[colin_mccabe]

> > Here in CA we've already have the government ask electric car owners to alter charging behavior due to power grid problems.

> This bit of propaganda has been surprisingly successful, especially since it doesn't really make any sense.

What doesn't make sense to you? California did ask electric car owners to charge off peak hours to avoid overloading the grid. See https://www.newsweek.com/california-facing-power-crisis-fret...

It was not a legally binding request, so electric car owners could ignore it if they wanted to.

[peter422]

Very few people charge their cars at the most expensive time of day anyways.

For about a 3-day period during a historical heat wave throughout California we all did have to reduce our electricity usage for about 3 hours each day, which was mainly AC but also charging electric cars.

To conclude based on this that the grid is broken or that electrical cars (which mostly charge at night) are going to result in the grid deteriorating further makes no sense through.

[robomartin]

> To conclude based on this that the grid is broken or that electrical cars (which mostly charge at night) are going to result in the grid deteriorating further makes no sense through.

No, that is not what leads to the conclusion. The conclusion is based on two things: Physics and mathematics.

What is happening now is merely a preview of things to come if we don't have the right conversations or people, as you are doing, dismiss the warnings some of us are issuing without making any real effort to understand.

About five years ago I designed and built (as in, I did it myself) a 13 kW solar array at home. Far more than we needed to supply the house. The plan was to use some of that for electric vehicles once they became viable. Note I didn't just say "affordable". The term "viable" is meant to include the entire ecosystem. As a comparison, a gasoline-powered vehicle is viable because you can easily refuel it without even thinking about it and it can be maintained and repaired anywhere and almost by anyone.

Anyhow. This led to me devoting a lot of time for about a year to try to understand energy, climate change and electric transportation realities. What I mean by that is that I invested time and effort seeing just how well the math and physics of what we were (and are) being told, actually align.

What I discovered was a surprise to me: They do not.

I wrote some code to simulate power requirements for a varying scale of EV adoption, all the way up to 300 million vehicles --our current fleet. The simulation predicted a need of between 900 GW and 1400 GW in addition to existing capacity. The current US capacity is 1200 GW. In other words, we need to double our power generation capacity and double (or more) our ability to transport power. As it turns out, this prediction was reasonably accurate.

One of the often hand-wavy things people talk about or write in articles is energy, rather than power. This is a huge mistake. Energy is power delivered over time. One can make outlandish claims about energy while ignoring the time element.

When, in a state like California, you have 31 million [0] EV's plug in to charge at, say, 6:00 PM every night, what you need is power, instantaneously, not energy. The fact that you generated <pick a number> of energy in the prior n days means nothing in that moment unless the energy was stored for delivery as power to each car in that instant.

What I discovered is that, at the end of the day, the hand-wavy stories just don't hold up. As a hypothetical, if you consume ten days worth of stored energy in one to nine nights, you are still short. The truth turns out to be that the EV problem, ultimately, is about power, not energy.

One way I think of this is that all 13 million+ households in CA [1] suddenly get TWO 5-ton air conditioning systems that are turned on every night at 6 PM for several hours. That's what we are talking about. And, no, we don't have the power and, if we had it, we could not deliver it.

So, yes, very much so: The grid is broken (in that it just can't cope with these loads) and a large installed base of electric cars will cause severe grid deterioration in multiple ways.

We can stick our heads in the sand an pretend this isn't so today because EV owners live in a privileged environment where they can take as much power as they need from the system and people, for the most part, don't notice any issues. I am going to guess that if we double the installed base of EV's in CA --which is mostly concentrated in large urban areas-- people will start to notice and this will lead to very interesting outcomes. I could get ugly for EV owners in so many ways.

I don't know how else to say it. I have written a lot about this. People prefer to be dismissive and continue to exist in ignorance of our future reality. We can't even build a high speed train and now we are talking about a transition to EV's that will require a doubling our our power generation and delivery capacity (this is absolutely indisputable). Why aren't we talking about mass adoption of nuclear power? It's because the easy political gains are not there, that's why.

[0] https://www.fhwa.dot.gov/policyinformation/statistics/2010/m...

[1] https://www.census.gov/quickfacts/fact/table/CA/RHI725221

[nostrademons]

Why do you assume that every EV is going to charge immediately at 6 PM? Any reasonable software controlling those chargers is going to a.) at least delay that till 9 PM so that you aren't at peak rates and b.) stagger charging so that not everyone's EV charges at once.

[robomartin]

Please do the math?

Please think it through?

I'll try to create a very simple hypothetical case to illustrate the problem and how to think about it:

Assume the average vehicle is driven 50 miles per day.

This includes commercial vehicles, long and mid haul semi trucks, work trucks, delivery vehicles, etc. In other words, the average daily per person numbers do not apply here (that would be around 25 to 45 miles per day, depending on location). I feel an average of 50 miles per day, regardless of vehicle class, is a reasonable number to use as a thinking tool to try to get a ROM (Rough Order of Magnitude) of the problem. The models I developed years ago were far more accurate than this, however, that kind of detail in a simulation is hard to convey in a post like this.

Assume, then, this to represent an average for all 30 million vehicles in CA.

The question:

How much POWER would this require?

Let's assume we use a Level 2 charger that would replenish 60 miles in an hour at 7 kW. Again, we are super-simplifying things here. For example, a semi truck or delivery van will be far less efficient and require charging at a much higher power level and longer charge duration. I am just trying to simplify this for the purpose of illustrating the problem.

Assuming 30 million vehicles charged simultaneously, this means we would need 210,000,000 kW

Let's have them charge with a uniform distribution across 24 hours. That means we need 8,750,000 kW

That's 8.75 GW.

A typical nuclear power plant produces 1 GW. In other words, in this evenly distributed scenario we would need the output of 9 nuclear power plants for 24 hours to charge all vehicles in CA.

We need 9 NEW nuclear power plants in CA. I would round that to ten.

This is power over and above current generation and transport capabilities.

How long does it take to build just one nuclear power plant? Well, certainly longer than a high speed train. I think the range is between 25 years and impossible.

How about 10 of them? Never. Unless we stop talking about EV fantasy and start discussing reality. And that is: If we want EV's to take over we need to get serious about being able to massively expand power generation and delivery and we need to do that immediately.

No, it cannot happen by 2030. That's preposterous.

And, no, solar isn't going to do it. That's wishful thinking. A solar installation that can match a 1 GW nuclear power plant and deliver 1 GW 24/7 has to be built with a peak capacity of at least 10 GW. This is massive and more than most people can imagine in terms of land use, materials, batteries, etc.

And, BTW, the above super-simplified hypothetical isn't even close to just how bad things will be in reality. For example, if you assume that, say, 25% of vehicles will need fast or high power charging, the power demand will skyrocket. Remember that I said the problem is power, not energy. Power is what you need when you have to charge a bunch of cars simultaneously. That's because you have to do it given the time constraints of the task. You don't have 48 hours to charge a semi truck that just completed a thousand mile journey. At best your might have eight hours. And that requires power. A typical truck stop might have fifty to one hundred long-haul trucks in need of charging. What they demand is power in order to deliver the requisite energy in a given amount of time. The other thing it does not take into account is concentration. A city like Los Angeles will require a staggering amount of additional electricity to deal with EV's and it will have massive peaks that will dictate the size and shape of the required feeds.

Again, we can go head-in-the-sand or understand we have a very serious that requires at least a doubling of our power generation and distribution capacity. If we don't wake up to that right away it will be an absolute mess.

I could get into your comment about delaying charge and staggering. I have including that sort of thing in my models. It does not change peak power demand. Here's the simplest explanation: Imagine you slow charge 30 million cars for 12 hours and stagger 1/12th of them every hour as you proposed. Well, 12 hours into this charge methodology you have 30 million cars charging simultaneously. And, because cars are used every day, you pretty much end-up with 30 million cars charging 24/7. I am over-simplifying. The point is that the stagger idea seems to be an intuitive solutions (I thought so before I modeled it), yet it does not eliminate the fact that you have to deliver so many kWh (now talking energy) to so many cars within a narrow window of time. In real use very few will adopt EV's if they have to spend 24 hours charging.

[nostrademons]

So build 9 new nuclear power plants? In 1972 alone, 13 new nuclear plants were ordered. [1]

Be careful with mixing physics/mathematical arguments and economic ones. If you want to talk physics, assume your (fairly generous) numbers of 8.75 GW. That's 9 nuclear power plants, as you mentioned. Or for solar, mean solar flux in CA is about 5 kWh/m^2 over a day, solar panels are about 20% efficient, that's 1 kWH/m^2/day = 24 m^2 / kW of panels = 24 km^2 / GW * 8.75 GW = 210 km^2 = an approximately 21 x 10 km solar array in the Mojave desert. That's well within the range of plausible land use. For wind, a typical offshore wind turbine generates about 8 MW of power, so we'd need about 1000 of them, turbine blades are about 750 feet across, figure 1/4 mile spacing, we'd cover 250 miles ~= less than half of California's coastline.

The reason these haven't been built yet is because of economics: it's not cost effective to invest this much when the demand isn't there yet. But then we're not going to get 31M car owners suddenly switching over to EVs. We'll get maybe 2-3M each year, switching over as they retire their old vehicles, and then we build one nuclear power plant, or 2 km^2 of solar, or 100 wind turbines, each year until the transition is complete.

[1] https://en.wikipedia.org/wiki/List_of_cancelled_nuclear_reac...

[robomartin]

> In 1972 alone, 13 new nuclear plants were ordered.

You are making my point. We can't build them.

> Or for solar, mean solar flux in CA is about 5 kWh/m^2 over a day, solar panels are about 20% efficient, that's 1 kWH/m^2/day = 24 m^2 / kW of panels = 24 km^2 / GW * 8.75 GW = 210 km^2 = an approximately 21 x 10 km solar array in the Mojave desert.

I am so incredibly tired of this argument. The only people who reach for this are those who know nothing about the reality of solar. They think in terms of the fantasy they've been sold and, therefore, know nothing about what happens in real life.

> and then we build one nuclear power plant, or 2 km^2 of solar, or 100 wind turbines, each year until the transition is complete.

Please. I beg you. If you have a Excel or something equivalent and have at least a high school understanding of Physics and mathematics, slow down, do some research and try to understand. You really do not. You are confusing a google search for reality.

I'll provide with a quick fantasy vs. reality education as a starting point. The rest is up to you. You can either continue to believe in fantasies or start to understand.

Here's a graph showing the power output of my 13 kW array about a month ago:

https://i.imgur.com/aNnbmDp.png

Notice the parabolic shape with a peak at about 8 kW.

Wait, what? Not 13 kW?

Right. Output changes through the year. I have yet to see it reach the full rated panel output. The most I've seen is around 10 kW. Do you know why? Because the fantasy you quote in terms of efficiency (and everything else) is a rating had under ideal laboratory conditions, starting with an operating temperature of 25 degrees C. This is great for marketing and laughable for real-life conditions.

It doesn't end there. Check this out:

https://i.imgur.com/pB1WgQ0.png

This was the very next day!

What happened? How did the array go from 8 kW all the way down to 2 kW, then back up to about 7, down again, up again, etc.? How did that happen?

Clouds!!! That's how that happened. F-ing solar idealists make me sick. I was one of them, BTW, until I built this system and learned that my fantasy did not match reality at all.

Clouds!!!

Do you think that's it? Check this one out. One day later:

https://i.imgur.com/FiaENVI.png

Clouds. Again! Are you starting to understand? Does this start to paint an image of why all these hand-wavy solar flux arguments are complete and utter nonsense?

Do you know when peak solar production occurs? Which month of the year? Most people will say June/July.

Nope, it's April/May. Here's a full year:

https://i.imgur.com/EQc8EDD.png

Because solar panels have a negative temperature coefficient. That's why! Which means their output is reduced as the panel temperature increases. In June/July it's just too hot. April/May happen to be the right balance between solar input, temperature and other factors.

Remember the graphs showing power generation loss due to clouds? What does that look like through the month. Well, here's what my output looked like this last April:

https://i.imgur.com/8lYKImD.png

See that? On any given day your power output can be reduced by anywhere between 25% and 50%. And that's in a good month. Look at what happened in January:

https://i.imgur.com/bGuCH2F.png

80% reduction in power output! 80%!

For goodness sake, abandon this fantasy and take the time to learn about reality. What's even more frustrating is that people like you will actually engage in intense arguments armed with nothing more than fantasies. Please.

I have no problem with someone not knowing something. We all have tons to learn. I certainly did not have the level of understanding I have today until I built my own solar array and started to try to understand why my output did not match my expectations. What a lesson that was.

What rubs me the wrong way is when people pretend to know something. I have never acted in that manner in my life. If I don't understand something to a good degree I keep my mouth shut and try to learn from those who actually do. That does not mean I don't make mistakes, but I try really hard not to say anything I don't know or have researched to a reasonable depth.

Let's talk about the consequences of the above graphs and your "and then a miracle occurs" calculations (because that's what they are when compared to reality).

The parabolic power output curve means you have to build a solar array 1.5 times larger in order to deliver the same energy over a roughly 12 hour period as that of a constant-power system (nuclear) producing your peak power.

Why?

Because energy is the integral of the power curve over time. The integral of an inverted parabola is 2/3 the area of the enclosing rectangle (the constant power curve). Therefore, my solar array produces 2/3 the energy of a source that can deliver constant power in the 8 kW to 10 kW range. In order to deliver that energy I have to grow my system by the reciprocal of that, which is 1.5. And, of course, I would have to add batteries if what I am after is power. In other words, I have to fill the areas outside the parabola with power I have stored in batteries.

Wait. There's more. This only covers, say, 12 hours of the day. Now I need to overbuild the system yet again in order to provide power at night. That means, at a minimum, a 2x multiplier. I am now up to 3x (1.5 * 2). In other words, my humble 13 kW system would have to triple in size to 39 kW.

Are we done?

No.

Why?

Remember the damn clouds? Here's a graph from March of this year:

https://i.imgur.com/yvTdNX0.png

Horrible stuff. You have to account for this. Believing that one is going to have perfectly shape parabolic output 365 days per year is part of that fantasy I have been referring to. That's not reality.

How do we account for that? If there are no nuclear power plants and no coal (whatever) power plants and we depend 100% on solar (please don't say "wind"), well, there are days when you could fully lose 80% of your output. Heck, you could lose 80% of your output for days or weeks due to weather or fires.

How to size a system to mitigate such events if an entire city and all of the EV transportation in that city depends on locally generated solar energy to exist?

This is where statistics comes in. If we had to mitigate that day when output was 1/5, we would have to build a solar power plant 5 times larger yet. That's not sensible. Reality probably lies somewhere around 50% to 100%, this being a guess and something that is very highly dependent on geography, weather and statistical probability of fires and other events. Build it in the desert? How much output do you lose to sand storms and sand on the array? Build it where there's lot of rain? You might need to overbuild by 10x to get constant power at the required level.

If I assume a 50% overbuild, we go from 3x to 6x.

So now, in order to be able to deliver 1 GW of power 24/7, you need to build a 6 GW photovoltaic solar array with a massive amount of storage.

The real number, when other factors are taken into account, is likely to be closer to 10x overbuild or more. What factors? Failures, maintenance, fill ratio, etc.

Connecting it to my prior post, if you need to add a true 10 GW of power generation capacity that is available 24/7 to support EV's you probably have to build at least 100 GW worth of photovoltaic solar generation and so many batteries I hesitate to count them.

Instead of this fantasy, we need to get our heads out of our collective behinds and build nuclear power plants. That's the only way. Solar alone can't do it, it can (and should be) be a supplemental add-on.

There's more to the story, of course.

[onos]

Thanks for this detailed comment and that above. Have you got links to any blog posts or code that we could look at to review in detail?

[robomartin]

I don't, but I am so sick and tired of the nonsense that I have been thinking of creating a site with more detailed data on this, methodologies, calculations and perhaps a better version of the simulation that anyone can try, reproduce and challenge (it would not be science without it being open to honest review).

BTW, I built my model many years ago, way before anyone was talking about this. While it seemed to be a reasonable rough-order-of-magnitude model, it wasn't until the end of last year that I finally got confirmation that my model produced reasonable numbers. This by non other than Elon Musk himself:

https://www.youtube.com/watch?v=TcI6FaaDp8g&t=3510s

In this interview he says we at least double our power production capacity. We (US) have 1200 GW of installed capacity. That means we need at least another 1200 GW. Which means 1200 nuclear power plants.

That number, for me, gives the problem a dimension, a scale, that is difficult to understand in any other way. 1200 nuclear power plants is nothing less than daunting. In the context of not even being able to build a high speed train, I am not sure what the reality of nuclear power in this nation might look like in the next 50 to 100 years.

Of course, Elon is pushing solar. Great for some areas, not so for a good deal of the nation. Imagine, for example, if Florida depended 100% on solar. Yes, that's an extreme example, of course. Sometimes these are necessary to jolt people away from thinking about the fantasy of something and focus on reality. Solar in Maine or Illinois has very different prospects when compared to solar in Southern California, Arizona or Texas.

Solar isn't the solution. It is part of it, of course.

[colin_mccabe]

In theory, electric vehicles could help stabilize the California grid, if we could charge them at times of peak solar production. There are definitely logistical challenges with doing this, though.

[robomartin]

> In theory, electric vehicles could help stabilize the California grid, if we could charge them at times of peak solar production

Can you clarify what you mean by this?

I read it as "EV's are storage that can feed energy back to the grid". The problem with this is that it assumes you charge your car and use it as a battery, rather than drive it. I don't think that's realistic at all. In addition to that, they still have to be charged. Which means we need additional power, over and above current demands, in order to do so. The energy has to come from somewhere, and we don't currently have it.

Also, the idea of charging at peak solar is a fallacy. This is what solar production looks like during an ideal and non-ideal day (source: My own 13 kW array):

https://i.imgur.com/aNnbmDp.png

https://i.imgur.com/pB1WgQ0.png

The peak lasts minutes, if not seconds. Peak solar, in this context, is pretty much useless. What you need is steady power delivery over a period of many hours (for slow to mid charge rates).

What a lot of people tend to ignore is that the current grid and power generation capacity is pretty much built to supply current needs. A large EV installed base expansion requires an equally large expansion of power systems at all levels. Solar isn't the solution. It's part of it, of course, just not the solution. The same is the case for wind. We need nuclear. Lots of it.