Lots of good points. To go completely renewable with no baseload (coal, nuclear, natural gas) would require massive storage like JRP mentioned. And unless you want to make EVERY end point power user be able to both generate and store all the power they would ever need, there will have to be at a minimum a local massive storage facility or a local/regional baseload power supplier with some kind of grid connections still in place.
The way I think things might play out, would be as follows. Obviously wind and solar are getting installed at ever increasing rates.
The wind is falling out of wind's sails in the US. "
Wind capacity additions (1,032 MW) dropped sharply in 2013 to less than one-tenth of the capacity added in 2012.". And while solar was up sharply, continuing the growth trend, just more than 50% of new utility-scale generation burns natural gas. (This doesn't account for small-scale PV, which was also big.) What changed with wind? The big issue was the end of investment tax credits, but also (a) the best sites are used or have over-crowded transmission and (b) increasing public resistance to more on-shore wind.
This cuts into the baseload power suppliers revenue, and also makes it tougher for the maintaining of the grid. Someone from England suggested that the running of the grid be separated from the generation of the power. I guess that was done there. That seems to me like a good idea. The grid would be nationalized, and run for the benefit of all connected to it. Then the power companies can compete against each other concerning who can supply power when and at what cost. They would also be competing against the individual homes that supply power via their solar panels. This would be a major change, no doubt.
What you describe is almost exactly what we have in much of the United States: from Maine to Maryland to Ohio, and Texas. The grid is owned and maintained by transmission companies that own no generation, and the grid is controlled by Regional Transmission Operators that are independent and accountable only to the Federal Energy Regulatory Commission. California and the Midwest also have RTOs, but they lack the full separation of transmission from generation. You'll notice all the fighting about roof-top solar has been in areas that still have the old vertically-integrated utility model.
From strictly a "lets emit as little CO2 as possible" point of view, I think something like this might happen: solar and wind keep ramping up, lessening the need for baseload power. Especially once personal or regional energy storage come into play over time as costs for that decrease. As less baseload power is required, they could start turning off the more polluting plants first (i.e. coal). Any specific plant wouldn't be shut down until the surrounding community that uses it's power can reliably get all required power from other sources or stored power from renewables.
You are implicitly assuming a carbon tax, which I agree with but will be very controversial to implement. The current rule used by RTOs for dispatch is "least cost," so units with the lowest offer price are dispatched first. What *is* hitting coal units, even under this system and without carbon pricing (or with low carbon pricing in RGGI states) is how inflexible they are: typically these units need 16+ hours to start up, so once you fire them up, you leave them on-line. When prices fall low, these coal plants can ramp down to ~30% of their maximum output, but they're still losing money (cost > market price). When supplemental power is only needed in a few hours, these coal plants will likely shut down for economics.
Since baseload will still be required for a long time, the choice would be nuclear. One problem I could see doing this would be that should renewable and storage come down so fast that everyone is doing it, you could end up with a very expensive nuclear plant that runs for 10 years and then can't compete with the renewable/storage suppliers. Who then eats the cost if the plant is no longer economically feasible?
The only new nukes being built are by vertically integrated utilities (in the US, Southern Company). Their ratepayers will continue to pay these costs until the whole cost is recovered (with interest). For example, Long Islanders are still paying charges related to the abandoned Shoreham nuke, even though it never generated one kWh.
Sunshine is the most diffuse of all sources of renewable power. The energy density of wind is far higher, and moving water (ocean wave, tidal, etc.) higher still. These other "fuels" have the advantage, too, of being more predictable and higher availability than solar. So, while I certainly see an important role for solar in the future, any 100% renewable scenario will almost surely have a mix of solar, wind, biomass, and marine. Having this diversity of generation sources will reduce (but not eliminate) the need for storage.
The cheapest form of storage is hydro--which is a component of nearly every region's power supply. The challenge is integrating the dispatch of hydro with the need to balance renewables, while at the same time respecting the water-release requirements to keep riverine environments healthy.