Storage

Storage technologies such as batteries, pumped hydro, and coordinated consumer energy resources are modelled as flexible assets that can shift energy across time. The model reflects both their physical capabilities and economic behaviour within the NEM.

Capacity and Technology Types

Each storage unit is defined by its power capacity (MW) and duration (in hours), as well as ramp rates, round-trip efficiency, and bidding behaviour.

Energy Limits and Cycling

For each storage plant in the model, the state of charge is tracked over time and constrained to stay between zero and its maximum energy capacity.

BESS lead times

Battery energy storage investment in the CEM uses a uniform 3-year lead time across all durations (2h, 4h, and 8h). This lead time is sourced from the AEMO ISP and represents the development and approval period before construction can begin. The CEM takes the maximum lead time across all BESS durations and applies it uniformly.

Operational Limits

Ramp rate limits are applied based on technology type to ensure ramping is consistent with real world behaviour. Even though BESS can ramp effectively instantaneously, the fleet ramp rate is restricted so it takes 10 minutes for the full fleet to ramp to maximum discharge. This captures the fact that, even though an individual battery can ramp fully from one settlement period to the next, this rarely happens for the whole fleet simultaneously.

Economic Behaviour

Storage dispatch is cross-optimised with the rest of the model. This means that it will dispatch in whichever way minimises total system cost, subject to it meeting its minimum cost of cycling or SRMC, whilst factoring in economic withholding of capacity.

For batteries, these SRMCs are derived from capital cost assumptions (e.g. cost per MWh of storage). Battery energy storage systems also have S-curves applied to their SRMCs based on historical battery bidding behaviour, to reflect diverse bidding patterns. This means that some battery capacity will take a low spread, as long as it still pays back this energy capex (capex/MWh spread across an assumed 8,000 cycle lifetime). Meanwhile, some battery capacity will be withheld to wait for a higher spread, reflecting opportunity cost.

As battery penetration grows, batteries will increasingly compete with each other and set the price more often. The S-curve approach captures this dynamic naturally: as more battery capacity enters the market, the aggregate supply curve flattens and price spreads compress. However, the S-curves used are calibrated on a period when batteries were largely price-takers. As the fleet grows and batteries become marginal price-setters more frequently, the model accounts for this through the CEM feedback loop - lower spreads reduce the economic case for new BESS builds, which in turn slows the buildout rate until an equilibrium is reached.

A low fixed cycling cost is used for pumped hydro based on public sources. However, it also has a round trip efficiency of 75% which means that it has to discharge at 33% higher price. For example, if it wants to discharge 75 MWh then it has to charge by 100 MWh as it loses 25%, so discharge price must be at least the charge price * 100 / 75. Some pumped hydro capacity is also withheld to wait for higher spreads to reflect opportunity cost.

Near term battery storage buildout

Near term battery storage buildout is modelled using a probabilistic view on which projects currently in the pipeline will be built, and when. Battery projects in the pipeline are separated based on project status, with an informed view on likely delays and total capacity to come online within each status group based on historical project timelines in the NEM and market expertise. These are the assumptions for the central scenario, which represents the most realistic worldview.