Storing Energy For Days Rather Than Hours

Grid storage today is dominated by systems that discharge over a few hours, which suits daily cycles. Covering a multi-day lull in wind and solar output is a different problem with different economics.

Power and energy are priced separately

A storage system has a maximum output, measured in power, and a total quantity it can deliver, measured in energy. These are supplied by different parts of the system.

In a lithium battery, both scale together because the cells provide each, so extending duration means buying proportionally more of everything.

Technologies that separate the two, storing energy in cheap tanks or reservoirs while sizing the conversion equipment independently, become cheaper as duration lengthens.

Round-trip efficiency matters less than it appears

Short-duration storage cycles daily, so losses accumulate quickly and efficiency dominates the economics.

A system that cycles a handful of times per year loses less in absolute terms, so a technology with poor efficiency but very low cost per unit of stored energy can still be the rational choice.

This inverts the usual ranking and explains why approaches dismissed as inefficient remain under serious development.

Pumped hydro remains the benchmark

Moving water between reservoirs at different elevations is by far the largest form of grid storage in operation and is well suited to long duration.

Its constraints are geographic and permitting-related rather than technical, since suitable sites are limited and projects take many years to approve and build.

Closed-loop designs that avoid rivers, and proposals using existing mines or reservoirs, are attempts to widen the set of feasible sites.

Chemical and thermal routes address different needs

Flow batteries hold energy in liquid electrolyte in tanks, so capacity scales with tank size, while thermal systems store heat in inexpensive materials and convert it back when needed.

Hydrogen produced from surplus electricity can be stored in very large quantities for very long periods, at a substantial efficiency penalty in both directions.

Each of these fits a particular duration and cycling pattern rather than competing directly with the others.

The revenue problem is harder than the technology

An asset that runs a few times a year earns little from ordinary energy arbitrage, so its value lies in reliability during rare events that markets compensate poorly.

This is the same gap that capacity mechanisms address for generation, and how long-duration storage is credited within them will determine whether any of it gets financed.