Power for baseload has a 99.99% uptime requirement.
The batteries needed to make variable power sources even 95% capable of meeting demand are economically unfeasible as well.
As a couple examples.... The laegest solar plant in Arizona produces only 10% of its peak summer power in the winter. ...similarily, in Texas the wind turbines are usually not turning during Texas peak demand because it's during high pressure systems where there's not wind.
the problem is distribution the power over long distances.
and there is no grid available that spans the whole world. that would take a huge amount of work for all the countries, but countries do not think global, so it would be a pita to get everybody on the same feets.
So "baseload" isn't some expecial part of the electricity. It is an accounting number. The challenge is to provide the correct power to the grid and for that, different technologies are used. Coal and nuclear power plants have a reasonably high availability, but as others have written, far away from 99% uptime. Their downside is, that their output can only be regulated very slowly. Both properties made them ideal for the amount of power which is called baseload. The problem is, they can't deliver much else due to their limitations. That is also the reason, that for decades the architects of the power systems tried to drive the baseload up. To increase demand at night, Belgium installed floodlights on their freeways, Germany for a while pushed electric heaters which would draw their main power at night at a reduced price.
Solar and wind behave inverse, they might not be as reliably available, indeed the sun doesn't shine at night, but they can be switched off in an instant. But any time they produce enough electricity, they can carry the baseload. So currently in Germany wind easily provides large parts of the baseload (https://www.electricitymap.org). Coal plants are almost shut down, even solar right now produces as much energy as coal.
So without coal and nuclear, the challenge still is to match the grid demand. This is independant from the baseload number. And with the shift to reneweables, the baseload number can be driven down quite considerably, if incentives are given to use electricity in times of high supply.
It probably will be supplemented by natural gas, coal, oil for the foreseeable future. These source have much less capital requirements and much better understood than novel nuclear energy; in a world where renewable energy just gets cheaper and cheaper the big capital spend of nuclear just does not make any sense.
Nuclear also requires 12-18 month cycles of refuelling, not to mention malfunction downtimes. These requires several weeks: https://www.eia.gov/todayinenergy/detail.php?id=1490
For baseload, I'd rather vote for a compact Tokamaks: https://www.tokamakenergy.co.uk/st40-has-been-rebuilt-and-su...
https://skepticalscience.com/print.php?r=374
Summary
Arguments that renewable energy isn't up to the task because "the Sun doesn't shine at night and the wind doesn't blow all the time" are overly simplistic.
There are a number of renewable energy technologies which can supply baseload power. The intermittency of other sources such as wind and solar photovoltaic can be addressed by interconnecting power plants which are widely geographically distributed, and by coupling them with peak-load plants such as gas turbines fueled by biofuels or natural gas which can quickly be switched on to fill in gaps of low wind or solar production. Numerous regional and global case studies – some incorporating modeling to demonstrate their feasibility – have provided plausible plans to meet 100% of energy demand with renewable sources.
Daily insolation in Arizona varies by less than a factor of 2 from June to December, so I call BS on this statement.
Assume the worst case of only solar- that baseload also continues during the day, so about 30% of it would be powered by solar. That leaves 28% of average load that would need to be supplied overnight, or 3 TWh for the entire US.
The current installed cost for grid scale batteries is <$500[1]- $483/kWh for the 129 MWh battery in Australia. Batteries in cars are <$200 at the pack level. At $500/kWh that's 1.5 trillion for the entire US. Well within manageable for several years- far less than has been spent on Iraq, for example.
There are many reasons to believe that batteries could easily last for tens of thousands of cycles if properly specified[2], but current batteries are specified to ~5000 cycles to 80% degradation. Adding 20% capacity every 5000 cycles would cost 6 cents per kWh of system throughput.
By comparison, the EIA says that Nuclear costs 8 cents per kWh today[3]. By current measures battery storage is cheaper than nuclear (not counting initial generation), and that's not even counting the massive reductions in capacity that would be enabled by other generation and HVDC transmission.
> As a couple examples.... The laegest solar plant in Arizona produces only 10% of its peak summer power in the winter.
That's a lie. The lowest winter output is 37% of peak summer output[4]. The variation is HIGHER in the summer than the winter; +55% in summer and -45% in winter. Power consumption in Arizona also varies at the same time (because that's how the plant was designed to produce power) by 25-30%, so the majority of that variation is simply proportional to normal load.
[1]: https://www.theverge.com/2019/6/25/18715585/tesla-australia-...
[2]: http://jes.ecsdl.org/content/166/13/A3031
[3]: https://www.eia.gov/outlooks/aeo/pdf/electricity_generation....