Renewables can’t replace natural gas until we develop effective ways to store the energy for a baseload.
Since we won’t use nuclear power, it’s going to be here for quite some time.
https://www.cnbc.com/2021/06/11/bill-gates-bullish-on-using-...
Since we won’t use nuclear power, it’s going to be here for quite some time.
https://www.cnbc.com/2021/06/11/bill-gates-bullish-on-using-...
(Obviously you don’t want to do antipodal HVDC until you need to, because even if the HVDC lines were free, pointless transmission losses are still pointless even if you can afford them).
Well there's the crux, ain't it? They're bloody expensive.
Also, to your 3.5% loss / 1000km line loss add in the transformation losses. Then consider that a 20k line must be a DC line (to avoid reflection losses).
Also, why did you use 20k km? Such a curious number, half the world's circumference, that I suspect you mean to propose to wrap the world in power lines to get power at midnight from noon elsewhere
Well 20km for your loss calculations is not nearly enough since you cannot connect two sites in a straight line. For starters the earth' topography won't allow it. But then there's the issue that you won't connect every city to every generating site, but have a hub-spoke model (like ethernet). To give you an idea of the added distance compare the straight line distance of Atlanta to Pittsburgh to the interstate distance (520 mi to 680 mi).
Finally, the proposal of connecting two sites on opposite ends of the earth is fundamentally hegemonist. What are you going to do with the fiercely independent people that tend to live in deserts when they oppose you wallpapering their desert with panels? When the Bedouins of Algeria are blowing up the electricity needed in Japan, will the EU send storm troopers to punish the locals?
Which is going to be a fixed amount regardless of distance? While I don’t have the numbers, I would expect HVDC to be a dead end already if it made a difference to my conclusion.
> 20k km? Such a curious number, half the world's circumference, that I suspect you mean to propose to wrap the world in power lines to get power at midnight from noon elsewhere
Indeed, though this is likely excessive: summer/winter is significantly harder for storage to deal with — and hence the mostly likely to be solved with transmission — than day/night, given most people sleep at night but nobody hibernates though winter.
> But then there's the issue that you won't connect every city to every generating site, but have a hub-spoke model (like ethernet)
Surely a web is the best topology? Existing grids don’t all run on hub-and-spoke models: https://openinframap.org/#2/19.82/13.02/L,P,S
> When the Bedouins of Algeria are blowing up the electricity needed in Japan, will the EU send storm troopers to punish the locals?
While true, I’m not even going to try to play at global geopolitics; all I can say is that while deserts are a good place for PV, and while PV is the best thing I (being neither a biologist nor a geologist) know what to do with a dessert, the land area required for PV is low enough that even the UK — entirely north of the US-Canadian border — would be able to power itself in winter with no transmission and just a day/night storage system using 1% of its land area [0], so tiling the entire Sahara is unlikely: The same equations say that even if we electrify all world energy consumption, Algeria alone could supply everyone 6 times over, but you’d only do a global grid like this precisely if you didn’t want the huge storage system I’m implicitly adding to worsen the performance of my opinion and give your criticism the best possible chance.
[0] illumination = (cos(51 north + plus axial tilt) / π) * 1kW/m^2 ~= 87 W/m^2; efficiency = 20% ~> 17.5 W/m^2; 35 GW / (17.5 W/m^2) ~= 2000 km^2; U.K. land area ~= 242,495 km^2.
I think this is a bad idea, but economics beats aesthetics and someone would need to run the numbers that I don’t have to say which of storage and transmission is actually better. Perhaps the minima for the UK is PV in Algeria and HVDC, perhaps it is local x*PV + y*wind + z* batteries. It probably varies from any one country to the next.
Sure, but knock off another 10%
"I would expect HVDC to be a dead end already if it made a difference to my conclusion."
I didn't get that. Could you re-phrase?
"Surely a web is the best topology? "
My point you're underestimating the line length since you cannot connect generation to consumption with a straight line. Also, the hub-spoke or a web network would need the connections to be much larger than needed for that city (since the city would also need to transmit to it's neighbors)
"While true, I’m not even going to try to play at global geopolitics; [..] worsen the performance of my opinion and give your criticism the best possible chance."
You cannot dismiss cultural concerns. Look what happened in Kabul last week!
> "I would expect HVDC to be a dead end already if it made a difference to my conclusion."
> I didn't get that. Could you re-phrase?
If the conversion losses were as high as 10%, why did anyone bother using HVDC for e.g. a mere 85 km stretch in the Trans Bay Cable?
But even if they were 10%, that’s a multiplier not an adder so it makes a difference of 10% to final energy prices, it wouldn’t break anything (though it might shift the minima to batteries, as those are also cheap and getting cheaper).
> My point you're underestimating the line length since you cannot connect generation to consumption with a straight line. Also, the hub-spoke or a web network would need the connections to be much larger than needed for that city (since the city would also need to transmit to it's neighbors)
I think you’re overestimating how much the losses matter. You only lose half your power from 20,000 km, adding an extra 10,000 km from wiggling around a bit takes you down to 34% of original generation, and even then webs usually have high-capacity connections between important locations and lower-capacity links from the high capacity nodes to the low use nodes. Is anywhere in the US more than about 2400 km from the coast? Because 2400 km is a multiplicative loss of 0.918 compared to whatever losses you get from a trans-oceanic link to wherever (0.77 from Algeria to DC, according to distance from Wolfram Alpha, likewise 0.66 from Queensland to San Francisco). And the USA already has (several) grids, so you’d probably just want to link any global HVDC connection like this into each grid, rather than to each city.
> You cannot dismiss cultural concerns. Look what happened in Kabul last week!
“I am not qualified to discuss them” absolutely is not the same thing as dismissing them.
But my main point is more of we don’t need to tile the Sahara (or any single particular place), so it shouldn’t even come up in the first place.
1. Because AC also has transformation losses (2-3% per transformer. You need at least three (medium voltage, high voltage, medium voltage, low (house) voltage).
2. The answer is in the name, ain't it? It's underwater.
AC coupling in a dielectric increases losses very significantly. The longest underwater line AC line in the world is only 135 km, in Greece, for this very reason. The Greeks probably couldn't afford the capital to instal a DC line instead and settled for higher losses.
You pay for more expensive HVDC lines for three major purposes:
1. Underwater cables - see dielectric loss
2. Cheating on the synchronization
3. Long AC lines. This has many reasons:
- Take the speed of light, and divide it by 200Hz (50Hz * 4) = 300 000 km/s /200Hz = 1500 km. That's as long as an AC line can be before you start to deal with impedance reflections.
- Lowering losses. AC was famously adopted to lower ohmic losses by adopting high V. But, AC lines are coupled to ground (they're a massive capacitor to ground), so at a certain length you have to consider AC coupling losses that don't exist in DC
"But even if they were 10%, that’s a multiplier not an adder so it makes a difference of 10% to final energy prices, it wouldn’t break anything (though it might shift the minima to batteries, as those are also cheap and getting cheaper)."
I love how eagerly ppl dismiss 10% loss. I'm aware its a multiplicative - it's also an 11% increase in the amount of panels needed (1/0.9 = 1.1). You have to think of your losses in the inverse - how much added land you have to cover (and loose CO2 sinks to photosynthesis)
The things that are "cheap and getting cheaper" are things at a certain scale. Everyone loves things getting cheaper at scale, but forget that things also get more expensive at scale. Batteries don't get cheaper once you start running against the bounds of how much electrode material is available.
"I think you’re overestimating how much the losses matter. You only lose half your power from 20,000 km, adding an extra 10,000 km from wiggling around a bit takes you down to 34% of original generation"
so from you're original 50%, to 34%, a loss of 34/50, implying an increase in land covered of 50/34, or about 50% more land you have to cover.
50% more land if I accept your hand waving argument that you'll only need an extra 10Mm to get from point to point.
"“I am not qualified to discuss them” absolutely is not the same thing as dismissing them."
"I'm not qualified to discuss them" absolutely is used, de facto, as a license to ignore. It maybe the most common human rationalization.
"But my main point is more of we don’t need to tile the Sahara (or any single particular place), so it shouldn’t even come up in the first place."
You have to tile something. The Sahara makes most sense lacking vegetation. But sure, tile Ireland and see how much the local appreciate it.
Also: since previous post Google says HVDC transformer/converter losses are about 1-3%, not 10%.
> The answer is in the name, ain't it? It's underwater.
Could’ve put traditional AC cables on towers on the sides of any of the large selection of road bridges. They didn’t. Why?
(Likewise: as the waters are demonstrably spanned by bridges, they can be spanned by cable towers)
> so from you're original 50%, to 34%, a loss of 34/50, implying an increase in land covered of 50/34, or about 50% more land you have to cover.
In the context that one country can supply 6 times global demand an extra 50% (a hand-waving factor I expect to be excessive given most of the world is ocean and therefore not going to get so much wiggle) makes no real difference. It’s not going to be the limiting factor.
> You have to tile something.
Same point. Land area isn’t the limiting factor.
I see that you use 35GW - presumably the electricity use in the UK. However, with transport and heating, the total power use in the UK is about 190GW on average, and probably about twice that average in the middle of winter when everyone has their gas heaters on.
If you propose to replace fossil fuels, you need to account for electric heat pumps replacing gas heater, and electric cars replacing ICEs - which means an electricity consumption of 100-150GW, which means 4% of the UK area rather than 1%. (Assuming electricity storage is lossless which it isn't obviously.)
Hydro storage and battery storage have both been proven at scale and battery storage in particular is dropping in price.
Nuclear is very expensive in comparison.
There is a reason that even though we've had the technology to make moving mass batteries for decades, the only form of such energy storage in operation today is pumped hydro.
How many more decades before we can replace that with a renewable option?
I’ve watched people like yourself be wrong for 2 decades. I suppose in another 5 decades you might be right.
In the meantime, being wrong has a price as greenhouse emissions increase
IMO existing nuclear plants should be kept until they can be replaced easily, and the US may well take decades to do so as they are very slow on the uptake. Other countries are doing a lot better, some have reached 100% power from renewables already for short periods of time and are on track to do it for long periods in the coming decades.
Greenhouse emissions are not rising in the US, they are gradually falling as renewable options become cheaper and are built out and EVs spread:
https://en.wikipedia.org/wiki/Greenhouse_gas_emissions_by_th...
I'd expect them to keep falling given transport and electricity generation are the main sources.
Emissions are increasing globally because we are building a lot of coal plants.
The fact remains that we are at least 2 or 3 decades from renewables being able to replace fossil fuels for power generation
That’s half a century of waiting
Why do you insist on attempting to set renewables and nuclear in opposition?
Scotland for example recently hit generation of 97% of demand from renewables in the year 2020. As part of a larger grid with appropriate balancing and storage, 100% renewables is absolutely doable, and as part of a grid with some base load generated from nuclear (as the US and UK for example already have and is not going away soon), it's definitely doable right now with sufficient will and resources:
https://www.bbc.co.uk/news/uk-scotland-56530424
The only reason the US is trailing so far behind other countries is a lack of will to transition off fossil fuels, a chronic lack of investment in infrastructure and a balkanised electricity grid run for profit, all of which is fixable.
Roughly one. According to Wikipedia, the US added 35 GW of each of PV and wind between 2016 and 2020, and has 95 GW of nuclear. Accounting for capacity factor, the quantity of renewables TWh produced roughly doubled, and is currently 14% vs. 20% for nuclear, but (worldwide at least) renewables are close to an exponential trend.
The main constraint the USA should have here is storage and/or long-distance transmission.
Such is exponential growth.
Though you’re right: needlessly closing nuclear reactors is bad for the environment as the most important problem right now is the CO2. Coal needs to go first, then oil, then natural gas.
None of which work beyond a few hours. The sheer number of batteries and the amount of pumped hydro needed is phenomenal. As one data point, Elon Musk's much heralded $100m battery in South Australia stores a few seconds of Australia's total energy use. A few tenths of a second for the US.
When it is cloudy and there is no wind for a week, renewables are a complete failure.
> is dropping in price
Batteries have been dropping in price for 140 years. But the best batteries have only been getting better at about 7% per year. Specific technologies have improved faster than this early on in their lives (e.g. LiIon) but during most of this improvement they were far from the best.
It also doesn’t produce a ton of power. The Three Gorges Dam is a colossal project, displaced literally millions, cost a frightful amount of money, and produces 1% of China’s power. Also, if that thing ever breaks, many millions will die.
Should have been a couple of small nuclear plants.
Regarding hydro storage, it’s the same exact thing, except now you need two dams, one high and one low. This isn’t a practical solution everywhere, certainly not at a national scale.
It produces 1% of the electricity used by CHINA the premier industrial and most populous country in the world and you dismiss it? Three Gorges was a mistake, but it generates power used by 13 million people. That is a ton of power.
Most gas plants are in the 300 MW size, i.e. the 3GD is about 80 of them.
Either way, no more of them will be built - there are not many more gorges like that left.
We need heavy-duty nuclear plants for base load, solar for daytime peaks, and probably some natural gas here and there