Wind turbines that aren’t spinning may actually be facing aerodynamic forces more dangerous than anything they encounter in motion
A parked wind turbine looks inert — blades locked, rotor still, no energy being generated. But stillness doesn’t mean safety. During extreme wind events, stationary blades can be exposed to powerful, erratic aerodynamic forces that engineers are only beginning to map with precision.
The problem lies in what happens when wind strikes a blade from the wrong direction. Under normal operation, airflow moves predictably from leading edge to trailing edge. When a turbine is parked and the wind shifts, that flow can reverse — and the aerodynamics that follow behave in ways that standard engineering tools were never designed to handle.
When turbines stand still, the aerodynamics get stranger
Wind turbines spend more time than you’d expect in a parked or idling state. During extreme wind events, yaw misalignment, or grid disconnection, operators lock the rotor and wait. The blade doesn’t rotate — but wind still hits it, sometimes from unexpected angles.
Accurate load prediction during standstill feeds into edgewise stability assessments and ultimate load calculations that determine whether a blade survives a storm.
The critical issue is reverse flow. Under normal operation, air moves from a blade’s leading edge to its trailing edge. When a turbine is parked and the wind shifts, that orientation can flip: air strikes the trailing edge first, travels toward the leading edge, and creates a pressure environment the blade was never designed to handle.
This isn’t a rare edge case. Field data cited in the study suggest that blades on turbines rated above 15 MW can become aerodynamically unstable in reverse flow at wind speeds as low as 8 m/s — well below the extreme conditions engineers typically plan for.
Dynamic stall: the aerodynamic phenomenon that complicates everything
To understand why standstill conditions are so difficult to model, you first need to understand dynamic stall. When an airfoil’s angle of attack changes rapidly, flow separation and reattachment don’t happen instantly. There’s a lag — and that lag creates hysteresis loops where lift and drag forces during the upswing of a pitch cycle differ from those during the downswing, tracing an enclosed loop rather than a single curve.
The study examined four distinct stall-induced vibration (SIV) zones: SIV-1 covers normal flow in both positive and negative angle ranges, while SIV-2 covers reverse flow in both directions. Researchers used the S814 airfoil — a 24%-thick profile common in wind turbine applications — running CFD simulations through OpenFOAM across the full 360° angle-of-attack range. Three engineering models were tested: the Øye model, the Beddoes–Leishman (BL) model, and the IAG model.
What the simulations found: asymmetry, vortices, and wide hysteresis loops
In normal-flow (SIV-1) conditions, the CFD results matched wind tunnel measurements well, accurately capturing hysteresis loops in lift, drag, and pitching moment. The IAG model performed best among the three. Both the Øye and BL models significantly underestimated hysteresis — predicting smoother, smaller force swings than actually occur.
Reverse-flow (SIV-2) conditions told a different story. Pronounced asymmetries emerged between positive and negative angle-of-attack ranges, tracing directly to how vortices form differently depending on which direction reversed flow approaches the asymmetric airfoil geometry. The simulations also showed early flow separation, irregular vortex shedding, and hysteresis loops substantially wider than anything seen in normal flow. Higher reduced frequency produced fuller, stronger hysteresis loops; Reynolds number had comparatively little effect.
Why existing engineering models fall short in reverse flow
In SIV-2, all three semi-empirical models struggled. The Øye and BL models produced over-smoothed predictions — missing the vortex-driven force fluctuations that CFD captured clearly, with hysteresis loops that were too narrow, too regular, and too small.
The IAG model came closest, but still underestimated the amplitude of hysteresis peaks. Setting the critical stall normal force to zero improved stall-onset prediction without fixing the deeper problem of underestimating peak loads once stall was already underway. In SIV-2, the airfoil effectively behaves as if stalled across nearly its entire operating range — a condition that conventional stall-onset definitions, built around normal-flow assumptions, simply don’t address. An L2-norm error analysis confirmed IAG outperformed the others across all four SIV zones, but even that advantage proves insufficient for the most demanding reverse-flow scenarios.
What this means for the future of large wind turbine design
For engineers designing the next generation of ultra-large turbines, these findings carry direct structural implications. Accurate load prediction during standstill feeds into edgewise stability assessments and ultimate load calculations that determine whether a blade survives a storm.
The study also flags a data quality problem. Engineering models depend on input polar data describing airfoil behavior across the full angle-of-attack range. Experimental wind tunnel data produce the best results, while CFD-extrapolated data — the most common industry approach — rank lower in reliability. For SIV-2, that gap matters.
Future dynamic stall models should treat normal and reverse flow as fundamentally distinct regimes, with separate time-delay and flow-memory representations for each. Higher-order harmonic terms, absent from current first-order semi-empirical models, may also be necessary to capture vortex-driven force fluctuations fully. Until those tools exist, the aerodynamic behavior of a parked turbine in reverse flow remains one of the least understood gaps in wind energy engineering.
The complete study is available here: Bangga G, Syawitri TP and Chairunnisa (2026) Dynamic stall modeling of a wind turbine airfoil at various stall-induced vibration zones. Front. Energy Res. 14:1761590. doi: 10.3389/fenrg.2026.1761590
Carlos is an engineer with strong expertise in technical and industrial topics. He previously worked at international companies such as Siemens and is multilingual.
