when EVs need it - when most of them will be in company parking lots. Subsidise companies to install lots of PVs on their roofs and bingo - green EVs.
I appreciate that this is a bit simplistic. Not all companies have vast warehouse type buildings suitable for PV. And sales reps, for example, are always in their cars during the day. But it's horses for courses, and there are some horses for which this course will be very suitable indeed.
And electricity can be easily transported, on existing infrastructure, to where it is needed. Can someone do the sums to work out what power would be generated if every roof in the country was covered in PVs? The only challenge is how to use it efficiently - which to a large extent means storing it, in EVs or otherwise.
To get an idea whether roof-top solar panels could provide enough energy, let's take the annual electricity consumption per household: 4648kWh in the UK. If you added an EV as per my previous calculations, it would need another 3400kWh for average UK mileage. You would need 11kWp of panels or about 70m^2 of south facing roof (250W panels are about 1.6m^2). That's more than double the size of a typical UK roof.
The US consumption is 11698kWh and assuming 2 EVs per household at 8400 miles per year you'd need between 10kWp in LA to 17kWp in Seattle, or 64m^2 to 109m^2. Despite more sun, consumption is still too large, so panels won't fit on an average US home.
So the conclusion is that with current efficiency and usage solar panels need far more than your roof area to supply all your electricity. If overall efficiency doubles in the next 10 years or if solar films become cheap enough to use on every wall and window then it might become feasible.
@Wilco1....you know your stuff, thanks for that. Couple of points....
This is obviously not feasible right now....however there are not too many EVs right now. We *should* get more energy efficient in our homes (though that is debatable, our appetite for energy seems to outstrip this), and solar panel efficiency should rise.
The average home can't make enough for the home use AND EV charging with current efficiencies. However there are vast roofs on shopping malls, industrial premises, etc which are often unused
Every little bit helps - especially if it can help reduce or even out base load generation requirement - which is where most of the fossil fuel usage is.
We shouldn't be scared to start something here because it looks like we'll never be able to fully meet our requirements....
IIRC, at a recent Embedded Systems Conference the CTO of Tesla said in the keynote address that you need about one carport's worth of solar panels to provide all the usual charging needs of a Tesla car.
Blog Doing Math in FPGAs Tom Burke 24 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...