Anyone know the efficiency? Batteries are in the high 90s, but are so expensive that they don't make sense for stockpiling. (Charge in the summer when days are long, discharge in the winter when everyone needs to turn on the heat.)
That pilot project claimed less than 25% efficiency, but the authors seemed confident efficiency could be raised. A 5 megawatt demonstration plant has been running in the UK since 2018:
Efficiency needs to be viewed with target storage times in mind. For example, the reason batteries are not used for seasonal storage is not just because of their cost, but the fact that they lose their charge pretty quickly. They are 90%+ efficient in the short term, but that drops pretty quickly.
On the other hand, something like pumped hydro is less efficient and takes longer to bring online, but can store the energy for months or even years.
For anyone wondering how long that is, the largest pumped hydro plant in the UK can be at full capacity in 16 seconds from standby, or 75 seconds from cold.
Note that "standby" is burning power to keep the turbines spinning without flow, so that they don't have to accelerate when the stopcocks are opened. Dinorwig's purpose is smoothing demand spikes, so its 75% roundtrip efficiency is acceptable and its duty cycle somewhat predictable, but it must be a tricky optimization problem to maximize efficiency while never failing to meet the load.
(Interesting side note - when I took a tour of Dinorwig some years back, the guide said that when the facility was built it rarely needed to run more than 1 turbine - now they routinely fire up all 6. At some point it will run out of capacity...)
Efficiency is only one measure of a storage system's overall usefulness.
Other dimensions are total capacity, total cost, ramp-up / ramp-down times, storage and delivery rates, and various elements of technical complexity.
Batteries offer moderate-scale, fairly-expensive, moderate density, and relatively low storage and delivery rate, energy storage. They're exceptionally useful for mobile uses (from handheld to vehicular), but may not be as desirable for stationary systems.
There's a spectrum of energy storage options, ranging from capacitive storage (very responsive, but also expensive, inefficient, high-rate but low-yield.), flywheel, battery, physical mechanisms (pumped hydro, compressed / liquified air energy storage), and chemical (electrolysis, fuel cell, fuel synthesis).
If you're storing energy when it's abundant and releasing it when it's highly scarce, net efficiency doesn't matter nearly as much as matching generation and load capacity. Losing energy because you don't have the capacity or capture rates means a theoretically highly-efficient process is actually lower in net efficiency.
The duration of energy storage, ranging from a few seconds (capacitors) or minutes (flywheel) to proven-at-hundreds-of-million-years (hydrocarbon chemical energy storage) also matters. If you're trying to smooth out transients, capacitors or flywheels (mimicking "spinning reserve" or inertia) are your tool. If you're shifting over a period of hours, batteries, CAES, and pumped hydro come into play, with the latter offering long-term capacity of months if needed. Fuel synthesis could in theory allow shifting by years or decades, with very stable storage, though a net round-trip efficiency of ~15-20% (losses in synthesis compounding Carnot efficiencies of thermal energy generation, if used for electrical generation).
Anyone know the efficiency? Batteries are in the high 90s, but are so expensive that they don't make sense for stockpiling. (Charge in the summer when days are long, discharge in the winter when everyone needs to turn on the heat.)