Connecting spatio-temporal turbulence spectra to wind farm power fluctuations

Reliable prediction of wind-farm power variability is essential for grid integration and control. We describe a physics-based framework that maps a previously developed space–time turbulence spectral model to the aggregate power spectrum of a wind farm, using only mean geostrophic velocity at boundary layer height, effective surface roughness, and turbine layout as inputs. A fully parameterized wavenumber–frequency description of boundary-layer turbulence is coupled with a layout-dependent sampling transfer function that represents the farm as a spatial filter on the flow. Turbulence parameters (friction velocity and hub-height mean wind) are obtained from a top-down wind-farm boundary-layer model, thus no direct wind-field measurements or empirical tuning are required. We evaluate the approach using a high-fidelity large eddy simulation dataset of a 60-turbine wind farm in a conventionally neutral atmospheric boundary layer. The dataset is publicly available as part of the Johns Hopkins Turbulence Database – Wind (JHTDB-wind) open numerical laboratory. We find that the model accurately reproduces the low-frequency scaling of aggregate power fluctuations and the distinct spectral peaks associated convective travel times between turbine rows. These results establish a direct, quantitative link between spatio-temporal turbulence physics and wind-energy power variability. They illustrate the use of turbulence theory–based models for potential future engineering applications such as layout design, control, and grid integration of wind farms.c