It's called the Duck Curve because the graph kind of looks like a duck if you squint https://en.wikipedia.org/wiki/Duck_curve
It's a familiar problem to everyone in the energy industry. Solar alone isn't enough, you also need energy storage to time-shift the generation and the load.
On an individual scale, if you're into home automation you can do some tricks like pre-cooling the house below your target temperature while you're gone for the day. At one house where we had significant excess solar I briefly entertained the idea of having a cooling vent into the attic to try to make that space cold on days with excess generation available during the day.
My attic has spray foam insulation and the HVAC system (also in the attic) already does this to a mild extent. I can go up there in the middle of an August day and not feel it at all.
Allowing the ceiling above your living space to reach 140F+ starts to get into radiative heating issues. You don't need to cool the attic to 72F. Anything would make a huge difference. Radiative heating scales quartically with temperature.
https://www.energy.gov/cmei/buildings/articles/unvented-cond...
My idea was to use excess electricity generation to cool the attic somewhat below the temperature of the living space. That space would then sink some heat after the sun went down, reducing the cooling demands in the evening.
An analogous idea would be to have a large water tank in the basement with a chiller attached. You run the chiller when you have excess electricity to get the water colder than the house, then let it passively soak up heat later.
It was all shower thoughts, but the idea of storing excess energy generation during peak as coldness (wording which makes pedants cringe) for evening cooling is not new. You can pre-cool your entire house, but having to wear a sweater for a couple hours every day during the summer is not fun after a while.
Freezing also introduces a lot of expansion stresses.
Having a giant ice cube in the basement would be cool, though.
10kWp needs around 60m2 of roof space, and that's gonna be more than enough power to cool a single floor house four or five times that size. Multi floor buildings are more efficient, as there are less exterior walls/ceilings.
Except what happens is not actual cooling, but spending energy to move the heat outside. In total we heat the planet, when we try to cool a small part down.
But what about panels over empty land? Claude with my simpleton prompting says panels win over grass -- grass albedo isn't that high (0.25 vs 0.07), panels offset it with generated electricity (unless your grid is wind/hydro/no waste heat source).
That said I think in practice it is pretty close to free. My own numbers show that if the Earth’s landmass were entirely grass, and we entirely replaced it with solar panels, it would be equivalent to roughly doubling atmospheric CO2 in terms of retained heat energy. Powering all of civilization would require less, and only be a few hundredths of watts per square meter.
Well, you are using that power for your own comfort when it could be used to reduce carbon emissions. Instead of cooling your house, you could allow that power to be used by industry, say an aluminum smelter, and thereby reduce the amount of carbon being burnt to power that smelter. Or, instead of running your AC, that power could be sent to the factory that builds solar panels... making new panels cheaper for everyone.
If producing solar panels is profitable, then your purchase gives the manufacturer capital it can use to expand production. And your purchase demonstrates demand, encouraging expansion of production and attracting investment in expansion.
You also become part of economies of scale that reduce the cost for everyone.
And the industry also gains experience from your purchase, pushing it further along the experience curve[1] which tends to increase efficiency.
Buying solar panels to power your air conditioner probably improves the situation for everyone by some small amount. Of course, it's possible that some other choice might improve the situation more, but most people aren't going to feel guilt as long as they haven't made the situation worse.
---
If we're talking about winter heating in for example German climates, you'd have to over build so much solar capacity to handle cold winters (even if we're willing to turn on fuel burning supplementary heaters for a statistically single digit number of days per year) that you have severe excess in summer.
You'd turn most factories (those with decently high electricity draw per capital expenditure) off for about a month worth of days in winter, usually combining this with scheduled maintenance, and:
To further shift large amounts of country-scale power consumption from year-round/winter to only sunny summer days: try to set up electrical refining of (purified) iron ore into (metallic) iron at larger solar farms where you could recycle their grid connection and their AC/DC converters together with already normalized (for there) local batteries (to shift delivery of half the production from around noon to the night, which reduces the AC/DC converter capacity and the AC grid connection/transformer size requirements). There are realistic low-temperature electrolysis options for electro-chemically splitting the oxygen off of the iron (purified iron ore is iron oxide) like Siderwin [0] and even near-room-temperature options like Electra [1].
For scale, [0] claims replacing the production of the existing coal-based iron ore refining plants of Europe with Siderwin would require 471 TWh/year in additional electricity in the grid; [2] claims the EU (without the overall insignificant island Malta, as it's not connected to the continental grid in a way that'd be readily suited to inclusion in these common-measurement-definitions statistics) produced 2497.3 TWh in the entire year of 2024, of which 237.6 TWh were solar.
Thus, it's overall "not that much" extra production capacity that'd even be spent on this (assuming an essentially-fully renewable grid, let's call it 99% by yearly aggregate electricity production): during heavy solar days, you'd import solar power from the grid at night, aim for something like 5~30% yearly average utilization of the electricity-guzzling iron-deoxidizer, and feed excess summer production into the iron-deoxidizer day and night via the local batteries. There's an overlap zone in grid daily solar yield, where you'd even import from other solar plants, depending on how expensive a kW of iron-deoxider capacity is vs. a kW of solar panel (with MPPT, but without AC inverter). Expensive solar and AC-grid means you'd import more on very sunny days (use the other solar plants instead of just discarding their power, as you'd run with no exported AC during daytime and suck the other plant's batteries dry at night, with more local iron-deoxider kW than day-average peak solar kW; local batteries do the flattening of the local solar power across the day/night cycle), while expensive iron-deoxidizer kW force you to run it more continuously across the summer months with only brief exceptions during some sporadic weeks of clouds (where you want big industry to just power through instead) despite this wasting solar power during summer daytime that can't be stored in the local batteries and can't be exported to the grid (on such peak sunny days, the plant would have to throttle the panels during the afternoon once it's local batteries hit 100% around noon).
It's around 1/6th of current yearly electricity production; at say 33% duty cycle however it'd approximately look like 1/2 of electricity production (during the third of the year where solar is in excess), so we'd be talking about a massive grid capacity impact. Hence the proposal/suggestion to move it to the DC side of solar farms that are large enough to warrant the staffing/infrastructure overhead.
[0]: https://industrydecarbonization.com/news/making-steel-with-e... [1]: https://www.electra.earth/our-technology/ [2]: https://eepublicdownloads.blob.core.windows.net/public-cdn-c...