OK, Google got CPR and is back on the job...
This is a "could we run a real world grid with only wind, solar and storage and do it for a reasonable price?". It's not an attempt to design the efficient/least expensive 100% renewable grid.
Researchers at University of Delaware used four years of weather and electricity demand/load data in one minute blocks to determine 1) if a combination of wind, solar and storage could meet 99.9% of demand and 2) the most cost effective mix of each to meet demand.
The data for 1999 through 2002 came from the PJM Interconnection, a large regional grid that services all or part of 13 states from New Jersey west to Illinois, from Pennsylvania south into Tennessee and North Carolina. This is the world's largest competitive wholesale electricity market, serving 60 million customers, and it represents one-fifth of the United States' total electric grid.
They used currently available technology and its projected price in 2030. They included no subsidies for wind and solar in their calculation. They did not include hydro, nuclear, tidal or other possible inputs. They also did not include power sales to and purchases from adjacent grids. They used three existing storage technologies - large scale batteries, hydrogen and GIV (grid integrated vehicles).
They found that by 2030 we could obtain 99.9% of our electricity from renewable energy/storage and the remainder 0.1% from fossil fuels for about what we currently pay “all-in” for electricity. The all-in price of electricity which includes coal and oil produced health costs currently paid via tax dollars and health insurance premiums.
During the four year period there were five brief periods, a total of 35 hours, when renewables plus storage were insufficient to fully power the grid and natural gas plants came into play. These were summer days when wind supply was low and demand was high. The cheapest way to cover these ~7 hour events was to use existing natural gas plants rather than to build additional storage. Adding in hydro, tidal, etc. would further reduce this number.
After 28 billion simulations using differing amount of wind, solar, storage and fossil fuels they found the best solution was to over-build wind and solar and at times simply "throw away" some of the produced power. Building "too much" wind and solar turns out to be cheaper than building more storage given the storage solutions we have at this time. Finding markets for the extra production, selling electricity to offset natural gas heating for example, further reduced costs.
Budischak, Sewell, Thomson, Mach, Veron, and Kempton Cost-minimized combinations of wind power, solar power and electrochemical storage, powering the grid up to 99.9% of the time Journal of Power Sources 225 (2013) 60-74https://docs.google.com/file/d/1NrBZJejkUTRYJv5YE__kBFuecdDL2pDTvKLyBjfCPr_8yR7eCTDhLGm8oEPo/edit
Remember, this is a "worst case" study. Add in hydro, tidal, geothermal, and residual nuclear and the price drops because less storage will be needed. The same happens when there is exchange of power between grids.
And those 2030 wind and solar prices? The authors used somewhat dated (2010) projections. Current prices have already brought the cost of wind down to their projected 2030 level and solar is already cheaper than their 2030 numbers.
We almost certainly will have better/cheaper storage than they used in their study. It's coming on line now.