Dispatch Strategies

How re-optimisation enhances revenue

The two-stage optimisation process allows the model to layer additional revenue streams on top of a baseline energy arbitrage strategy.

Example: Capturing a real-time price spike

Imagine the initial dispatch plan schedules the battery to fully discharge at 7 PM to capture a forecast evening peak price. However, at 6 PM, a new short-term forecast predicts an even higher price spike at 7:30 PM due to an unexpected power plant trip.

The re-optimisation process will identify this new opportunity. It will adjust the plan to hold the energy in the battery for an extra 30 minutes and discharge at the higher price, significantly increasing revenue.

When opportunities can be missed

Conversely, the initial dispatch plan can sometimes constrain the battery, preventing it from capturing later opportunities.

For example, on a day with high solar generation, the initial plan might schedule the battery to fully charge during the middle of the day to take advantage of low prices. If an unexpected network event causes prices to fall even further (potentially to the negative price floor) later in the afternoon, the battery may be unable to take advantage. Because it is already full based on the initial plan, it has no remaining 'footroom' to charge more and get paid for it. Our modeling approach accounts for these real-world operational limitations.

Strategic capacity reservation

To mitigate the risk of missing out on short-term opportunities, the dispatch model can be configured to strategically reserve a portion of the battery's power (MW) and energy (MWh) capacity.

This reserved capacity is ring-fenced during the initial dispatch plan and is only made available during the real-time re-optimisation stage. This strategy ensures the battery maintains the flexibility to capture high-value, volatile opportunities that often only appear close to the dispatch time, such as extreme price spikes or lucrative FCAS events. This is a key lever in balancing a reliable baseline revenue stream with the potential for high-upside from market volatility.

Optimising FCAS stacking

The model's sophistication extends to its handling of ancillary services. Not all eight FCAS services can be provided at the same time from the same battery capacity. The model incorporates AEMO's rules on which services can be 'stacked' together.

It intelligently assesses the prices across all eight FCAS markets and finds the most profitable, permissible combination of services to provide in each 5-minute interval, ensuring it maximises revenue from the battery's available capacity for grid services.

Key modelling assumptions

The revenue forecast is underpinned by a set of core assumptions that reflect real-world trading conditions and battery operation.

• Two-stage optimisation: The model generates an initial day-ahead dispatch plan which is then refined through a series of rolling real-time re-optimisations.
• Capacity reservation: A configurable portion of the battery's MW and MWh capacity can be reserved specifically for trading in the more volatile real-time window.
• Market saturation: To ensure realistic outcomes, the volume of FCAS the battery can provide is limited by market saturation parameters, reflecting a feasible market share.
• Price foresight: The model assumes access to accurate price forecasts ahead of each optimisation step (day-ahead and real-time). The quality of these forecasts is a key driver of model performance. You can find out more about our price forecasting methodology [here].
• Transaction costs: A minimum price spread is required for the model to execute an arbitrage cycle. This prevents the model from attempting to capture unrealistically small profits and simulates implicit transaction costs.