Posters

Abstract

Wind turbine power production is typically calculated as a function of density-corrected wind speed measured at hub height. However, atmospheric stability, whilst its effect on wake losses has become well-documented, also plays an overlooked role in wind turbine performance as a primary driver of temporal variations in vertical wind shear, where stable conditions produce strong shear. As turbines are increasingly sited in non-temperate climates with strong stability cycles, the non-linear impact of atmospheric stability on wind profiles challenges the validity of power curves measured in moderate wind conditions, leading to systematic deviations from warrantied performance. This study aims to quantify the impact of atmospheric stability in multiple climates, via its influence on vertical wind shear, on turbine power performance. By analysing operational data from over 500 turbines across 32 sites in 9 countries (spanning latitudes 35°S to 59°N), we investigate how shear affects turbine power output. SCADA data were cleaned to exclude invalid samples and curtailments. Temporal shear variation was characterised using the shear exponent, α, derived from ERA5 reanalysis data, using the 10m and 100m wind speed signals. This work assumes that shear at the ERA5 node and site is highly correlated, though local terrain effects may introduce a constant bias. Power production was binned by density-corrected hub-height wind speed (as measured by nacelle anemometers) and split into five shear categories, with curves normalised for turbulence intensity. This methodology isolates the shear-specific impact on performance, while acknowledging its root cause in atmospheric stability. Results confirm a clear trend: turbines produce less power relative to hub-height wind speed as shear increases. Across 80% of sites, power output in partial load dropped by 1–5% between α=0.14 (typical of unstable conditions) and α=0.5 (a commonly seen value in stable conditions), with an average 1.1% decrease per 0.1 increase in α. However, the difference in power output between α=0.14 and α=0 (unstable conditions) was not statistically significant.  As the distribution of α is positively skewed, neutral conditions are not representative of mean performance. As they are often used as the baseline for warrantied power curves, but fail to represent real-world performance in regions with strong stability cycles, this can be a strong but underappreciated factor in energy production bias. These findings challenge current practices in two ways: * Given that many pre-construction assessments consider only neutral, or average, shear profiles for turbine power production calculations, there is likely a cause for bias in location with strong stability cycles. Furthermore, the IEC’s rotor-equivalent-wind-speed (REWS) method, which estimates that high shear produces more power for a given hub-height wind speed, appears over-optimistic – meaning that when variable shear is considered using this method there may be further cause for bias. * Performance monitoring should explicitly account for shear (and by extension, stability) for power curve efficiency assessments. Addressing this bias is critical not only for improving energy yield predictions but also for ensuring the financial viability of the industry in an era of rapid global expansion into diverse climates.