Oops, don’t let a ‘greenie’ read this — they will try to do it.
Too late, it's already being done. Well, not exactly like what you describe, but close to it. It works best with "low head turbines" coupled to motor/generators. A dam is built across the mouth of a bay, impounding a large storage volume. During times of low power demand, excess is used to drive the turbines as pumps, raising the volume of water behind the dam. As demand for power increases water is released back through the turbines, driving the generators and making more power available to the grid.
One advantage a "low head" system has is that if tidal forces coincide with the power demand cycle (high tide/low demand and vice versa) they may be used to add to the amount of water moved into and out of storage.
There have been "high head" systems tried as well but the greater the height of the storage reservoir above the source the lower the overall efficiency of the process.
That is, converting electric power (current x voltage) to mechanical power (torque x speed) involves losses due to electrical resistance, magnetic hysteresis, windage, and bearing drag. Converting mechanical power to hydrodynamic power (pressure differential x flow rate) involves losses due to pump slip and bearing drag. Delivering the hydrodynamic power to the storage reservoir involves wall friction which causes pressure drop. Each of these loss factors has a coefficient of performance, the overall efficiency is the product of all of the COP's of the various stages, divided by 2 since you lose power pumping water uphill and again when you let it flow back down.
A lot of processes are reversible but there are no free lunches. If you lose 25% of the input energy going from "A" to "B", you will lose another 25% coming back to "A". The only time something like this makes sense is when your electric grid is largely powered by large, central generating stations with steam powered turbines. It doesn't matter how the steam is generated, atomic energy, geothermal, coal, oil, or bio-mass. All large central stations run most efficiently at a constant load, unfortunately consumer demand varies on a 24 hour cycle.
In the final analysis, gas turbine driven generators are simple, efficient, quick to start and can be quickly "throttled" up and down to mach a varying load. It would seem to me at the ideal grid would have 75% of it full power requirement supplied by nuclear plants with the remaining 25% by gas turbines to handle "peaking" loads. Forget about wind, solar, and tidal as they cost more then thermal based generation and none of them is available 24/7. They can only compete when the Government shovels in tons of taxpayer money.
Regards,
GtG
But "NUCLEAR" power is expensive you say...
The power isn't expensive, the investment in the infrastructure is and largely because of the long, dragged out licensing process. Every plant built has to start from ground zero, answering the same objections every time. The environmentalists have effectively stopped any development of nuclear power in this country.
What we need to do (if we had a DOE w/ balls) is design a modular plant utilizing best available technology. Nothing off the table, pebble bed reactors, thorium fuel cycle, &c, &c. Then revamp the licensing process to allow "type certification" which would permit private companies to build the modular plants where needed w/out years of fighting with the envirowheenies.
Problem solved!
G