Hydro Plant Optimization – Major Gains from Non-Invasive Technology
CLEAN CURRENTS 2026
Time: 11:00 AM - 12:00 PM
Day: 9/23/2026
Room Number: CC Central: Intelligence Theater
New: Add Sessions to Your Calendar
Interested in a session? Click “Add to My Calendar” to save it and stay on schedule. Available on mobile and desktop. Calendar Note: The event may download as an .ics file first. On mobile, tap the downloaded file to open it in your calendar app. On desktop, open the file to add the event to your calendar.
Eagle Creek Renewable Energy has installed a new technology -- the "Optimizer" (AxHydro-Opt) -- at its flagship facilities (totaling about 200 MW in capacity) and is seeing impressive results -- additional energy yields of as much as 15% and efficiency gains of up to 13%, depending on daily generation level!
During this special presentation, you'll gain insight into how the technology works and how Eagle Creek is benefitting from its application.
Background - What Is the Opportunity?
Getting more from the existing hydropower fleet is an immediate and highly attractive proposition-one that requires no new dams, only smarter operation of existing facilities.
The opportunity is substantial: with roughly 1,283 GW of conventional hydropower and 189 GW of pumped storage installed worldwide, a 5% efficiency improvement across the global fleet would avoid an estimated 110 Mt of CO2 annually.
What Are the Challenges?
Capturing this potential requires solving the dispatch problem as it actually is, rather than as a convenient approximation. The physics are unforgiving. Turbine power depends nonlinearly on flow, hydraulic head, and efficiency surfaces that shift with operating condition and wear; reservoir levels couple to flow decisions at every timestep; and, in cascaded systems, water released upstream alters head and inflow at all downstream stations. Unit commitment adds integer-valued on/off decisions and realistic operating patterns. Common simplifications-linearizing efficiency, ignoring head variation, or optimizing one dam at a time-leave meaningful energy and flexibility on the table, systematically underusing the very assets they aim to schedule.
Details about a Solution
AxHydro-Opt is a nonlinear optimization framework that retains the full nonlinear physics across the entire cascade and time horizon while honoring flow bounds, reservoir limits, environmental minimum-flow requirements, unit-commitment logic, and water-balance dynamics. The solver returns dispatch-ready schedules in minutes on standard hardware and scales through spatial and rolling-horizon temporal decomposition to fleet-size systems.
It is paired with a continuously calibrated physics-based digital twin-an active-learning layer that detects efficiency drift and recalibrates model parameters as equipment ages-and orchestrated through an agentic AI layer that translates results into explainable, dispatch-ready guidance while keeping operators in control.
Results To Date
The Optimizer has been applied to a 200-MW multi-station cascade. Multi-year historical validation showed additional energy yields of 14.8% and 7.0% in successive years; schedules running in production since late 2025 have delivered consistent efficiency gains of 5 to 13%, depending on daily generation level.
What's Next?
This integrated framework-optimization, learning, and agentic AI working in concert-provides the foundation for a growing set of capabilities. The optimizer could be extended to include hydro-plus-battery hybrid optimization, automatic generation control, uncertainty quantification, and life-extension-aware dispatch that balances energy production against asset longevity.
