But when the percentage of wind power becomes so large that you need more spinning reserve you need to build a MW of reserve for each MW of wind power and keep is spinning wich is very bad for the economy. You can work around this if you have large loads that can be shed quickly like hydrogen production or heating of houses with a large thermal mass.

Nuclear net load rejection: (I have have gotten it right, I have not worked with it. )
When the load disappears the turbine runs like a vacuum cleaner with a clogged pipe, it overspeeds due to lack of resistance to its movement. If you quickly close the steam valve to the turbine you get a preassure spike in the steam lines due to the living energy in the steams movement, as in the "thud" noice in a water pipe when you quickly close av faucet.  You have to close the valve and get rid of the steam so you dump it directly to the condenser thru another valve.  But the condenser is built to get rid of the left over heat and condence the steam to water after the turbine has taken all that it can so this is marginal for the condensers capacity. The preassure in the condencer rises and this is not good for the turbine. You quickly have to get the reactor to deliver less steam. Closing the valves helps partly since it gives a higher preassure and with a higher preassure you get less boiling and more of the energy stays in the hot water. Then you need to get the nuclear reacton from adding so much energy to the water. In a BWR you slow down the recirculation pumps that forcing water into the reactor core. This gives a smaller percentage of liquid water in the core, less moderation, neutrons dissipate out of the core and the power output goes down. Phew!

I do not know how it is done in a PWR. In a BWR the control rods are mostly used to form the active region in the core and I think it is the same with a PWR and the power output is regulated by adding a neutron absorbent to the water, boron.  Inserting the control rods untill power output falls and then withdrawing them about as much must be the quickest to do. Whitehall, is this correct?

This is obviously a tricky procedure where all parts of the control system has to do the right things within fractions of a second or you get overspeed on the generator or overpreassure somewhere in the system and then an automatic shutdown, it also shuts down if you overcompensate.

There has been two large grid failures in southern Sweden since I started to learn about different kinds of engineering.  During the latest the nuclear reactors had about 75% success with the net rejection.

Another problem is that you cant run the reactor on two low power our you will get an accumulation of neutron absorbing substances that hinders a power increase even if you  withdraw the control rods. Then you have to shut down and wait for them to deacy naturally, you get the same need for a pause after an emergency shutdown.  So a nuclear powerplant leaving the grid can not come back immediately.