Hacker News Reader

Comment by seiferteric | Hacker News Reader

It's a great point. Even in my spot in northern California this winter, it's been so cloudy this year, I can't imaging an all solar+storage grid would have worked out well this year. In this case having electricity that costs 45% more when it's cloudy is a lot better than none at all. Actually this gave me an idea, it would be really neat to have an comprehensive simulator of power grids that incorporated weather, demand spikes etc. to play around with different energy source mix to get an idea of what actually works and when it fails as well as total cost, environmental impact etc.

View on HN

> Even in my spot in northern California this winter, it's been so cloudy this year, I can't imaging an all solar+storage grid would have worked out well this year.

The electricity demand in the winter is much lower than the summer demand, especially in California due to residential AC.

> Actually this gave me an idea, it would be really neat to have an comprehensive simulator of power grids that incorporated weather, demand spikes etc. to play around with different energy source mix to get an idea of what actually works and when it fails as well as total cost, environmental impact etc.

There's a gentleman in Australia who does ~this for their market -- 5 hours of storage is enough to get to a 99% renewable grid there;

https://twitter.com/DavidOsmond8/status/1640959088887730176

With heat pumps coming online and gas being phased out that might not be true for very long. Also is 99% even close to good enough? That would be like 4 days a year without power right?

The premise isn't that you'd just sit in the dark when there isn't enough renewable + storage -- you'd run nat gas peaker plants or some carbon-based source for those 1% of times when it's needed -- would still almost entirely decarbonize the grid.

How do you make it worthwhile for someone to have a natural gas peaker plant ready to run at any time for only 1% of hours in a year? Where does that cost figure in the tradeoff between renewables + batteries vs nuclear?

We have really good insight into week-ahead curves for wind and solar, and high quality prediction for day-ahead production, so it wouldn't need to be ready "at any time" like its usage would be a surprise. And you incentivize it like anything else, providing either a $/kwh figure that makes it economical for them to capture their costs or subsidize the construction and standby nature of them like we currently do with black start generators (https://www.kut.org/energy-environment/2021-08-05/if-the-tex...).

It's hard to make a reliable cost prediction comparing nuclear vs. wind/solar + batteries since we don't know how to build nuclear any more.

Vogtle 3/4 are going to cost maybe $40 billion when all's said and done? With OpEx, you get to something like $0.18/kwh. That's more than 5x the cost of unsubsidized wind or solar installations which would buy you a bunch of storage.

I agree the reserve mechanisms provide the incentive to pay generators to exist with an entire gas turbine, generator, switchgear, operations staff, health and safety staff, maintenance staff, transformer, interconnection, land, environmental permits, hr and accounting departments to be “available” but not generate except in the time that renewables can’t. The only difference in cost for the plant to exist and not run but be ready and to run is the cost of the fuel.

So we still have to pay for dispatchable generation if we want to have power on a calm cloudy day week. We can add that cost to the cost of the storage and overbuilding of capacity that allows solar or wind to deliver rated power overnight. Or live with blackouts

> The only difference in cost for the plant to exist and not run but be ready and to run is the cost of the fuel.

Which is over 50% of the cost of running a combined cycle plant (page 12: https://www.lazard.com/media/sptlfats/lazards-levelized-cost...) -- but also misses that you'd need far fewer plants if you build renewable + storage generation to match the 99% use case so the total cost spend on peakers would drop dramatically even if some were still needed to provide backup generation.

Also has the added benefit of almost entirely decarbonizing power generation.

And induction cooktops, and home charging BEVs overnight…

my back of the envelope calculation based on current human energy usage and battery production, it would take more than 150 years to manufacture 5 hours storage capacity. Of course manufacturing capacity is increasing and efficiency improves with electrification, but so is energy consumption, plus you have to factor in battery replacements during that time... Ya its going be tough to do that.

No offense to your envelope, but that feels like a silly assumption. We produced ~15GWH of batteries in 2010 and now that takes us something like 3 weeks? Planned capacity additions will make that a daily production amount by 2030, and each subsequent cell is cheaper than the one before. This also assumes Li Ion as the only game in town. Plenty of other more likely options (flow batteries, hydro, etc) are feasible on a 15-year horizon.

5 Hours of electricity is something like 13 TWh, so if we get to 5,500 GWh annual production by 2030, it would take ~2.5 years to provide 5 hours of global electricity storage. Handicap it all and double the electricity requirement and halve the annual production figures and it's still only 10 years' capacity to go to a 99% carbon-free grid.

No 13 TWh is just our current _electricity_ consumption for 5 hours. We are talking about replacing all energy consumption like heating with electric heat pumps etc. This would be closer to 100TWh. And google says our current production is around 600GWh of batteries per year. 91.89 TWh / 600 GWh/year = 153 years.

Yes there is pumped storage, but I wonder how much more of that there really is to develop? Plus damns are environmental disasters of their own. The other storage methods are speculative at best. I just wanted to push back a bit on the idea that "just" need solar + storage as many people seem to believe. It is a big ask, and I think we will need more. If we don't want gas plants, then nuclear could be a good option.

Nah, that's ignoring that this hypothetical is in a world with mostly renewable power generation. E.g. most of the energy in that mix you're referring to is oil/natural gas/coal -- and over half of that is lost before it reaches the first power pole or the drive wheels. Vastly more efficient electric devices will greatly reduce the amount of energy needed. E.g. a heat pump with COP of 4 needs 1/4th the energy as burning natural gas in a high-efficiency furnace. Yes it shifts that energy demand from natural gas to electricity, but total energy consumption still drops. Similarly, a gas Corolla getting 40mpg equals about 1.2 miles/kwh, whereas a Tesla Model Y is more like 3.5 miles/kwh, so again, even though there's an increase in electricity consumption, total energy consumption decreases by ~2/3.

My other point stands though -- even if you grant a much larger electricity demand, you can't just look at today's battery production capacity. There are pipeline projects that will 10x that by 2030 (which is what happened in the past 10 years). Even with pessimistic assumptions, you get to pretty reasonable time frames pretty quickly.

Yes efficiency improves with electrification (I said so above) but step back for a second. The premise of this was that 5 hours storage would be "good enough" but my original point is that with change in electricity usage brought on by bevs, heat pumps etc, that this may not be the case forever. What if it is 10 hours? Or 20? Also, you have to keep in mind developing countries rapidly growing energy demands. And yes, battery production will increase, probably exponentially.. until it doesn't, everything is an S-Curve in reality, so it all depends on what we can realistically produce.

> everything is an S-Curve in reality

The only thing that limits battery production is demand. There will be plenty of materials available and plenty of machine-makers exist. That S will go as big as we need.

That's the kind of logic that got us here.

The "there will be plenty of oil" logic got us here because it was right, so I'm not sure what you're implying. It doesn't reflect negatively on batteries unless ability to scale up is inherently bad, and I don't think it is.

It's a more complex calculation than that because not all the energy storage is going to use battery electric. Pumped storage hydro is often pointed to as one option for mass quantities over a long durations, and somewhat more esoteric stuff like low-friction flywheels comes up for "capacitor" type loads where what's needed is a cycle of a few minutes.

The advantage of the batteries is specifically in power density, which makes them suited for mobility and consumer convenience. But the more you get into seriously optimizing electrical storage for scale, the less it's going to be about one specific mode.

> The electricity demand in the winter is much lower than the summer demand, especially in California due to residential AC.

My mom's house doesn't need A/C but it does need a lot of heat in the winter. We looked at moving her to solar/battery and an electric heat pump instead of her gas furnace. With even a tiny bit of trees nearby and otherwise pretty good exposure we were told it was going to be hard to make it worth it. Her demand in the winter would be pretty high.

We generate 3x our needs in summer; 1/3 of our needs in winter, when we are running electric air source heat pumps at 6000' in northern NM. If we had a passivhaus-level home, I think it could be possible to satisfy the winter heat requirements with the solar we have, but not with anything less (and even then, it would be challenging).

Interesting, I appreciate the information. What is the nominal capacity of your solar array?

6.7kW

We typically generate 800kW - 1300kW per month (more in the summer). Our panels are angled at 30 degrees to slightly preference winter generation.

Great, thanks!

Reply on news.ycombinator.com