The owner drove just over 15K miles in 12 months and bought about 5200 kWh that year.
The reason the average numbers I use are higher than this number is simple: I can guarantee you that vehicles like the electric Ford F150 will require far more energy, probably twice as much. We already know that the Mustang Mach E requires more energy per mile than a Model 3.
When we say 300 million vehicles, we are not talking about 300 million Tesla's. There will be a wide range of vehicles, from large trucks to minivans, cargo vans, large and small SUV's, performance cars, commuter cars, etc. And then we have to include semi trucks and large commercial trucks. The average energy-per-mile figure isn't going to be that of a Model 3.
However, for the purpose of a discussion, I'll go with the number from the article for a Model 3. In fact, because I want simple numbers, I'll round it down from 5,200 kWh to 5,000.
So, the for the entire year we need to pump back 5,000 kWh into the car.
How long is a year?
Hmmm. People don't drive the same distance every day. We have to be VERY careful not to average ourselves into an artificially low power requirement number. Please note I am using upper case for highlight because this doesn't really do the job. Not yelling at you.
This point deserves highlighting: This issue is about POWER generation, not energy.
For those who might not be comfortable with the concepts:
A 1,000 W light requires 1 kW of POWER every instant it is on. If it happens to be on for one hour, it will have used 1 kWh of energy. Ten hours, 10 kWh.
If I have a thousand 1 kW lights I need 1 GW of POWER every instant the lights are on. It doesn't matter if they are on five seconds or three days. I need a gigawatt. Energy is a function of how long the lights are on. A kilowatt-hour means you used 1 kilowatt of power for one hour, not five minutes.
OK, back to cars.
This is why "How long is a year?" matters. There are three possible answers to this:
200 days, 365 days or some complex formula that accounts for average weekend driving.
There are roughly 200 working days in an year, which is when most of the driving happens. Put a different way, this is when people will most likely drain their batteries the most.
The difference between 200 and 365 is massive. It's almost double. This is significant because the next question we have to answer is:
When do people charge and how much?
The Model 3 owner from the linked article didn't have one massive 5,000 kWh charge on January 1st. to then drive the entire year. That would require a battery the size of a large building.
For the purpose of modeling we have to either choose to develop a complex model, one where we divide the population into behavioral groups and assign a wide range of utilization and charge scenarios to each group. That's a lot of work when all you are trying to get is a sense of proportion rather than an accurate answer.
I think it's sensible to pick 200 days as a starting point. That assumes all of the driving is done during the week and ignores the weekends. We can look at the 365 day case as well and compare notes. This is why throwing this into a spreadsheet is useful.
I spent some time discussing power because this is what we are after. If I now take the annual 5,000 kWh baseline from the article and divide it into 200 days, we get 25 kWh per day.
That's the energy you need to pump into the car every one of those 200 days. However, you are not going to be plugged in for 24 hours. The more likely scenario is that you are going to plug in after work while you sleep. The assumption I have made is that the average car will be charged in eight hours.
So, that means you have to deliver 25 kWh in 8 hours. Which means you need 3.125 kW of POWER. Here's where the difference becomes important, if you could charge 24/7 you would only need a little over 1 kW of power. Because we can't do that, our power delivery system would have to provide us with THREE TIMES the power when compared to averaging over 24 hours.
What if we charge every day for 8 hours each day? 5000/365 = 13.7 kWh per day. That means 1.7 kW of power for eight hours.
Now we have two scenarios we can compare:
200 day year -> 3.125 kW power for 8 hours
365 day year -> 1.7 kW power for 8 hours
How many vehicles would be charging simultaneously?
My simple model states that the minimum is around 188 million and the maximum would be 300 million. This is a model where I divide the fleet into time-zone groups, each of which starts charging the entire fleet for that time one hour after the prior time zone. A simple model, yes. I am just trying to get a sense of proportion here. A more accurate answer would require regional as well as behavioral modelling. For example, places like New York and Los Angeles are going to behave differently from Las Vegas, NV or Wakefield, MA.
What matters is the peak power requirement, not the minimum --unless we are willing to ration electricity.
Two scenarios then, 200 day and 365 day year and 300 million car peak utilization.
200 day year -> 3.125 kW (power) x 300M cars = 938 MW
365 day year -> 1.7 kW (power) x 300M cars = 510 MW
In other words, at a minimum, given this model, we would need an additional 500+ megawatts over and above current utilization, with the peak being over 900 megawatts.
Now it is necessary to put this into context. A typical nuclear power plant produces 1 MW of POWER. In other words, we are talking about needing somewhere in the range of 500 to 900 new nuclear power plants.
We can't say something like "air conditioning alone uses x megawatts" because we don't get to use that power for cars. That power is an existing requirement. In order to have 300 million electric vehicles we need to ADD power generation AND transmission capacity throughout the country.
That's the other part of the story that is often waved over. Moving an extra 900 megawatts of power isn't something one can assume the current infrastructure can handle at all. Here in CA we are already struggling with blackouts and forest fires caused by a range of issues, including the aging power grid. We should not hand-wave our way around the realities of what we are facing if we want this electric vehicle future.
So, 500 to 900 nuclear power plants. That means, just guessing, somewhere between 5 to 20 per state (some states will need more than others). Well, we can't build ONE nuclear power plant in, say, ten years. If we want to go full electric in 30 years we would have to build 15 to 30 nuclear power plants per year, every year, for the next thirty years.
I'll repeat what I said in my prior post: Even if I am off by a factor of 2 to 10, the problem is of massive proportions. At the low end it means we need 50 nuclear power plants (off by a factor of ten and using the low estimate). At the high end we need 900 of them, if not more.
All I see out there is hand-wavy, blue-sky, just install solar panels wishful thinking. Not a single honest mathematical model in sight (that I know of).
The important take away is that, while range is about energy, the reality of charging is that it is about power. It is far too easy to fabricate numbers that are artificially low by making assumptions like 24/7/365 charging and that nobody ever wants or needs to pump the aforementioned daily requirement of 25 kWh into their vehicle in just one hour, something that requires EIGHT TO TEN TIMES MORE POWER (due to losses) than when charging in eight hours.
> A typical nuclear power plant produces 1 MW of POWER.
I think you've got things confused by about an order of magnitude here – 1 GW per nuclear power plant sounds more like it.
(Also for comparison's sake – an electric train can use up to a few MW when accelerating under full power, and you certainly don't need multiple nuclear power stations to power just one measly train)
That was a typo. If you look at the amount of work and research I have done on this subject I think it is pretty easy to determine "confused" is far from where I am. Yes, it is 1 GW. And my math uses this number, not 1 MW.
I invite you to run through your own calculations. I actually WANT to be wrong. I just don't see what I am missing. Again, this is about developing a ROM (Rough Order of Magnitude) model. The difference between 50, 100 and 300 nuclear plants is almost irrelevant. Why? Because we can't even build a single nuclear plant in 10 to 25 years, which means that a ROM requirement of ten, twenty or a hundred nuclear power plants might as well be a million.
In the US, we are at a point in history where we can't build anything of any real scale. The best example I have of this is the failed high speed train in California. A project sold to voters as a ten billion dollar price tag. It is now at a hundred billion, only about ten miles have been built. These ten miles are unusable (not in service as far as I know) and are far from being high speed by any definition of the term. Some think this thing will be a trillion dollar disaster, if it is ever completed.
In this context, we actually think we can add hundreds of gigawatts to our power generation system? The only way to do is is through nuclear power. Which means it is a fantasy. Unless our culture, philosophy and politics changes radically we just can't do it.
Here's another ROM calculation. Let's keep to California. We have just over 31 million cars and trucks [0].
Since this is a ROM calculation, I'll start with the assumption that everyone gets home and plugs into a Type 2 charger.
Type 2 chargers typically deliver 3 to 5 kW of power.
How much power will we require at 6:00 PM PST when everyone gets home and plugs in?
3 kW x 31 million = 93 GW
5 kW x 31 million = 155 GW
It does not matter if people charge for one hour or eight, if they are all pulling 3 to 5 kW from the grid, you have to have the ability to supply this kind of power instantly or power outages and other ugly things will be the consequence.
OK, this was a ROM calculation. Right? What if only 10% of these vehicles plug in every day at 6 PM. What then?
3 kW x 3.1 million = 9.3 GW
5 kW x 3.1 million = 15.5 GW
How much power generation capacity do we have in CA? [1]
Refer to the table titled "Installed In-State Electric Generation Capacity by Fuel Type (MW)"
In 2020, it's about 80 GW. Also, note that half of it comes from burning natural gas and only just over 2 GW from nuclear.
Here's where we have to understand that we don't build infrastructure to be able to supply two times the power we need. The cost of doing so would be staggering. In other words, we don't have another 2 GW of nuclear power sitting around waiting to be used. Hence the blackouts and other issues we have throughout the state.
My guess is that we are likely at 80% peak utilization. For many years now we have been asked to limit use of air conditioning and power in general or risk blackouts. In fact, rolling blackouts are kind of a normal thing in CA these days.
Is it realistic to assume to only 10% of all electric cars and trucks in CA will plug in on any given day? Likely not. Put a different way, the longer these vehicles wait to recharge the worse the power deliver problem becomes. If they plug in every day they might only need power for a couple of hours. If everyone waits until the weekend to plug in, they might need to sit on that charger for eight to ten hours and the stress to the power grid would be compounded.
I like to use nuclear power plants as my unit of measure because they are about 1 GW. The ROM calculation above says we need from ten to 155 new nuclear power plants to be built in CA in order to support simultaneous class 2 charging by some portion of a vehicle fleet where every single vehicle has been switched to electric power, no more gasoline or diesel at all.
I think the low end of this ROM calculation isn't reasonable. The same is the case for the high end. The answer likely lives somewhere in the middle of this range. One thing is certain, we need to add a very serious amount of generation capacity, likely in the many tens of gigawatts.
We might have to DOUBLE our current power generation capacity. Double it.
How do we do that?
Well, being that half of it comes from natural gas, maybe we build more plants and burn more of it. How is that for being "green"?
Solar?
I don't think so. Most of the solar capacity in CA is installed on homes. In general terms, these rooftop systems are sized to cover the energy needs of the home. Most of my neighbors have systems that are barely adequate enough to cover their needs, which means they have nearly zero excess capacity. So the rooftop-solar-powered electric car charger is mostly a fantasy for most. Most of the systems in my neighborhood are around 6 kW. This does not mean they actually deliver this kind of power, not even at the peak.
I designed and installed my own system, which consists of 40 panels, for a theoretical total of 13 kW. At the absolute peak of the season I might see 10 kW. Yesterday the peak was just over 8 kW [2]. This is due to a combination of the time of year, clouds, shade and dirt on the panels. As you can see, the curve has a nice 45 degree-ish slope both going up and down. By 6 PM (coming home time) I am at about 2 kW. Most of my neighbors would be lucky to generate 1 kW at that time.
As for energy generation, this is July [3], a peak of about 57 kWh. June [4] was a little bit better, with a peak at 65 kWh. May [5] had a peak at 69 kWh. April [6] was the best month, with a peak at 72 kWh.
You might note that every single month had several days of really low energy output. This is usually due to weather, clouds or such things as fires reducing the photons that can reach the panels. For example, while April provided a nice 72 kWh peak, it also had a day where the best we did was 24 kWh. May was the best month so far this year, with 1.9 MWh total energy generation [7].
This is all to say that the reality of solar is very different from the fantasy of solar. Most people who do not have solar think of it as some magical energy source that gives and gives and gives. Not so. And, when it comes to electric vehicles, the problem becomes that you don't have it when you need it. Sure, there are all manner of accounting credits that come into play. Explaining the mess that is TOU billing in the US would take-up another post, if not five.
What do you do when you expect to generate 72 kWh and you only make 24? The power grid has to be ready to supply your needs. No problem, solar energy sharing from neighbors will take care of it locally, right? Not so. Most homes have small systems. In my neighborhood there are probably only two or three homes generating at my level. Which means we are the only ones with real excess capacity. When the clouds come in, nobody has power to share, not even those of us who produce twice as much as the rest.
Without storage solar isn't very useful for electric car charging. And storage at the home is a foolish investment from a ROI perspective. Not there yet. Believe me, I want to turn my home into a massive solar-powered UPS. I studied this when I engineered my system and it is ready for it. It just does not make sense at all at the moment. It would be far more logical to add another ten panels than to buy batteries.
Anyhow, not to go on a tangent here. The point is that the step change in power requirement to support millions of electric cars isn't going to come from solar at residential rooftops. And, even if we installed massive grid-scale solar, the curves and generation issues you see from my system will still create issues. The only way they might be able to mitigate this would be through energy storage and, at this time, this is a high cost fantasy. I have high hopes that iron-based battery technology --which stands to be 10x cheaper-- might become a reality we can all benefit from.
The other issue with solar is that neighborhoods are turning against these massive installations anywhere near their town. A few weeks ago I read a story about a massive installation that is being taken down (at great loss to all investors) because the neighborhood sued with a claim of taking a hit to home values due to the unsightly visual of thousands of panels on a hill. As is always the case, reality is far more complex than most think and it can't be reduced to a single variable.
Let's start with a useful number. How much energy does something like a Tesla Model 3 require, on average, per year?
I am using this article for the data:
https://cleantechnica.com/2019/08/12/tesla-model-3-owner-dri...
The owner drove just over 15K miles in 12 months and bought about 5200 kWh that year.
The reason the average numbers I use are higher than this number is simple: I can guarantee you that vehicles like the electric Ford F150 will require far more energy, probably twice as much. We already know that the Mustang Mach E requires more energy per mile than a Model 3.
When we say 300 million vehicles, we are not talking about 300 million Tesla's. There will be a wide range of vehicles, from large trucks to minivans, cargo vans, large and small SUV's, performance cars, commuter cars, etc. And then we have to include semi trucks and large commercial trucks. The average energy-per-mile figure isn't going to be that of a Model 3.
However, for the purpose of a discussion, I'll go with the number from the article for a Model 3. In fact, because I want simple numbers, I'll round it down from 5,200 kWh to 5,000.
So, the for the entire year we need to pump back 5,000 kWh into the car.
How long is a year?
Hmmm. People don't drive the same distance every day. We have to be VERY careful not to average ourselves into an artificially low power requirement number. Please note I am using upper case for highlight because this doesn't really do the job. Not yelling at you.
This point deserves highlighting: This issue is about POWER generation, not energy.
For those who might not be comfortable with the concepts:
A 1,000 W light requires 1 kW of POWER every instant it is on. If it happens to be on for one hour, it will have used 1 kWh of energy. Ten hours, 10 kWh.
If I have a thousand 1 kW lights I need 1 GW of POWER every instant the lights are on. It doesn't matter if they are on five seconds or three days. I need a gigawatt. Energy is a function of how long the lights are on. A kilowatt-hour means you used 1 kilowatt of power for one hour, not five minutes.
OK, back to cars.
This is why "How long is a year?" matters. There are three possible answers to this:
200 days, 365 days or some complex formula that accounts for average weekend driving.
There are roughly 200 working days in an year, which is when most of the driving happens. Put a different way, this is when people will most likely drain their batteries the most.
The difference between 200 and 365 is massive. It's almost double. This is significant because the next question we have to answer is:
When do people charge and how much?
The Model 3 owner from the linked article didn't have one massive 5,000 kWh charge on January 1st. to then drive the entire year. That would require a battery the size of a large building.
For the purpose of modeling we have to either choose to develop a complex model, one where we divide the population into behavioral groups and assign a wide range of utilization and charge scenarios to each group. That's a lot of work when all you are trying to get is a sense of proportion rather than an accurate answer.
I think it's sensible to pick 200 days as a starting point. That assumes all of the driving is done during the week and ignores the weekends. We can look at the 365 day case as well and compare notes. This is why throwing this into a spreadsheet is useful.
I spent some time discussing power because this is what we are after. If I now take the annual 5,000 kWh baseline from the article and divide it into 200 days, we get 25 kWh per day.
That's the energy you need to pump into the car every one of those 200 days. However, you are not going to be plugged in for 24 hours. The more likely scenario is that you are going to plug in after work while you sleep. The assumption I have made is that the average car will be charged in eight hours.
So, that means you have to deliver 25 kWh in 8 hours. Which means you need 3.125 kW of POWER. Here's where the difference becomes important, if you could charge 24/7 you would only need a little over 1 kW of power. Because we can't do that, our power delivery system would have to provide us with THREE TIMES the power when compared to averaging over 24 hours.
What if we charge every day for 8 hours each day? 5000/365 = 13.7 kWh per day. That means 1.7 kW of power for eight hours.
Now we have two scenarios we can compare:
How many vehicles would be charging simultaneously?My simple model states that the minimum is around 188 million and the maximum would be 300 million. This is a model where I divide the fleet into time-zone groups, each of which starts charging the entire fleet for that time one hour after the prior time zone. A simple model, yes. I am just trying to get a sense of proportion here. A more accurate answer would require regional as well as behavioral modelling. For example, places like New York and Los Angeles are going to behave differently from Las Vegas, NV or Wakefield, MA.
What matters is the peak power requirement, not the minimum --unless we are willing to ration electricity.
Two scenarios then, 200 day and 365 day year and 300 million car peak utilization.
In other words, at a minimum, given this model, we would need an additional 500+ megawatts over and above current utilization, with the peak being over 900 megawatts.Now it is necessary to put this into context. A typical nuclear power plant produces 1 MW of POWER. In other words, we are talking about needing somewhere in the range of 500 to 900 new nuclear power plants.
We can't say something like "air conditioning alone uses x megawatts" because we don't get to use that power for cars. That power is an existing requirement. In order to have 300 million electric vehicles we need to ADD power generation AND transmission capacity throughout the country.
That's the other part of the story that is often waved over. Moving an extra 900 megawatts of power isn't something one can assume the current infrastructure can handle at all. Here in CA we are already struggling with blackouts and forest fires caused by a range of issues, including the aging power grid. We should not hand-wave our way around the realities of what we are facing if we want this electric vehicle future.
So, 500 to 900 nuclear power plants. That means, just guessing, somewhere between 5 to 20 per state (some states will need more than others). Well, we can't build ONE nuclear power plant in, say, ten years. If we want to go full electric in 30 years we would have to build 15 to 30 nuclear power plants per year, every year, for the next thirty years.
I'll repeat what I said in my prior post: Even if I am off by a factor of 2 to 10, the problem is of massive proportions. At the low end it means we need 50 nuclear power plants (off by a factor of ten and using the low estimate). At the high end we need 900 of them, if not more.
All I see out there is hand-wavy, blue-sky, just install solar panels wishful thinking. Not a single honest mathematical model in sight (that I know of).
The important take away is that, while range is about energy, the reality of charging is that it is about power. It is far too easy to fabricate numbers that are artificially low by making assumptions like 24/7/365 charging and that nobody ever wants or needs to pump the aforementioned daily requirement of 25 kWh into their vehicle in just one hour, something that requires EIGHT TO TEN TIMES MORE POWER (due to losses) than when charging in eight hours.