Power cables are getting cheaper and cheaper. The expensive part used to be the voltage conversion stations at the ends, but with mass production of MOSFETs for EV's these have now become far cheaper than the JFET's and other exotic silicon that used to be used.
In turn, that means voltages can be higher, letting one use more of the cheaper PVC or XLPE insulating material and less expensive aluminium for the same amount of energy delivered a large number of kilometers.
To be honest, I don't think we're many decades away from the cable+conversion stations themselves cost being irrelevant, and the administration costs, land purchase costs, etc dominating.
> The expensive part used to be the voltage conversion stations at the ends, but with mass production of MOSFETs for EV's these have now become far cheaper than the JFET's and other exotic silicon that used to be used.
Why do you believe these things are related?
HVDC lines operate in the hundreds-of-kilovolts range. For example, https://en.wikipedia.org/wiki/Basslink operates at 400kV. There are no MOSFETs or JFETs directly involved in stepping down that power.
Semiconductors are stackable to get higher voltage. They're parallelizable for more current.
Cost scales linearly with voltage and current, and is therefore constant WRT to system power.
Thyristors require you have at least one transformer operate at AC line frequency (50/60Hz). That costs a lot, since you need enough steel to store 20 milliseconds of your total power as a magnetic field. Thyristors are on-off devices (like most semiconductors when used for power conversion), but cannot turn off without zero current, which precludes a bunch of high frequency designs which are better for harmonics and weight-of-steel.
Overall, they were a popular choice in the 90's and 2010's, but I don't think we'll see any new designs installed with them.
I've never heard of MOSFETs being used in extra-high voltage systems, but I have not been following the industry for a while. Do you have any links? I've only seen IGBTs or older technology used.
Nah - the insulation material costs ~ $0.80/liter, whereas aluminium conductor costs $6.50/liter.
If you can have the conductor 1mm^2 thinner (capable of carrying less current for the same heat production) and the insulation 1mm^2 thicker (capable of handling a higher voltage) and transfer the same power, then you'd save money.
It only works up to a certain limit obviously - the relationship is non-linear and there is an optimal point.
The actual tradeoff involves a lot more modelling, because you need to consider all kinds of other factors, not just the costs of the conductor and insulator.
The problem with long distance AC is the reactive power component caused by the capacitance, and the voltage rise caused by the Ferranti effect.
The reactive component has significant impact on the generation equipment and grids. It also causes the Ferranti effect, where the voltage along the cable rises. This can make managing the voltage within the cable difficult because at no load, the load end has a higher voltage than the source, and when loaded, the middle of the cable has a higher voltage than both ends.
During stable operation these effects can be managed with Statcoms, shunt reactors and voltage regulation tap changers. However during transient operation you will be relying upon the static protective devices such as surge arrestors, depending on how large the transient is.
DC transmission does not suffer from the same reactive power component and has less losses, but it does require large convertor stations at both ends.
It doesn't seem like anyone directly answered your question. As far as I am aware, all long distance undersea power cables are high voltage DC. I believe this has to do with the efficiency of power transfer over long distances.
AC loses power by inductively and capacitively coupling to nearby objects. It's manageable at medium distances above ground, cheaper than a pair of converter stations. However, water is much more conductive than air and losses from an underwater AC cable would be much greater.
AC is a sine wave, of which the peak is a factor of Sqrt(2) higher than the DC voltage. That means your insulation needs to be sqrt(2) thicker - ie. 41% more insulation material.
On top of that, you also have losses to the cables capacitance with AC.
But DC has the cost of the conversion stations to consider - both capital cost and efficiency causing operational cost.
> But DC has the cost of the conversion stations to consider - both capital cost and efficiency causing operational cost.
I suppose you mean AC-DC conversion stations. Assuming only solar energy will be "pumped" over the wire, then the "only" conversion stations that are needed are at the consumer, right? I said it before, I don't know much about electricity, so please correct me if I'm wrong.
> It’s really difficult to make solid state components that work at million+ volts.
You can split (or add up) the million volts as transmitted at either end so the individual components only work across a small fraction of the 1MV potential difference. This is how can get 12V from 1.5V batteries or use 1V LEDs from a 12V line.
Amusingly (IMHO), The band's line-up remained the same for 20 years until 2014 when Malcolm retired due to early-onset dementia, from which he died three years later; additionally, Rudd was charged with threatening to kill and possession of methamphetamine and cannabis. Stevie, who replaced Malcolm, debuted on the album Rock or Bust (2014). On the accompanying tour, Slade filled in for Rudd. In 2016, Johnson was advised to stop touring due to worsening hearing loss. So a rocker's fate: forgot what planet they were on, went mad on drugs becoming threats to society, lost their hearing, or kept touring indefinitely with a changing lineup cashing in on past glories.
is there any thing special about the nature of such project that makes you ask this question? By default, long range transmission is always DC for that exact reason.
Aluminium is far less dense, which in turn makes the whole cable bigger, which has other costs (eg. fewer kilometers of cable fit in a boat). Usually it's still the best choice overall though.
> I would have guessed there must be enough domestic customers or in Indonesia that would make more sense.
Australia is a big place. The northern tip of Australia, where this project is based, isn't really that much further from Singapore than from the Australian population centres in the South East of the continent.
Indonesia is much poorer than Singapore, and has awfully inefficient bureaucracy and regulatory environment.
> proposing unrealistic nuclear solutions to seriously focus on renewables.
they're doing unrealistic nuclear proposals, because they know it takes a long time to ramp up, and in the mean time, their buddies' investments in the coal industry gets time to exit and profit properly. It's designed to prevent losses in fossil fuel investments.
Not to mention that australian nuclear cannot be profitable imho - not when solar is so cheap. Their current proposals for nuclear basically requires taxpayer subsidies.
50% of Australia lives in Brisbane, Melbourne and Sydney. Having a nuclear power plant for each would make sense. Melbourne would make the most sense first as it gets a lot less sun than the others.
Meanwhile nuclear is feasible in China, South Korea, maybe in the UK (who are well into sunk cost on their next reactor already), and probably in the US.
My understanding is that I the time it takes to build a nuclear power plant, a helluva lotta solar power generation can be built and up and running and generating power.
And in that time span as well, solar power will increase its efficiency.
And then batteries, to store and deliver that power outside of generation hours, are a parallel to that.
If a nuclear power plant could be built quickly and simply, the equation would be different.
Unfortunately, from the limited amount that I've read, nuclear power plant projects often run over time and over budget, exacerbating the time scale issue I described above.
I don't think that's actually true. US Navy and their contracting shipyards had consistently built nuclear subs in 3 year strides for decades. One set of fuel lasts is good for 1/5th century, after that the sub needs to be cut up and refueled. It's not something that take years after years of permitting and change of plans and suspected acts of arson of unknown motivation if it's literally operated by US Army or Navy(but not NASA).
Solar power is just amateures littering compared to that.
there has been an unfortunate "phase shift" since 1970 in the nuclear energy industry/ecosystem, mostly because the risk engineering principle/mandate called ALARA (as low as reasonably achievable), and of course reasonable does not mean profitable. (which makes sense, we want safe reactors not just "there was a safety budget, and we spent all of it" >>safe<< ones, right? sure, but the real world is stubbornly full of cost-benefit trade-offs, and apparently we crossed it somewhere during the 70s.)
Nuclear is held to a much higher safety standard (eg in terms of deaths per Joule) than any other form of electricity production. And that includes photovoltaic!
Nuclear is so safe--even fully factoring in the accident at Chernobyl--that people very occasionally falling off rooftops when installing solar panels is a bigger health hazard per Joule produced.
Sure, please adjust the numbers for when we had to evacuate cities for nuclear scares. You can do calculations in 'quality adjusted life years' or some other ways to convert deaths and injuries and the cost of evacuations. It doesn't really change any conclusions, even with very pessimistic estimates. I just picked deaths, because they are relatively easy to get clear numbers for.
And don't get me wrong: solar is mostly fine anyway. It's coal that's really obnoxious. Both in the mining and in the burning, and in the accidents. (And to a lesser degree other fossil fuels.)
Photovoltaic is great! On a purely technical level both solar and nuclear can work well, nuclear perhaps a bit better and we had the technology for longer. On a practical level, solar will win, because people fear nuclear.
All electricity generation methods have engineering challenges. Eg solar has some big problems with daily variations and seasonal ones. We can solve the former with batteries, and the latter via big cables to (sub-) tropical regions.
Wind is also great! And we've only just started tapping waves and tides, too. And geothermal.
nuclear safety has changed a lot. even though "walkaway-safe passively cooled" is not a technical term, but that's the design goal nowadays.
the real problem with nuclear is that the market is small, fragmented, US regulations are bad (as I elaborated upthread), so there's no real volume, no economies of scale, no healthy competition and there's basically no innovation even around the safety critical core...
1) The risk of evacuations happening is tiny and I'm not even convinced it is still a factor. We've not yet seen a messy meltdown of any plant designed and built after Chernobyl in 1986 and designs have changed a lot since then.
2) We don't know what a large-scale solar disaster looks like yet, but they might happen. For example I recall the Wikipedia page for the Year Without Summer [0] - we know that sometimes nature puts things in the atmosphere that might hamper solar in a way that nuclear can be designed around. IE, we might find we now have a risk of our power stations just deciding to produce less one year because of a usually unrelated disaster. Or maybe even stop if there is enough volcanic ash.
Plus renewable projects have had a more noticeable association with grid failures and mishaps than nuclear projects. We really don't have much experience with what mass solar failures (if they do exist, but they probably do) look like or how common they are.
How much bigger of a health hazard is manufacturing/installing solar panels compared to nuclear? Let's say, per one terawatt-hour of produced energy, how many people die doing each?
First you’re going to need reliable worker safety data and population cancer rate data out of China (which makes almost all panels), which…. Good luck.
Nuclear power plants are unrealistic to build in short time frames, such as trying to meet agreed green energy targets. Part of the Nuclear proposal being put forward by Australian conservatives includes dropping out of the Paris Agreement and refocusing on a 2050 time frame (ie. past the politicians' retirement age)
If we had the renewables to replace the coal politicians would love it to retire in a heartbeat. The reason it’s sticking it around longer is because politicians fear the backlash from blackouts and high prices more than the backlash from the bad PR of delaying closures of coal.
I'm unsure about Indonesia, but domestic customers in that region would be pretty limited. The closest major power users would be in Queensland (>1000km) away.
If Australia refined _all_ of the 40,000kt of Bauxite we export each year into "frozen electricity" Aluminium, that'd only require about 600GWh, or about 4% of the 1.7GW 24x7, or 15,000GWh per year this would send to Singapore.
Large datacenter are in the 100MW sort of range, so only single digit GWh per year.
Australia generates a few hundred TWh per year. 272 TWh in 2021/22 - or 272,000GWh, around 20 times what this project will export to Singapore.
Data centers and Aluminium and Iron smelters are big electricity consumers. But they barely even move the needle compared to cities with millions of households.
Approximating bauxite as pure aluminum oxide [1], 40 million tons of bauxite contains about 21 million tons of aluminum. A ton of aluminum takes about 14 megawatt hours of electricity to produce [2]. That would be about 294,000,000 megawatt hours (294,000 gigawatt hours, or 294 terawatt hours) to turn Australia's bauxite exports into aluminum. Australia could easily double its electricity production/consumption to refine bauxite into aluminum metal instead of exporting the bauxite.
You're off 3 orders of magnitude, 40,000,000,000 kg x 15,000 Wh/kg = 600TWh (you likely tripped on the kt, which is 1000x1000kg, at least I did the first time I ran your numbers). That's not 0.2% of Australia's energy use but 200%.
Ha! It figures. Further down that thread I wrote: "Also, I'm notorious for dropping three orders of magnitude when doing mental math using kilo/mega/giga/tera prefixes."
Turns out when you do the math right, Aluminium _is_ frozen electricity.
Natural resource sales send USD to Australia. AUD is now worth more because it is backed by more USD. Manufactured exports are also traded in USD, so Australian exports become much more expensive because workers and local materials are paid for in AUD.
For that, you'd need to make massive investments in a part of that world that has mostly untouched nature.
It might or might not be a good idea. But you need to then compare those massive investments to the relatively modest investment of the power cable to bring the electricity to a part of that world that already has all the other infrastructure needed, and also already has lots of water.
Singapore has no strategic depth anyway, becoming dependent on importing power isn't some extra vunerable vector vs building domestic generation that likely can't be protected long term. Current is Singapore military vs region is like PRC:TW in the 90s... back then TW with US equipment was one of the more potent forces in the region and could stomp far larger/poorer countries with inferior hardware. But advanced equipment can only scale so far vs quantity, and as rest of ASEAN gets wealthier they're going to build out more modern capabilties, at scales that rich but small Singapore won't have the resources to defend against. If anything integration with AU, with military infra (and future US B21s) is probably more secure / geopolitical hedge against other's meddling.
While Singapore is a surprisingly martial country, if they get into a war with anyone in SEA they're running a very real risk of being destroyed. Indonesia alone has 5x their GDP and 20x their population. There isn't much difficulty choosing which city to target first when going up against Singapore either.
In Singapore's situation, they can probably invest assuming that they are not in a military conflict with anyone. If they get into a war with anyone who can cut that cable they will be returning to the stone age anyway. If Indonesia objects to them they will go, if someone with the power to coerce Indonesia objects to them they're in deep trouble.
> While Singapore is a surprisingly martial country, if they get into a war with anyone in SEA they're running a very real risk of being destroyed. Indonesia alone has 5x their GDP and 20x their population.
Wikipedia gives an estimate of $1.47 trillion for Indonesia's GDP in 2024. The estimate for Singapore is $525.228 billion. The factor seems to be less than 3x. Where do you get 5x from? Are you going by PPP or so?
> In Singapore's situation, they can probably invest assuming that they are not in a military conflict with anyone. If they get into a war with anyone who can cut that cable they will be returning to the stone age anyway. If Indonesia objects to them they will go, if someone with the power to coerce Indonesia objects to them they're in deep trouble.
You can't make those assumptions, if you don't want to be bullied. Singapore doesn't have that cable right now and we ain't in the stone age. That situation ain't no different from having a cable, but it being cut.
I was looking at the PPP figures. By accident as it happens, I was looking at the first box in Wikipedia with "GDP" in it. But I think that is still fine in this context.
> You can't make those assumptions, if you don't want to be bullied. Singapore doesn't have that cable right now and we ain't in the stone age.
You aren't at war either as far as I know, and hopefully it stays that way. But if Singapore happens to be at war with someone who thinks cutting that cable is a good option then the stone age beckons. And not because of the cable.
Yeah - from a purely technical point of view it seems strange that you'd run a power cable 2000 miles to Singapore to service 4 million people, running alongside the coast of Bali, Java and Sumatra - population 210 million.
Presumably those in Singapore have a lot more buying power though. And the politics are more favourable for big capital investment projects.
Yeah, they also have zero room left so I guess the option was between more dirty power stations in Malaysia or this. Seems like a wise, forward-looking initiative.
Singapore has plenty of room left, and we are making more via land reclamation. The question is just one of opportunity costs: what else could you do with the land?
I'm pretty ignorant about Singapore, but... I get the impression it's quite small. Wikipedia says 750 sq km.
The solar farm powering this Suncable project is 12,000 hectares, or 120 sq km. So the solar farm is 1/6th the size of Singapore. Although Singapore is only planning to buy around 1/3rd of the capacity, so maybe this'd be equivalent to only 40 sq km, or 1/20th the size of Singapore.
I suspect there are more profitable uses to the Singapore economy for land reclamation than dropping solar panels on it?
No one applies economic cost benefit analysis to buy a Louis Vuitton bag for $50,000. Prestige, signalling, membership to exclusive club, etc dominate the consideration.
Reserves are cash in hand and represent immediate and hard spending power.
> Prestige, signalling, membership to exclusive club, etc dominate the consideration.
So? These _benefits_ also fit into a bog standard cost/benefit analysis. For example, Singapore would need to weight this project against buying everyone a luxury handbag..
Btw, in any case keep in mind that the project is privately financed and will make money selling electricity to Singaporeans. The electrons that power my gadgets at home don't have any colour, so I can't even tell if my electricity comes from a particularly prestigious source. It's all intermediated by the wholesale market.
> Reserves are cash in hand and represent immediate and hard spending power.
That's about on the same level as arguing that having a money printing press represents raw spending power.
Most central banks around the world conduct monetary policy via domestic interest rates and affect these interest rates by buying and selling domestic government bonds. Thus they will have lots of government bonds on their balance sheet. But it doesn't mean that they can just take these bonds and use them to buy solar farms.
The Monetary Authority of Singapore is (almost?) unique in foregoing interest rate as a channel of monetary policy, and instead working via the foreign exchange rate. They affect the foreign exchange rate by buying foreign currencies via freshly minted Singapore dollars (or selling them to remove Singapore dollars from the market).
And just like the American Fed keeps the government bonds they buy on their balance sheet (and pretty much has to!), our Monetary Authority of Singapore keeps the foreign currency on the balance sheet, and they show up as reserves.
By design, Singapore has at least as much in foreign exchange reserves as we issued domestic currency.
In a sense, most of the eg Euros in our reserves are already 'spent', but they are spent in the form of SGD in circulation. (I say 'most', because we have more reserves than we issued SGD. Singapore is cautious like that.)
Indonesia is ~ 17 thousand islands, many steep equatorial jungled volcanic slopes and at 275.5 million is the fourth highest population for a country globally.
Land is in tight demand with food a priority over panels and issues that may not be apparent (clear slopes leads to instability, and keeping them clear is a Sisyphean task, etc).
I do not believe that we can design a system that will withstand waves and wind from tropical monsoons or even most tropical storms or cyclones in the pacific. I can't speak for other oceans or areas of the world, but I believe that this design requirement will probably make it a non starter.
> The approval paves the way for the next phase of development to deliver industrial-scale electricity to customers. But it still has some way to go, with a final investment decision not expected until 2027.
and
> However, SunCable still needs to negotiate Indigenous land use agreements with a number of different traditional owner groups along the transmission line route to Darwin.
In turn, that means voltages can be higher, letting one use more of the cheaper PVC or XLPE insulating material and less expensive aluminium for the same amount of energy delivered a large number of kilometers.
To be honest, I don't think we're many decades away from the cable+conversion stations themselves cost being irrelevant, and the administration costs, land purchase costs, etc dominating.