One workspace for AI-integrated energy management: a 17-stage technology pipeline, a data guide for every engineering problem, and a lab that runs on your own data.
Open a stage, then any technology inside it. Your progress is saved in this browser.
Start with Your data to solve a real problem with your own CSV or live weather. Explore and Advanced EMS are teaching models; Economic Dispatch and Net-Load use illustrative numbers, while Real Data uses actual yearly Ember figures.
Three thermal units with quadratic cost C = a + bP + cP². The solver finds the single marginal cost λ at which total output equals demand, holding units at their limits. Default data is the textbook example from Wood & Wollenberg (Example 3A); expected result at 850 MW is about 393 / 335 / 122 MW and λ ≈ 9.148 per MWh.
Baseline = share demand in proportion to each unit's operating range (a simple non-optimal rule). The saving is what equal-λ dispatch gains over it.
Net load = demand − solar − wind. Add renewable capacity and watch midday net load fall and the evening ramp steepen. The red area is energy that cannot be absorbed because thermal must-run units cannot go lower.
Illustrative synthetic shapes (not measured data): double-hump demand, bell-shaped solar, gently varying wind. Ramp = change in net load over 3 hours divided by 3. The "India 2025" preset loads real Ember solar and wind capacity; peak demand there is an assumption (see Real Data tab).
Actual yearly data for any country or region, 2000 to 2025. Pick a place and a metric; KPIs show the latest year. Optionally layers in plant-level capacity from the WRI Global Power Plant Database when that file is connected.
Data source: loading…
Net-Load tool needs a peak demand, which Ember does not give. It is assumed as 1.15 × average demand (demand ÷ 8760 h); must-run floor = 35% of average demand and ramp capability = 15% of peak per hour are also assumptions.
Source: Ember yearly electricity data. Capacity is not reported for every source and year; gaps are skipped. Plant columns (GPPD) appear only if a plant file is configured.
Rank every country by a metric. Small systems are filtered out by a minimum generation so the list stays meaningful. Click a row to open that country in the Real Data tab.