Fifteen-minute resolution
A demand value for every quarter-hour, across the full day and every configured level of the portfolio.
Load forecasting
Load forecasts for the full portfolio, at 15-minute resolution.
The closer a balance group's forecast tracks its actual load, the smaller its imbalance and the lower its cost to settle it, short or long.
A balance group's exposure is determined by the accuracy of its demand forecast. Underestimate load, and the group is short precisely when the system is short and imbalance is most expensive; overestimate it, and the resulting long position earns little when the system is already long. Standard load profiles overlook the specific drivers of a portfolio's demand: its customer base, its connected assets, and the weather across its region.
How it works
Orvio trains a dedicated machine-learning model on the portfolio's metered history and the weather that drives its load, then forecasts demand for every quarter-hour across the portfolio. The forecast is constructed bottom-up at each configured portfolio level, so the figures reconcile from a single meter group up to the entire balance group.
The model retrains daily on the latest readings, and every run is scored against recorded meter data, so accuracy is measured rather than assumed.
Capabilities
Trained on metered data
The portfolio's metered history and regional weather, not a generic load profile.
Updated daily
The model retrains on the latest readings, tracking actual consumption as it shifts across seasons.
Structure-aware aggregation
The forecast aggregates from meter group to balance group without losing the underlying detail.
Forecasts are delivered on the platform and via the API, with the same timestamping and scoring as the system forecasts, so every run remains fully auditable.
What changes in practice
For a balance group, demand-forecast accuracy is the difference between purchasing short into an expensive system and holding a balanced position when it matters. Orvio provides schedulers with a forecast they can defend, built on the portfolio's own data, alongside a daily score that confirms how it performed.
See what flexibility could earn.
Book a demo and see the signal on live data, applied to real assets.
Pricing scales with the portfolio. Book a demo for a quote.
FAQ
What data do you need from us?
The portfolio's metered consumption history, uploaded directly into Orvio as CSV, Excel, Parquet, or ZIP; Orvio detects the schema, the column mapping and aggregation level are confirmed, and the files are consolidated into a single dataset from which the first forecast is produced. Weather is sourced automatically, so only the metering needs to come from the customer.
How long until the first forecast?
Onboarding takes a matter of days once a clean metered history is available. The model improves as it retrains daily on new readings.
At what level do you forecast?
At each configured portfolio level, from a single meter group up to the entire balance group. The forecast is built bottom-up, so the levels remain consistent.
How is accuracy measured?
Every run is scored against recorded meter data, with the same timestamping as the system forecasts, so forecast performance is observable rather than taken on trust.