Yesterday we released the latest upgrade to our revenue forecast for battery energy storage in Great Britain, version 3.3.
Our latest upgrade adds:
- intraday prices into our fundamental model; and
- an entirely new modelling approach to enable realistic optimisation between all revenue streams.
You can read our full forecasting methodology here.
This upgrade, on top of updated commodities prices, updated timelines for the retirement of gas power-generation, and improved capacity expansion modelling means you can model BESS revenues with increased clarity and confidence.
The upgraded revenue forecast
Every quarter we update our model with refreshed commodity prices and adjusted capacity projections; however, we’re also always improving our forecasting models themselves.
Creating intraday prices
The Day-Ahead price of electricity is informed by forecasts for demand, wind and solar generation, and plant availability. These forecasts are critical to building the supply stack to meet demand from one half-hour to the next -and setting prices.
As time progresses from the closure of the day-ahead market to delivery, forecasts for each of these variables evolve into the outturn values. Power can be bought and sold in response to these changing forecasts, up until 20 minutes before the delivery window starts.
While the Day-Ahead price and Intraday price are highly correlated, the evolution of these forecasts (along with human behaviours) causes intraday prices to diverge from their day-ahead reference point. These price changes represent an opportunity for storage to optimize against.
- More information on how we model demand, wind and solar generation forecast errors is here.
- Information on modelling plant outages is here.
- Information on how we model 'other' things that can change intraday prices, like human behaviour, is here.
Modo's intraday forecast represents the EPEX Continuous Intraday market reference price of products traded at half-hourly granularity, or RPD HH. The actual data of EPEX RPD HH is available via our API here.
A new price forecast means a new revenue stream for the dispatch model to optimize against...
Re-building dispatch
The intraday continuous market is continuous: it's an order book in which trades can be placed anything from ~18 hours out from delivery to 20 minutes beforehand (with most volume traded ~2 hours from delivery). In the real world, trading decisions made, say, 4 hours out will impact decisions 3, 2, or 1 hour out, with factors like liquidity an issue. It's often a very fluid picture. The reality is quite different to a 24-hour perfect foresight model.
To try and mimic continuous intraday trading, we perform multiple optimisations.
- First, we lock in the day-ahead position using the wholesale day-ahead price forecast and any ancillary contracts using the Dynamic Frequency Response price forecasts. These can introduce constraints on parameters like the battery's state of charge, which are respected in subsequent optimisations.
- This day-ahead position is then re-optimised in many steps across the delivery day. Multiple optimisations are run in increments of 2h, and the price signal for the rest of the day is adjusted to consider the intraday price for the first 2h of the new optimisation window.
This optimisation ends up with a revenue strategy that looks like this:
Intraday optimization doesn’t always provide uplift to the revenue forecast, and depends on the other prices available to the model. You can find more examples in our documentation, here.
Accessing Modo Energy forecasts
You can interact with revenues forecasts on the Terminal in two key ways. And as always, the best tool depends on the job you’re trying to get done:
To inform high-level strategies - a library of pre-populated revenue forecasts
There are more than 3,000 pre-populated revenue forecasts for battery energy storage in Great Britain our Data & Forecasts explorer. Our aim is that access to these forecasts enable sensitivity analysis and a quick answers to questions like: “How does duration impact long-term revenue opportunities?”, or “Which zone has the most significant long-term opportunities?”.
Alongside the library all of this data can be trended on our Plotter:
To value projects and existing assets - creating your own revenue forecast
The library of pre-populated forecasts is powerful when looking at a wide range of options for developing projects. However, if you’re looking at a specific project or existing asset, then you want a revenue forecast that has inputs matching the specific characteristics you’re looking at.
And so we’ve exposed our forecasting model, making it possible for you to create highly specific forecasts for battery energy storage in Great Britain from start to finish in less than 10 minutes.
