Two things. First, these charts are for energy generation, measured in billions of kWh. That is very different from the problem for electric vehicles, which is power generation and delivery.
Since I can't know if readers understand the difference (you might, others may not), here's a simple example:
You have to lift 1000 kg of metal to the top of a massive 1000 meter tall skyscraper being built. It's a thousand 1 kg pieces. Let's say it takes 10 minutes to walk up to the top.
You can do this at least two ways:
One person moves each 1 kg piece up to the top, one at a time. That will require 1000 times 600 seconds, or 600,000 seconds. That's roughly seven days.
The other way is to do it all at once. We get a thousand people, each one of them grabs a 1 kg piece and, ten minutes later the entire load is at the top of the building.
Anyone can understand that these problems require vastly different scenarios.
Energy is about the total amount of work done. In this case it is the same for both scenarios.
Power, on the other hand, is a measure of how quickly this work was done. In this hypothetical, the fast delivery scenario required a thousand times more power than the slower version.
Electric cars require power. Energy is delivered over time. The problem with our grid and power generation is that it simply cannot handle this beyond a very small deployment. People driving electric cars today are enjoying the fact that they do not pose a significant threat to the existing infrastructure and generation capacity. As the numbers increase we will start having problems.
The idea of having a million or ten million cars simultaneously fast-charging today is likely an unthinkable reality.
Put a different way (just pulling hypothetical numbers out of my behind, these are not real numbers):
If we could charge electric cars over, say, 30 days, we might not need even one bit of additional power generation capacity or changes to the infrastructure. Sure, nobody would be able to drive more than once a month, but, hey, we don't need to build nuclear power plants.
However, that is not a practical reality. Tens of millions of cars will need to quick charge in 20, 40, 60 minutes. And another set of millions of cars will want to be charged over, say, 8 hours, overnight. This requires the ability to deliver power --large amounts of it-- in a way we are not prepared to do. This is both true in terms of power generation capacity as well as infrastructure, from the cables all the way to the power plants.
What the page you linked shows is an increase in energy consumption. It says nothing about power. Also, the increase you point out is only 60%, not a doubling from 1980 to 2020. Still, this is somewhat meaningless. I can consume 60% more energy if I keep the lights on 12.8 hours per day, instead of 8. I would not demand 60% more power out of the grid, I would just use the available power for a longer period of time. That's why speaking in terms of energy isn't relevant to the electric car problem. The problem is power, not energy. People need their cars charged now, right now. That requires power we cannot deliver.
I will acknowledge the above chart is delivery not peak power generation, but to an extent the two are related. If you don't have generation capacity, you can't increase consumption. I wish I had a chart of generation capacity over time, but I couldn't find one. Still though, it still seems to me the change from 1980 to 2000 probably indicates a pretty large change in power capacity, it's not like they managed to deliver a ton more power without making any changes to the grid and production capacity. You really think the only change in our usage in that time period is we left the lights on 60% longer?
I will acknowledge I did only eyeball the chart and saw it was about 2k going to about 4k, the values are 2,290/3801 so about 60% as you say and not doubling. My bad.
Still though, your argument of leaving the lights on longer and power capacity points to how EVs are still possible even with this kind of usage vs power argument. Charging my EV at home doesn't increase my home's peak power much at all. It's another 50A circuit on at my house, but it's not like it's running all the time. It's mostly running when most the other appliances in my house aren't running. So there's already plenty of capacity to serve my house 50A, my neighbor's house 50A, on and on, even with the current hardware because it was already built with the idea a lot of people are going to get home from work and start cooking dinner with an electric range and electric oven together use considerably more power than charging my electric car. For those who can charge at home, it really shouldn't change the math on power at all.
I do agree power demands would increase a good bit with DC fast chargers, those would be the biggest change in our grid infrastructure. Some of the arguments of having a good bit of local battery storage to smooth out those peak power loads sounds incredibly expensive to me. However I don't think those would necessarily make up the bulk of actual charging usage. The majority of Americans live in single family homes, implying they have some kind of garage where they can charge their car and would thus have the same power argument made above. Decent L2 chargers at offices can provide adequate charging to those commuters without charging capacity at home. Obviously this wouldn't have the same time of use shift as above, but L2 charging wouldn't impose nearly as much power demands as DCFC. While DCFC is somewhat important for mass adoption of EVs, I think you might be overestimating the actual usage of DCFC. The majority of Americans would rarely need them. I haven't used one once in the many thousands of miles I've put on my EV so far. A huge chunk of the power charging so far was 12A@120V, would you really argue every home having a load like that running overnight is going to melt all the transformers?
And FWIW I'm still a proponent for using diesel/petrol where it makes sense. Please don't take me as an EV fanatic thinking that Elon's self driving pure EV semi's are going to completely change the trucking industry anytime soon.
> You really think the only change in our usage in that time period is we left the lights on 60% longer?
Of course not. This was just an example to attempt to drive the point that there's a very real difference between energy and power. You can deliver more kilowatt-hours (energy) without changing kilowatt (power) generation capacity. The bottom line is that energy metrics or charts are impossible to use in this discussion.
> your argument of leaving the lights on longer and power capacity points to how EVs are still possible
In small numbers, sure. At scale, 300 million vehicles. No. Nobody is going to wait 16 hours to charge their cars. The infrastructure has to be able to deliver double the power we can deliver today.
I read through your reasoning about the 50 A circuit at one or many homes. The problem with this reasoning or story is that, while it is easy to say these things, until you sit down with Excel and quantify it all with a reasonable model of reality, it is just words. Once you do that you will very quickly realize a full migration to electrics is absolutely impossible with the current power generation and distribution infrastructure.
As I said before, I wanted to understand this in some detail. I kept hearing these claims about going electric, yet nobody showed any calculations. As an engineer, I had to put numbers to the claims.
The model I created about five years ago simulated a fleet consisting of 300 million vehicles of various types. This being US-centric, I divided the fleet into six time zones based on current population distribution figures. And, within those time zones, I created variables to manipulate utilization scenarios. For example, a percentage of the population would only drive 20 miles a day and would be happy slow charging over eight hours. Another portion of the population required fast charging a few times per week (pretty much as one might get gasoline today). Some would be hybrid: plug in at home every night and top-off with fast chargers during the week as needed. Yet another part of the model looked at some percentage of those vehicles being mid and long distance trucks, which simply cannot afford to sit idle for eight or more hours while charging. Etc.
The point is, the model wasn't a simplistic statement of the kind that is often offered in these discussions, things like "We have more than enough solar to do this" while nobody bothers to put some numbers to it.
Having said all of that. The best confirmation of my numbers I can offer is Elon Musk himself explaining in that video that we need to double our power generation capacity and redo our entire infrastructure. Until Elon was asked this question I was, to be dramatic, entirely on my own making these kinds of claims. People want to religiously believe that these things are possible. They are not. So, don't believe what I say, but, please, do believe what Elon Musk said --which happens to confirm my findings from many years ago.
Assuming both Elon and I are not completely full of shit, well, reality is that the fully-electric transportation dream will require leaving this cult-like delusion behind to put into place a realistically attainable plan. This plan requires --absolutely requires-- cheap oil. Oil is what we have to use to build the very infrastructure required to support electric transportation. Without cheap oil we cannot get there. That is VERY important to understand. Without an admission and an understanding that petroleum is critical in making the all-electric dream a reality, there is no way to make it happen.
And so, the plan must include a well-defined and intensely funded period, about 50 years long, where we drive local petroleum costs as far down as possible in order to be able to engage in a massive construction project to revamp power generation capacity. This plan should also include a vey serious commitment and a solid plan to build hundreds of nuclear power plants. We cannot do it without nuclear. And, yes, solar and wind must be a part of this.
The precise proportions and sequencing of the above will be a matter of detailed analysis and, yes, math. None of this is particularly difficult --outside of the politics and ideological delusions we must leave behind. It's just a matter of starting to bring the truth to the forefront and having the adults in the room lay out a clear, attainable plan for a transition to all-electric transportation within the next 50 years.
To drive the point home: If you feel you still need to argue against my claim that we need to double power generation and infrastructure, you need to argue against Elon Musk. You are free to discount me as some nutcase on HN. Elon, on the other hand, well, he is confirming what I am saying. Not sure how anyone can argue against that one.
> a full migration to electrics is absolutely impossible
I'm definitely not one to argue the full migration to electrics. As mentioned elsewhere I think diesel/petrol/similar kind of fuel vehicles will be around in at least some capacity for quite some time. Its incredibly energy dense and the ability to just pump up gallons of it from pits in the ground is quite handy.
> Some would be hybrid: plug in at home every night and top-off with fast chargers during the week as needed.
I really can't imagine this would make up any large fraction of the EV fleet. If someone can charge from home, the extreme majority of their charging will be overnight or potentially during a weekend if they need to do some catch up. The vast majority of commute cars which can charge at home will not need this hybrid approach save for the extremely rare long road trip when comparing the total number of miles.
> percentage of the population would only drive 20 miles a day and would be happy slow charging over eight hours
That 20mi commute would generally use ~6kWh. Over 8 hours, that's 750W of power. During a time where power at people's homes is often at its lowest point. Delivering that power would barely be a blip. If our grid collapses because a few million people plug in a tiny 750W space heater overnight we've got some serious problems.
> If you feel you still need to argue against my claim that we need to double power generation and infrastructure, you need to argue against Elon Musk
I would, especially when it comes to long-haul trucks. Arguing against Elon is then easy in this regard, where's the long-haul driverless trucks which were going to revolutionize the entire trucking industry in 2017? EVs in this kind of capacity don't make sense without massive changes not only to the grid but also to our battery technology. Just making the batteries a little fatter isn't going to change the math on these kinds kinds of trucks. So you should pretty much just completely eliminate these kinds of vehicles from your model, its a moonshot to even think these kinds of things are really going to switch to EV-only anytime soon for many reasons.
I don't doubt we'll need to make a lot of changes to our grid. I agree it will take a lot of energy to make those changes and it will use a lot of petroleum based products to make those changes. There's oil in practically everything we touch, so I agree its short sighted to praise the idea of expensive oil. I also agree this won't be something that happens overnight, it'll be something that will take an effort over decades.
I agree moving to EVs is overly praised and there's quite a cult mindset of if we just put an EV in every garage we'll have solved some big problem. As you mentioned its often a question of power, and well at least down here in Texas overnight we often get to the point where there's not even enough load to match the generation available for free. On many grids there's already a good bit of slack in the power distribution in off-peak times, probably enough to support every average commuter car in every single family household. Note that term, average commuter car. Not the people which somehow are still sane after a 100+mi daily commute, they're extreme outliers and yeah it will probably take a lot of work to get to the point where all of those people are reasonably served with EVs.
I think there's a bit of a disconnect between us here. You're arguing about the full migration to EVs, while I mostly agree a full migration to EVs isn't really practical anytime soon. While a large part of the average commuter car and short haul vans can probably make the switch to EVs today, it seems there are plenty of use cases where EVs are still impractical vehicles. This is before even figuring in your grid arguments. So once you then eliminate those vehicles where EVs are currently a somewhat impractical case at the moment, the math of how reasonable grid upgrades get becomes a much more practical number.
Also, FWIW I do agree there will need to be some amount of an upgrade to the grid, and it will probably be a large one. Shifting all that energy which used to be handled by fluids running through pipes into cables overhead will inherently mean some amount of capacity upgrades. However, I think for average commuter cars, this really won't be that much of an impact. There are definitely cases where there will need to be a large change (DCFC corridors) but I doubt the average suburb would need some radical change. Its just adding load to a time of day that routinely experienced its lull.
> If our grid collapses because a few million people plug in a tiny 750W space heater overnight we've got some serious problems.
Again, if you want to understand this well, I urge you to stop and create your own simulation model. This can be done in Excel.
The point is that it is very easy to make statements such as your when one does not make an attempt to actually model reality and see what this might mean. When you do (if you do) I assure you there will be an "Oh. That's what he was talking about." moment.
I very much understand that a full migration to EV's isn't likely to be possible for perhaps as long as fifty years, if not more. Yet, that's not the point here. What we are being asked to do to this nation and the world is motivated by the delusion of clean energy and pink unicorns. They are selling this delusion as if it were reality. It is not.
Let's assume that only a third of the 300 million vehicles convert to electric power in, say, 25 years. That means we need to increase our generation capacity by about 400 GW. For context, that means 400 nuclear power plants, each operating at 1 GW. I use nuclear power plants because this is a good measure of the scale of things as well as the impossibility of achieving such a goal in just 25 years. We can't build one in 25 years, much less 400 of them.
Even worse, the power distribution system, the "grid", would have to be rebuilt to at least double, if not triple, its power transport capacity. This due to a few realities, one of them being that you have to build it for present (once the upgraded version goes online) peak demand as well as future requirements.
To be clear, I am not proposing that going electric is a bad idea or that we should not do it. In a very much for the transition. What I am saying is that the delusion we are selling does not describe a path to a reality where even one third of vehicles can be electrified. In order to get there we have to abandon ideology and focus on math and science. Then we can plan a path that will get us there. Whatever "there" means, whether that is 25% electrics or 75%. We can't get there without a stop to all this ideological nonsense.
The perfect example of this is the insanity of our current oil production policies. Here in the US, the current government is doing all it can to kill our oil production and processing infrastructure. They hate it. Ideologically. They just hate it. Irrationally. Because they have all bought into the delusion.
What's reality? Well, if we want, say, 25% of our vehicles to go electric in 25 years, we have a massive job ahead of us. We have to build a massive amount of additional power generation capacity. We have to execute a decidedly non-trivial upgrade of our entire power delivery infrastructure. We have to build somewhere between 50K and 100K charging stations throughout the nation. We have to manufacture batteries, solar panels, cables and all the materials and components that go into such systems. We also have to build nuclear power plants faster than ever in history. And more. I can't possibly list all the dependencies in the tree that leads to being ready for 25% of our vehicles becoming electrics. Not to mention actually manufacturing the vehicles.
What do we need in order to accomplish the above? Cheap oil. What I mean by "cheap" is $20 oil, rather than $130 per barrel. Why? Because EVERYTHING we will have to make, move, install, manufacture requires oil. Everything. You can't build a power plant without a massive amount of fuel. The cables that move the electricity? You can't make them without oil. You can't install them without oil. Manufacturing parts, and structural materials (beams, etc.) out of steel and aluminum? You can't do it without oil. Plastics, electronic components, microprocessors, displays, computers, motors, car bodies, etc. Everything is based on industrial processes that depend on oil. Humanity cannot exist in its current form without oil.
And so, the paradox here is that, if we want a future where ground transportation is significantly less dependent on petroleum, we have to make petroleum as cheap as possible for somewhere around twenty five years, if not more. Only after making that kind of a monumental effort will we be able to start detaching from oil. We can't get there from here because, at $100+ per barrel these projects become insanely expensive and likely unattainable.
We need a real plan of action based on real math and science. And then we do everything we must do to make it happen. We have to leave this delusion behind us in order to move forward. If we don't, electric vehicles will be doomed to being an oddity for decades upon decades.
> Let's assume that only a third of the 300 million vehicles convert to electric power in, say, 25 years. That means we need to increase our generation capacity by about 400 GW.
Can you point me to the actual math and not just a YouTube video of a guy habitually wrong about technology and talk of a spreadsheet model of probably inaccurate usages which really claims such needs? Because in practice I've got an EV in my garage tonight which doesn't increase our overall power demands at all. In fact, it's only finally using the negative generation costs that we've had in my area for years. That's right, right now as my EV charges the wholesale cost of energy in my area is negative USD. There's not enough load to meet the generation demand! My retail provider is making money buying power at negative cost on me charging my car at the moment. Adding more cars charging right now would only bring balance to that mismatch of generation to demand, not overload demand at all.
The model you've mentioned above has a lot of questionable assumptions baked into it. It didn't take a new transformer for my home to support an EV, nor my neighbor, nor the person next door to them. So when thinking of vehicles, and especially not miles or per pound of payload delivered, it's really not that crazy.
Please feel free to share your magical spreadsheet model which explains everything. I'd love to see it. It sounds like it makes a lot of assumptions which may or may not be realistic. But until you share that I'm going by my actual use cases I've actually seen in person with EVs, where it didn't increase my peak power at all, and between myself and my neighbors was absorbed by the grids capacity without any issue at all.
So let's take your example of building out a model based on usage I've actually experienced. How much peak load has my EV actually added to my home's usage? 0 additional kW peak? Cool, so we extrapolate that to all the other single family houses, so we can easily add over 15 million EVs to our fleet before we even need to upgrade a single line. Given we've only managed to produce a few hundred thousand EVs in several years it'll probably take many decades before we even each that 15 million figure, so we're probably fine to slowly roll out upgrades to our grid if we're really only targeting upgrading commuter vehicles at the moment.
> "Let's assume that only a third of the 300 million vehicles convert to electric power in, say, 25 years. That means we need to increase our generation capacity by about 400 GW. For context, that means 400 nuclear power plants, each operating at 1 GW. I use nuclear power plants because this is a good measure of the scale of things as well as the impossibility of achieving such a goal in just 25 years. We can't build one in 25 years, much less 400 of them."
Hinkley C is the first public nuclear power plant in the UK in decades. At current rate it will be 27 years between announcement and completion, at a cost of $30Bn USD equivalent, double the earlier cost and a decade later than planned. Brazil is building a 1GW solar farm [1]for a total cost of $750M USD equivalent. Algeria is aiming to build 15GW of solar in the next 15 years[2]; that's a country with a GDP of $150 Billion USD compared to USA's $21 Trillion.
+400GW of Solar in 50 years at Brazil's prices would be $20Bn/year. Solar has been dropping in price significantly in the past two decades and may do so in coming decades. Compared to +400GW of nuclear power at these prices, which would be $240Bn/year. Hinkley C has doubled in cost in a decade, and it's not certain it would be significantly cheaper if done again several times.
That's still not easily /doable/ for all the other reasons you've discussed about materials needed and power used in supply chains and manufacturing and limits of places to put it, etc., but it's a more plausible to commit a-COVID-response-per-year than an-entire-Apollo-Moon-landing-program-per-year for the next dozen presidential terms. The Elon Musk interview you linked, you say "We need to double power generation, it can't be done! It's impossible! Elon says so!" but Elon says "we need to double power production and distribution, which isn't going to work, that's why we make solar roof tiles, we need more local power generation, blah blah". He didn't seem to agree with you that it was impossible.
USA uses 21 million barrels of oil every day, Google tells me there is about 1.6MWh equivalent in a barrel of oil, or 33million MWh per day equivalent; over 24 hours makes about 1.3TW, approx the same as USA power generation. To replace all the oil use (in pure power terms not counting plastics, etc) with electricity would require doubling the power output of the country, yet a lot of the oil isn't used for transport, maybe half is used for transport. So replacing transport should be a chunk less than doubling power output. And it's not the case that transport needs to stay the same efficiency for the 50 years, e.g. how many of the 300M cars you talk about get less than 60mpg? Less than 40mpg? How many of them stop their engines at traffic stops? Elon commented that wing mirrors on Teslas cost up to 5% drag but can't replace them with cameras for regulatory reasons which could be changed. What if road speed limits were reduced to mandate more efficiency? What if delivery shopping was subsidised? What if work-from-home was given a tax break? What if double-car-ownership was taxed more? 50 years is enough to drive a lot of other changes in society if there was collective will.
I'm not handwavingly saying "it can be done easily", maybe it still can't; but I am saying that you are thrilling over the-sky-is-falling. It's not the case that the only possible variable is power production; power use can be incentivised to change.
The UK national grid is currently producing[3] 30GW, half of that from wind. There's 33M cars in the UK, if they all charged at the other commenter's suggested 750W, and needed charging every 5 days each, that would add 5GW of demand. The UK built 2.4GW of wind power in a single year in 2019. You're modelling 5x the population with 10x the cars and getting to 100x the power demand.
> "What do we need in order to accomplish the above? Cheap oil. What I mean by "cheap" is $20 oil, rather than $130 per barrel. Why? Because EVERYTHING we will have to make, move, install, manufacture requires oil. Everything. You can't build a power plant without a massive amount of fuel. The cables that move the electricity? You can't make them without oil. You can't install them without oil. Manufacturing parts, and structural materials (beams, etc.) out of steel and aluminum? You can't do it without oil. Plastics, electronic components, microprocessors, displays, computers, motors, car bodies, etc. Everything is based on industrial processes that depend on oil. Humanity cannot exist in its current form without oil."
I often think the only way out is through. We cant revert to pre-industrial life without billions dying and quality of life plummetting. Still, oil is finite even if you discount the CO2 burning. Making it 1/6th the cost would lead to increasing use; would you, and how would you, stop it being used for motor racing and plastic junk making, in 5x the quantity?
Since I can't know if readers understand the difference (you might, others may not), here's a simple example:
You have to lift 1000 kg of metal to the top of a massive 1000 meter tall skyscraper being built. It's a thousand 1 kg pieces. Let's say it takes 10 minutes to walk up to the top.
You can do this at least two ways:
One person moves each 1 kg piece up to the top, one at a time. That will require 1000 times 600 seconds, or 600,000 seconds. That's roughly seven days.
The other way is to do it all at once. We get a thousand people, each one of them grabs a 1 kg piece and, ten minutes later the entire load is at the top of the building.
Anyone can understand that these problems require vastly different scenarios.
Energy is about the total amount of work done. In this case it is the same for both scenarios.
Power, on the other hand, is a measure of how quickly this work was done. In this hypothetical, the fast delivery scenario required a thousand times more power than the slower version.
Electric cars require power. Energy is delivered over time. The problem with our grid and power generation is that it simply cannot handle this beyond a very small deployment. People driving electric cars today are enjoying the fact that they do not pose a significant threat to the existing infrastructure and generation capacity. As the numbers increase we will start having problems.
The idea of having a million or ten million cars simultaneously fast-charging today is likely an unthinkable reality.
Put a different way (just pulling hypothetical numbers out of my behind, these are not real numbers):
If we could charge electric cars over, say, 30 days, we might not need even one bit of additional power generation capacity or changes to the infrastructure. Sure, nobody would be able to drive more than once a month, but, hey, we don't need to build nuclear power plants.
However, that is not a practical reality. Tens of millions of cars will need to quick charge in 20, 40, 60 minutes. And another set of millions of cars will want to be charged over, say, 8 hours, overnight. This requires the ability to deliver power --large amounts of it-- in a way we are not prepared to do. This is both true in terms of power generation capacity as well as infrastructure, from the cables all the way to the power plants.
What the page you linked shows is an increase in energy consumption. It says nothing about power. Also, the increase you point out is only 60%, not a doubling from 1980 to 2020. Still, this is somewhat meaningless. I can consume 60% more energy if I keep the lights on 12.8 hours per day, instead of 8. I would not demand 60% more power out of the grid, I would just use the available power for a longer period of time. That's why speaking in terms of energy isn't relevant to the electric car problem. The problem is power, not energy. People need their cars charged now, right now. That requires power we cannot deliver.