Modelling the storage fleet
Energy storage is becoming a major part of energy systems as grids incorporate more intermittent renewable generation. The chart below shows rapid deployment of grid-scale battery storage in GB, Germany, and Spain. Germany also has a significant and growing number of residential batteries.
Technologies covered
We distinguish three storage types, each with its own operational rules:
-
Pumped hydro – Large-scale assets with water reservoirs, typically ~10 hours duration. They pump when prices are low and release when prices are high.
-
Grid‑scale batteries – Large (~10MW plus) battery energy storage assets grouped by bidding zone and duration (2h, 4h, 6h, 8h). Charging and discharging slows near full/empty energy levels and cycle costs vary depending on asset and age.
-
Prosumer batteries – Home and commercial batteries behind the meter. These maintain a minimum charge for backup, and only a fraction of the fleet participates in markets at any time. The portion of the fleet participating in the markets grows over time.
Storage transitions from price taker to price setter
Today, storage mostly follows prices — charging when cheap, discharging when expensive. It is usually a price taker. By capturing these spreads in price, it lowers the overall operating cost of the electricity system. But, as the storage fleet grows it will become a price setter: smoothing peaks and troughs in price and shaping price patterns.
A naive model for storage, particularly as the fleet gets large, can result in block pricing. The low price is set by the charging of storage (close to zero), and the high price at the cycling cost of storage (usually the minimum price you tell the model it needs to discharge). While this ‘minimises the cost’ of running the system in a system-wide optimisation model, it is unrealistic.
In the real world, operators do not have perfect foresight of prices. They must choose the best time to cycle individual assets, which will have different physical limits, commercial arrangements, and warranties. Grid limits can restrict power flows. Individual operators act to maximise their profit rather than minimise overall system cost. Real markets are messier!
Realistic decision-making & imperfect co-ordination
Our goal is to let storage move prices in a realistic way, without turning it into an all‑knowing, perfectly coordinated trader. Therefore, we add a few limits to the actions of storage fleets:
- Charge-rate tapering — Storage responds best at a state of charge ~50%. Near full or empty, response is limited.
- Cycle costs — Each cycle requires a minimum price spread to justify the wear on the battery.
- Coordination limits — Not all storage participates simultaneously, especially prosumer batteries.
These constraints produce behaviour similar to high-penetration markets like California’s CAISO.