Hacker News Reader

Comment by sealeck | Hacker News Reader

> Unfortunately, California is a terrible benchmark. It is as close to ideal for Solar as it gets. Most places are not going to see this kind of performance

We can also build power lines! Between different places! Such as the places with lots of sun, and the places without lots of sun!!!

View on HN

Solar has gotten so cheap that it's cheaper to just overbuild solar in cloudy places than to build long distance HVDC lines between sunny and cloudy locations.

https://caseyhandmer.wordpress.com/2020/12/27/the-future-of-...

The great thing about solar is that you can put it nearly everywhere. Avoid the transmission losses and put the panels close to the load, even if it’s not as ideal as one state over.

It’s easy to put extra solar panels into a system to make up for reduced average sunlight. It’s standard practice to have a ratio of PV capacity to inverter capacity of something like 1.2:1. In a low sun location you could bump that up to 1.5:1 or higher.

It consumes a lot of power to transmit over long distances. From what I understand it's basically always preferable to generate power as close to where you are using it as possible.

Absolutely, but besides rooftop solar we don't have a single power source currently that does that.

Traditional power generation was always centralized big plants. Most people wouldn't want to live next to them and from a health perspective you probably shouldn't.

So "close" has always been within air quotes.

Also small-scale fossil fuel, hydropower, wind, geothermal, etc. What solar does differently than those is be directly usable and not have significant negatives which made it undesirable to have near non-industrial users because it doesn’t make noise, air, or water pollution.

Batteries are significant amounts of "nearby risk" that is being handwaved away imo

They are fire/explosion hazards, heavy metals, etc.

They’re a fire risk but have you tried fossil fuels and power lines? The newer battery technologies have significantly minimized the heavy metal needs and especially the sodium-ion batteries really reduce the fire and explosion risk.

We aren't talking about installing coal power plants in people's houses the way we're talking about installing solar and batteries in people's houses

People use gas generators, though, and it’s not like you have to go back further than my grandfather’s lifetime to have factories or other industrial buildings with oil/coal plants on-premise or directly powering machinery. Pollution and logistics pushed away but we didn’t start with a pervasive power grid.

I have a gas line directly into my apartment building, as do the vast majority of buildings in my area.

Genuine question, what do you mean by small-scale fossil fuel

What power output in megawatt are we talking about here? I'm struggling to think of a fossil power source efficient at small scale

Well, if you go back far enough there were a ton of factories and other buildings which had their own small plants but the main thing I was thinking of was stuff like that xAI data center in Memphis with the methane-powered turbines where they’re avoiding grid limits and transmission losses at the expense of pollution.

Well, nearby is relative.

My point is that a nuclear power station near a city is probably better than a wind farm offshore 1000km away even if the wind farm and the nuclear generate the same

It's hard to define 'better'

Nuclear has significant downsides besides the waste and proliferation risks.

You wouldn't want to build a grid on just nuclear.

Let's assume that city takes 1.2MWh at its peak every day. That would mean you need to be able to supply that. So you build a nuclear plant producing 1.2MWh of energy.

Now you have the argument against renewables (the sun doesn't always shine) in reverse. The city doesn't always need that peak power. And nuclear is the slowest of all power sources to tune up and down in terms of output.

Nuclear for base load makes a lot of sense as it'll always be fully utilized. But nuclear to power a grid 100% doesn't exist anywhere for a reason either.

That wind farm needs to be like 4% bigger to cancel out transmission losses so the question has to involve the relative costs.

The EIA has it at $81 for advanced nuclear, which is competitive with offshore wind ($88) but not hydro ($58), PV+battery ($53), PV ($31), or onshore wind ($29). Now, both of those could see big improvements with scale but I think the uncertainty of the markets and our political climate are going to complicate that a lot. A big nuclear push needs a lot of upfront funding and while Trump has boosted it a bit and hates wind, I’m not sure how much that counts on a loan that big since right now that plant is guaranteed to lose money most of the time unless it’s near a big industrial user with high baseline demand.

> It consumes a lot of power to transmit over long distances.

No, it doesn't. This is plain wrong. Power loss for a 1000 km HVDC line is 3.5% [0], which is the same order of magnitude as battery losses.

If you had of said it cost a lot of money to transmit over long distances, then fine.

[0] https://en.wikipedia.org/wiki/High-voltage_direct_current#:~...

Who told you that? I build power lines. It consumes a couple percent.

RE ".....We can also build power lines! Between different places! ...."

We certainly can. However it is expensive. In my country (Australia) is is estimated to be cira $20 Billion

That's a steal compared to racing toward extinction.

You need to persuade the median voter about this, and looking at the state of Western politics right now, it seems to me that peak willingness to spend huge money on environmental projects is behind us.

I agree with you, but surviving is worth more than almost any amount of money and that fact doesn't depend on the median voter's opinion.

20 billion for infrastructure that lasts 50 years is 400 million per year hardly a major issue for a 1.6 Trillion dollar economy.

The US recently spent more than that on a single nuclear power plant.

> 20 billion for infrastructure that lasts 50 years

Average solar panels last 20-30 at most right now just fyi

50 years is practically right out of the question

There are several 50 year old solar installations that are still working fine. They've only degraded 10%-20% or so. The reason that they are typically replaced after ~20 years is that the technology has gotten so much better that you can replace the old panels with new ones and get 4X the power in the same space. The wires and the mounts are worth more than the panels, so reuse those with modern panels.

Do you have any real study on this?

Considering they work by basically taking electrons from the material, they are guaranteed to get much less efficient and I doubt it's a linear effect. I think once you have destroyed the first layers, it becomes much more complicated to get meaningful power depending on variables.

Solar panel talk always focuses on the ideal conditions like California, but you have to account that plenty the energy transition is necessary in place where solar panels are not that efficient to begin with.

“We can also build power lines! Between different places!”

The actual panels themselves are a different kind of infrastructure.

20 Billion is a rough guide. Apparently this estimate is approximate to 100% It also depends on lots variables, like - for a large powerline the towers can cost $1M each. I also assume the initial cost does not include maintenance costs.

Maintenance costs exist, but they can also extend the lifespan well past 50 years. So it’s all a big cost vs benefit tradeoff where maintenance occurs because it’s a net gain.

The point is the upfront costs aren’t actually that significant.

But if you're from Australia you should know that your country isnt exactly typical?

There's a few big metros the rest of it is sparse, where sparse is a stretch? Isn't Perth like the most isolated city in the world?

Perth and that part of WA has its own generators and independent electrical network. Darwin in NT Australia is probably the same situation too. Large expanse's of desert and very sparsely separated areas to the east coast.

You say that, but recent regulatory history suggests that building power lines in the USA isn't necessarily something which can be done.

one state supplying the rest is not what you want, especially if there are chances that something will disturb that states grid.

I remember the uni day discussions about Africa supplying the rest of the world with solar energy and that the material requirements for such an infrastructure should become a thing around 2025 - 2035 ... then someone explained climate change and hinted at the exponential function ...

back to topic: you'd have to maintain an "inert" backup, which isn't portfolio-communist-economically "viable". or you share the load "as much as necessary", which would still become an issue if any of the suppliers have a fallout ...

That was a real plan

https://en.m.wikipedia.org/wiki/Desertec

one of the projects listed is probably somewhere in my bookmarks.

and AFAIK it still is a real plan but, officially, we are not where we wanted to be material science-wise ... heat, distances and batteries are still somewhat an issue, meaning, not efficient and cheap enough to be a no-brainer (despite already being a no-brainer but, you know, wolfs of the wall street bangers are super cool and smart and who wouldn't be partisan with leBuffets et al???)

Reply on news.ycombinator.com