Decision making in the energy sector with innovative energy system models
Welcome to Greenville Analytics. Our mission is to revolutionise energy system modelling with AI-based speed, accuracy and simplicity, because the energy transition runs on the models we can afford to build. Explore the technology, and get in touch if you want to know more.
The model has become the constraint.
Runs take days, not hours
Compute sets the planning cycle.
Detail is traded for tractability
Hours and regions averaged away.
Workflows resist everyday use
A handful of scenarios, not hundreds.
Energy systems, AI and algorithms.
A TU Delft spin-off — passionate about the energy transition and ambitious to innovate, with roots in optimisation under uncertainty, machine learning and high-performance solvers.
Alongside the core team, a group of TU Delft student developers builds and tests the platform with us — keeping the work close to the research it comes from.
Different products and services based on our core technology.
Short-term forecasts. Day-ahead and intraday dispatch at full hourly detail — prices, flows and storage behaviour for the horizon operators actually trade on.
Long-term simulation. Capacity expansion and system pathways across weather years, fuel prices and build limits, with uncertainty swept rather than assumed.
Expert services. When things get complicated, Greenville offers on-demand model customizations and consultancy services.
Five kinds of decision, one model.
TSOs & grid operators
Placeholder — grid studies take weeks while the network keeps changing.
Adequacy and congestion, re-run as the grid changes.
Energy traders
Placeholder — price signals move faster than conventional models can run.
Portfolio and hedging views from the same physics.
Renewable developers
Placeholder — capture prices are uncertain and drive the investment case.
Siting and capture-price curves per weather year.
Policy & regulators
Placeholder — decisions need assumptions that can be scrutinised.
Transparent pathways with assumptions on record.
Industrial buyers
Placeholder — electrification plans hinge on future power prices.
Electrification cases against real system prices.
Active in the scientific community.
We are active in the scientific community, supervising and collaborating on graduate research at TU Delft. Find here a selection of relevant recent work.
Economic dispatch with no fitted parameters reconstructs Dutch day-ahead prices to 22.7–24.5 EUR/MWh RMSE, beating a 44-parameter calibrated merit-order stack; endogenous cross-border trade is the decisive structural lever.
WHY IT MATTERSPlaceholder — more reliable price forecasts for trading and hedging decisions.2026Exploring the trade-offs among flexibility options in hydrogen productionNour El Din, Verzijlbergh & De Vries · Frontiers in Energy ResearchA power system model coupled to a grid-connected electrolyser and storage, with a new method for mapping near-optimal regions — showing how flexible hydrogen consumption and storage investment complement one another.
WHY IT MATTERSPlaceholder — helps hydrogen project developers size electrolysers and storage with confidence.SEE MORE RESEARCH →Built in Europe, run in Europe.
Model code, training data and compute stay on European infrastructure. The critical infrastructure of the energy transition should not depend on software you cannot inspect or hosting you cannot reach.
