HIL Wind-Tunnel Testing of Floating Turbines

• Hardware-in-the-loop wind-tunnel methodology is an integrated approach combining physical tests and numerical models to replicate aerodynamic and hydrodynamic interactions in floating wind farms.
• The experimental setup uses real-time feedback to capture two-way coupling between rotor dynamics and platform responses, thereby revealing wake-induced excitation mechanisms.
• Quantitative findings, such as a 58% drop in downstream thrust and increased platform motion, inform optimal turbine array design, mooring tuning, and control strategies.

A hardware-in-the-loop (HIL) wind-tunnel methodology is an integrated experimental–numerical approach for analyzing the coupled dynamics of floating wind turbines and their wakes. This paradigm merges physical wind-tunnel testing of one or more scaled turbine rotors with a real-time software model emulating platform motion, hydrodynamic loads, and mooring systems, enabling the capture of two-way interactions between aerodynamics and platform kinematics. Such a framework is indispensable for studying wake-induced excitation mechanisms in floating wind farms, where platform motion and aerodynamic forces are intrinsically interdependent and cannot be reliably captured by prescribed or open-loop kinematic surrogates (Fontanella et al., 11 Jan 2026).

1. Definition and Core Principles of HIL Wind-Tunnel Testing

The HIL wind-tunnel methodology is defined by the tight cyber-physical coupling of physical wind-tunnel hardware (scaled rotors, actuators, sensors) with a high-fidelity numerical model (computing platform surge/pitch, hydrodynamics, and mooring dynamics) closed through real-time feedback. This integration enables direct measurement of loads and motions, model-based dynamic responses, and transfer of those responses back to physical hardware, ensuring both aerodynamic and hydrodynamic fidelity.

Key elements:

• Physical rotors mounted on robotic actuation platforms (reproducing surge and pitch DOFs).
• Real-time sensors reconstruct aerodynamic loads and platform positions at sub-millisecond intervals.
• Numerical solvers propagate the full-order platform dynamics under time-varying aerodynamic and hydrodynamic forces.
• The platform state vector is fed back to the robot controller at 1 kHz rate, maintaining synchronization to within sub-cycle accuracy.

This architecture is essential for fully capturing two-way coupling: platform oscillations imprint low-frequency signals in the near wake, which modulate the inflow and thus the aerodynamic load on downstream turbines. The framework is especially critical for floating offshore wind energy, where accurate validation of array-scale models and control strategies must account for this bidirectional fluid–structure interaction (Fontanella et al., 11 Jan 2026).

2. Experimental and Numerical Model Structure

The physical experimental setup consists of two 1.2 m-diameter rotors (scale 1:150) with robotic actuation platforms allowing independent surge (xₛ) and pitch (βₛ) motion. The wind tunnel provides steady, low-turbulence inflow. The numerical backbone encompasses:

• Aerodynamic load computation using measured/estimated local inflow Uₑ, tip-speed ratio λ, and rotor-specific Cₜ(λ), C_q(λ) mappings.
• An augmented Jensen-type model for the wake velocity deficit, parameterized by expansion coefficient k_wake, and incorporating a turbulence kinetic energy (TKE) term for downstream turbulence intensity.
• Platform motion is governed by

$$(M_{\rm fowt}+A_\infty)\,\ddot qₛ + R_{\rm hydro}\,\dot qₛ + K_{\rm hs,moor}\,qₛ = F_{\rm wt,n}$$

employing full-order mass, added-mass, stiffness, and damping matrices.

• Aerodynamic force feedback is realized by subtracting inertia and mooring compensation (from wind-off system ID) from tower-top load cell measurements, thus ensuring net aerodynamic loading is properly transferred from wind-tunnel to numerical domain.

The model and data are non-dimensionalized appropriately (e.g., time scale: λ_t = λ_L/λ_U = 1/60), and key dimensionless groups (Reynolds, Froude) are tracked, with compensating terms added numerically where scaling laws cannot be directly preserved in the laboratory.

3. Wake Deficit, Turbulence, and Platform Response

Downstream turbines (WT2) in the wake of upstream turbines (WT1) experience both reduced mean inflow velocity and enhanced wake turbulence intensity (TI_wake rises from 2% to up to 14% at wake edges over 3.5–4.3 D). The persistence of low-frequency energy in the wake TKE leads to increased spectral energy at both surge and pitch resonance frequencies for WT2.

Notable experimental findings include:

• Mean thrust coefficient Cₜ reduction for the downstream turbine from 0.82 (free stream) to 0.34 (wake), a 58% drop.
• RMS and mean platform displacements (x, β) for WT2 are driven higher at natural mode frequencies.
• Power spectral densities for WT2’s platform motions show 3–5× increases at the natural surge and pitch frequencies when compared to free inflow, consistent with direct resonance excitation via wake velocity fluctuations.

The mechanism chain is explicit: WT1’s motion imprints low-frequency oscillations into the wake; slow wake recovery ensures these energy components persist to WT2; WT2 aerodynamic loading is thus forced at resonance; platform surge/pitch modes are amplified (Fontanella et al., 11 Jan 2026).

4. Wake-Induced Excitation Mechanisms and Quantitative Results

The HIL method enables direct diagnosis of wake-induced excitation paths:

• Coherence analysis (not reported in detail but can be robustly estimated): γ²[U′, xₛ] > 0.8 at f_surge, γ²[U′, βₛ] > 0.7 at f_pitch with −90° phase lag, confirming that wake-induced inflow fluctuations drive platform resonance.
• Quantified reduction in mean thrust, amplification of RMS surge/pitch response, and spectral peak gains provide experimental closure for numerical code validation and for ocean-scale model parameterization.

The methodology enables experimental campaigns that would be intractable in the field (due to cost, scale, and lack of control), providing anchored data for model validation and optimization of wind farm layouts, mooring designs, and collective-control strategies.

5. Implications for Wake-Resonance Mitigation and Farm Control

Experimental results from HIL wind-tunnel testing directly inform array design and active control:

• Turbine array layout: Increasing spacing beyond 6–8 D reduces the residual wake energy at resonance frequencies, attenuating excitation of downstream platforms.
• Mooring system tuning: Shifting natural frequencies (via mooring stiffness) away from the dominant low-frequency wake components suppresses resonance amplification.
• Active/farm-level control: WT1 can be deliberately operated to "dither" wake frequencies off WT2's resonance, while collective downstream pitch and generator-torque control can be employed for active damping of resonance response.

Metrics derived from HIL testing (e.g., Cₜ reduction, ΔRMS, PSD-peak amplification) can be fed into layout optimization codes and dynamic simulation frameworks for floating wind farms. This capability is essential for supporting the design of next-generation floating wind plants that are robust against both mean-wake losses and dynamic resonance fatigue (Fontanella et al., 11 Jan 2026).

6. Methodological Advantages, Scope, and Limitations

Distinct from purely physical or purely numerical experiments, HIL wind-tunnel methodology offers:

• Real-time, bidirectional coupling between fluid-structure and platform dynamics at high temporal resolution, capturing both transient and steady-state phenomena.
• The ability to conduct physically accurate resonance studies under controlled inflow and boundary conditions.
• Systematic parametric exploration (platform design, control, wake geometry) not possible in the oceanic environment.

Limitations include incomplete scaling of relevant non-dimensional numbers (full Reynolds and Froude similarity not achieved) and reliance on numerical correction for certain hydrodynamic/elastic phenomena. Nonetheless, for problems dominated by aerodynamic–platform coupling and wake-induced fatigue, HIL wind-tunnel methodology is uniquely capable (Fontanella et al., 11 Jan 2026).

---

References:

• "Hardware-in-the-loop wind-tunnel testing of wake interactions between two floating wind turbines" (Fontanella et al., 11 Jan 2026)