Fatigue damage calculation of offshore wind turbines’ long-term data considering the low-frequency fatigue dynamics

Abstract

The mutual and simultaneous action of external variable forces on (offshore) wind support structures causes fatigue. Fatigue analysis in this context relies on 10-minute-long strain signals. Cycle-counting allows capturing the fatigue cycles nested within these signals but inevitably leaves some open loops called half-cycles or residuals. Some other loops are not even cycle-counted because the algorithm cannot catch low-frequency (LF) cycles spanning more than 10 minutes, as caused, e.g., by wind speed variations. Notwithstanding, LF cycles are also the most damaging since they always contain the highest range in the variable amplitude signal. Therefore, counting multiple 10-minute signals and merging the resulting histograms into one inevitably has some non-conservative effects, for it leaves the LF cycles uncounted. In this work, we avoid signal concatenation to recover the LF effect. To do so, we use the residuals sequence from the 10-minute signals as it embeds the LF information. As method validation, we compare the impact of LF fatigue dynamics recovery on the linearly accumulated damage using a real-life dataset measured at a Belgian offshore wind turbine. By defining a factor that incorporates the LF effect, we observe that after 300 days of observation, the factor converges to a fixed value.

Farm‐wide interface fatigue loads estimation: A data‐driven approach based on accelerometers

“…Then, by employing a rainflow counting algorithm, 47,48 holding the linear damage accumulation hypothesis as true (Palmgren-Miner's rule) 49 and through the employment of the Wöhler exponent (the negative inverse slope of the SN curve), 50 a damage-sensitive feature, the damage equivalent loads (DEL), can be calculated for any ten minute window. 51,52 DELs are widely adopted among load engineers to quantify fatigue rates (over, e.g., damage) and are furthermore used in design documents, which opens the door to comparisons. Therefore, given the Wöhler exponent, m, the number of cycles, n j , of a given stress range, Δσ j and the tower-transition piece outer and inner radii, r o and r i , respectively at strain gauge locations we can calculate the DEL, as given by Equation (1).…”

“…Through the 10‐min average yaw angle from the SCADA, the strain gauges enable the calculation of the bending moments and . Then, by employing a rainflow counting algorithm, 47,48 holding the linear damage accumulation hypothesis as true (Palmgren—Miner's rule) 49 and through the employment of the Wöhler exponent (the negative inverse slope of the SN curve), 50 a damage‐sensitive feature, the damage equivalent loads (DEL), can be calculated for any ten minute window 51,52 . DELs are widely adopted among load engineers to quantify fatigue rates (over, e.g., damage) and are furthermore used in design documents, which opens the door to comparisons.…”

Reconstruction of environmental site conditions by the integration of SCADA and reanalysis data

“…Linear interpolation may be sufficient for most atmospheric parameters. However, low-cycle fatigue damage is affected by variations in the mean wind speeds [11], which are present in the SCADA signal but not in the much smoother ERA5 signal, and that cannot be modelled by turbulence intensity.…”

Fatigue Assessment of Wind Turbine Towers: Review of Processing Strategies with Illustrative Case Study

“…Sadeghi et al [72] found differences of 10%, 25%, and 50% between cumulative damage for different wind lengths at a multi-megawatt offshore wind turbine support structure, considering S-N curve with Wöhler exponents of 3, 4, and 5, respectively.…”

“…However, this option is undoubtedly time/memory demanding. An approach that significantly reduces calculation time without losing accuracy in the final spectrum histogram was introduced by Marsh et al [73] and successfully applied by Sadeghi et al [72]. To obtain the number of cycles per stress level, the 10 min strain time series were processed with the rainflow cycle-counting method [53].…”