Turbulence in real-sea conditions and its impacts on tidal energy devices

Abstract

Tidal energy is unique among renewable technologies due to its high predictability, making it an essential part of the renewable energy mix. However, extracting energy from turbulent and energetic tidal sites poses durability and design challenges for developers. Existing standards and guides lack specific guidance on accounting for turbulence-induced fatigue loads so windbased turbulence models are commonly used. Uncertainties arising from this approach lead to higher safety factors and increased device costs. This work investigates the suitability of current methodologies and wind-based turbulence models for tidal applications. It aims to identify the most important considerations for fatigue loading, helping to reduce uncertainty in device design.

Comparisons of semi-empirical models to turbulence measurements from four Acoustic Doppler Current Profilers deployed across two tidal sites, show that in many cases the models are not representative. Both sites show significant deviations from the theoretical lengthscales, isotropy ratios and shear profile, and the agreement with spectral models is shown to be component and depth-dependent. The application of Fourier methods for analysing non-stationary phenomena such as turbulence is also examined. By novel application of wavelet analysis, it is shown that intermittent bursts of coherent turbulence are obscured by the averaging associated with Fourier analysis. The energy bursts have instantaneous turbulence intensities up to 80% higher than the average.

The consequences of using wind-based turbulence models in design are explored by testing the sensitivity of turbulence parameters to simulated loads, using the turbine design tool - Tidal Bladed. Varying the turbulence parameters profoundly impacts the loads with turbulence intensity, resulting in a 90% change in fatigue loads (for intensities 2−24%). Length-scales show a 49% difference in loads across the range tested (5−70m) and the load difference between shear profiles is over 20%. Additionally, load measurements from a full-scale, operational turbine demonstrate a load response to a broad range of turbulence scales, including scales much larger than the rotor, with blade pitching modulating this response. It is also shown that even when structure shadow loads are significant due to the downstream position of the rotor, stochastic turbulence is still the bigger driver of fatigue loads.

The findings highlight the need for clearer industry guidance for the treatment of turbulence in design and testing. Caution is raised against using wind-based models in tidal applications and the importance of accurate measurement and derivation of turbulence parameters, in particular turbulence intensity, is highlighted. If excessive conservatism in design is to be reduced, high-quality turbulence measurements for each site and location are required and appropriate measurement and analysis techniques must be employed.