Wind turbines on hilltops capture far less power than speed-up forecasts predict, and standard controllers are partly to blame
Place a turbine on a hilltop, the logic goes, and faster winds translate into more power — with output rising as the cube of the speed gain. It’s a foundational assumption in onshore wind planning, and it’s likely wrong in ways the industry has underestimated.
New high-fidelity simulations from a DTU research team suggest that neither the aerodynamics nor the turbine’s own control system behave as that assumption requires. Running detailed computational fluid dynamics experiments with a virtual turbine at three terrain positions — flat ground, the foot of a hill, and the ridge — the researchers found results that challenge standard planning models for complex terrain.
The cubic scaling assumption and why it matters
Wind power output scales with the cube of wind speed — that relationship sits at the heart of wind resource assessment. A hilltop that accelerates wind by 10% is expected to deliver roughly 33% more power, driving siting decisions for onshore projects worldwide.
The hill’s standard configuration had a height and width comparable to the rotor diameter — meaning terrain features and rotor size were on the same scale.
The problem is that this cubic relationship was derived for flat terrain. Applying it to hilly landscapes requires assuming that a turbine’s peak power coefficient, Cp,max, stays constant regardless of where the machine sits. That assumption, recent research suggests, breaks down in complex terrain.
The DTU study set out to isolate two underexplored contributors: how much the physical ceiling on power extraction shifts with terrain, and whether the turbine controller stays at its optimal operating point when surrounding flow changes.
How the simulations were designed
The researchers used Reynolds-averaged Navier–Stokes (RANS) simulations with the DTU 10 MW reference turbine modeled as an actuator disk, placing it in three positions: flat terrain, at the foot of a synthetic Gaussian hill, and on the hilltop.
The hill’s standard configuration had a height and width comparable to the rotor diameter — meaning terrain features and rotor size were on the same scale. When hill dimensions are far larger or far smaller than the rotor, the aerodynamic effect largely disappears, so this setup captured the interaction at its strongest.
Additional simulations varied hill height, width, and surface roughness to test whether core findings held across realistic geometries. The industry-default k-ω² torque controller was calibrated on flat terrain and applied unchanged across all three positions.
What happens to the power ceiling in complex terrain
At the foot of the hill, where the slope accelerates flow passing behind the turbine, the maximum achievable power coefficient increased by 2.46% relative to flat terrain. Flow acceleration behind the rotor helps the turbine extract energy more efficiently.
The hilltop result is more striking. Cp,max dropped by 18.25% there — flow decelerates sharply behind a hill’s crest, reducing how much kinetic energy the turbine can extract, regardless of how fast wind arrives at the rotor.
This breaks the cubic scaling law entirely. Power gains on hilltops tracked closer to a quadratic relationship with wind speed increase, and theoretical analysis suggests the scaling could fall closer to the power of 1.5 under the most locally confined speed-up conditions. The effect persisted across hill geometries tested, fading only when terrain features were much larger or smaller than the rotor.
Why the standard controller falls short
The k-ω² controller maintains a fixed ratio between generator torque and rotor speed squared, keeping the turbine locked onto its optimal operating point in flat terrain. It infers the right operating state from rotor speed alone, without measuring wind speed directly.
In complex terrain, the controller follows the same curve — but that curve no longer aligns with the true optimum. On the hilltop, it settles at a tip-speed ratio and induction level too high for the decelerating conditions behind the crest.
The measured power loss from this mismatch reached up to 1.96% on top of the aerodynamic ceiling reduction. Decelerating flow demands lower induction and a lower tip-speed ratio — adjustments the standard controller simply can’t make on its own.
Implications for wind energy planning and future controllers
The most direct consequence is for energy yield assessments. Speed-up factors assume cubic scaling and a stable power coefficient, and both assumptions fail in the cases this study examined. Assessments built on them will systematically overestimate hilltop output.
The controller-induced loss is roughly 2%, but it’s systematic. In complex-terrain projects already carrying high uncertainty, consistent underperformance compounds with other error sources and could shift project economics meaningfully.
Adaptive control strategies could recover some of that loss. Extremum-seeking control — an approach that slowly adjusts the torque constant by probing for performance improvements — is one candidate, though the researchers note that performance gradients are small and seasonal changes in terrain surface conditions could make sustained optimization difficult.
What comes next is a modeling framework that accounts for streamwise flow development around a turbine, not just local wind speed at hub height. As the industry pushes into more complex landscapes, updating that practice may become necessary rather than optional.
Reference: Zengler, C. P., Gaunaa, M., and Troldborg, N.: Aerodynamic analysis of standard variable-speed torque control in complex terrain, Wind Energ. Sci., 11, 2601–2619, https://doi.org/10.5194/wes-11-2601-2026, 2026.
Daniel García is an Editor-in-Chief with strong expertise in structural work and engineering principles. He combines this technical foundation with deep knowledge of energy, spatial design, and emerging technologies, bringing a forward-thinking and analytical approach to editorial leadership.
