Minutes from IEA Wind Task 52 General Meeting 2025

IEA Task 52 AGM Sam Cressall BEng, MScR Exploring The Impact Motion Corrected Floating Lidar TI On Annual Energy Production Slide 2 Presentation Scope and Motivations Motivation •FLS TI departs differs from point TI measurements; motion and averaging effects •Uncorrected FLS -> artificially higher TI in the resource assessment •Large push towards developing better correction methods for TI •How effective is this correction in applications such as energy production models? •Well defined KPIs exists for HWS and WD but TI is yet to be fully defined (Carbon Trust, 2025) Scope •Review concepts relevant to turbulence, FLS, and AEP •Monte Carlo: propagate measurement errors of TI distributions into AEP uncertainty. •Quantify change in AEP P50 (accuracy) and P90 / P50 ratio (confidence) •Present the contextual importance of this uncertainty Slide 3 (Lee & Jason Fields, 2021) Wind Resource Assessments Uncertainties Measurement uncertainty is Epistemic & Aleatoric Today we will discuss the epistemic component of floating lidar ✓A 3% loss factor equality 1GW has an estimated cot of 50-60 million EUR (DNV, 2019) ✓Increases in P90/P50 reduces the LCOE (DNV, 2016) ✓Reducing uncertainty by 1 % can result in USD 0.5 - 2 million of economic benefits (Brower et al.,2015; Bodini et al., 2022) ✓A change of 1 % in wind speed uncertainty can lead to a 3 % to 5 % change in NPV of a wind farm (Kline, 2019). ✓Modelling flow around masts can reduce wind speed measurement uncertainty from 2.68 % to 2.23 %, which translates to GBP 1.2 million of equity savings for a 1 GW offshore wind farm in the United Kingdom (Crease, 2019) Slide 4 Floating Lidar, Motion, And Turbulence •Floating lidar systems (FLS) have become a mature technology for measuring HWS for offshore WRA •Hydrodynamic forces on the buoy artificially increases the TI measurement •Motion correction strategies are now being validated, certified and accepted! Different designs have different motion Slide 5 FLS TI Motion Correction What do the standards say? DNV-RP-0661 (2023) •TI MRBE +/- 10% 0.5% change in AEP •Based on wake affects from experiment •Developed for onshore lidar The Carbon Trust OWA roadmap (2025) •Stage 2/3 + ; •XTI, CTI, Cti, R2ti, TIMBE, TIRMSE, TIRMBE, TIRMSRE, TIREP-D IEC61400 50-4 (2025) •Comparing the wind speed interval binned turbulence from reference and FLS for the test period. Do I need motion compensation? Slide 6 Floating Lidar and Turbulence •Deterministic method from (Watson et al., 2025) •Motion increases the mean and spread •Correction improved the mean but a high spread remains –We are normalising the regression to met mast residual distributions Fixed CW to Met mast Raw to Fixed CW Corrected to Fixed CW Slide 7 Normal Distributions of TI Errors TIbase = 0.1 TIbase = 0.2 TIbase = 0.05 Metric Fixed TI Fixed Lidar CW_Raw CW_Corr Pulsed_Raw Pulsed_Corr Slope to MM (mtot)1.158 1.208 1.08 1.459 1.082 Intercept to MM (Ctot)-0.007 0.028 0.001 0.054 0.005 RMSE 0.01 0.05 0.021 0.08 0.018 Wind speed Distribution based on CT KPIs While we do show both pulsed and continuous wave distributions, we are not comparing pulsed to CW, only corrected to raw Similar findings presented by (Kelberlau et al., 2020) 𝓝𝑻𝑰 (𝒎𝒕𝒐𝒕𝒙+𝑪𝒕𝒐𝒕,𝑹𝑴𝑺𝑬) Slide 8 Annual Energy Production AEP KPIS P90 / P50= (3879/3928) = 0.988 A measure of uncertainty or confidence 𝐴𝐸𝑃𝑝=෍ 𝑁𝑇𝐴𝐸𝑃𝑇𝑖=෍𝑛𝑡Δ𝑡න𝑃𝑇𝑖𝒙𝑖𝜂𝑖𝒙𝑖𝑓 𝒙𝑖𝑑𝒙𝑖 𝑛𝑡 = 8760 [hours] 𝑓 𝒙𝑖 = Joint distribution of wind condition at the turbine [U,Ѳ] 𝒙𝑖 = HWS and WD averaged over a time step ∆t Δ𝑡 = Averaging period 𝜂𝑖𝒙𝑖 = Wake losses at the turbine 𝑓 𝒙𝑖 = Long term wind resource distribution 𝑃𝑇𝑖𝒙𝑖 = Power output of the turbine 𝑁𝑇 = Number of turbines •These distributions are commonly created from Weibull wind speed distributions interannual variability (IEC 61400-12-1) •P50 and P90(shown as P10) values provides banks confidence in the estimation on the AEP of the farm. (DNV, 2016) Slide 9 Annual Energy Production AEP KPIS P90 / P50= (3879/3928) = 0.988 A measure of uncertainty or confidence 𝐴𝐸𝑃𝑝=෍ 𝑁𝑇𝐴𝐸𝑃𝑇𝑖=෍𝑛𝑡Δ𝑡න𝑃𝑇𝑖𝒙𝑖𝜂𝑖𝒙𝑖𝑓 𝒙𝑖𝑑𝒙𝑖 𝑛𝑡 = 8760 [hours] 𝑓 𝒙𝑖 = Joint distribution of wind condition at the turbine [U,Ѳ] 𝒙𝑖 = HWS and WD averaged over a time step ∆t Δ𝑡 = Averaging period 𝜂𝑖𝒙𝑖 = Wake losses at the turbine 𝑓 𝒙𝑖 = Long term wind resource distribution 𝑃𝑇𝑖𝒙𝑖 = Power output of the turbine 𝑁𝑇 = Number of turbines •These distributions are commonly created from Weibull wind speed distributions interannual variability (IEC 61400-12-1) •P50 and P90(shown as P10) values provides banks confidence in the estimation on the AEP of the farm. (DNV, 2016) Slide 16 Sensitivity Analysis P50 % Error P90 / P50 AEP P50 % Errors Corrected: 0.5% < P50 < 2% Raw : 2% < P50 < 12% P90 / P50 Confidence Corrected: 0.97 < P90 / P50 < 0.985 Raw : 0.89 < P90 / P50 < 0.98 Wind Resource Base TI All lidar cases increase the AEP P50 compared to a Fixed TI value In all cases corrected lidar offers improvements over raw lidar Fixed lidar confidence is equivalent to Fixed TI confidence Discussion Slide 18 AEP Uncertainty and Confidence Thinking broader: 𝜎𝐹𝐿𝑆 =1−𝑃90/𝑃50 1.282 ‘For offshore wind farms currently in development, DNV GL would typically expect to see a 10year P90/P50 ratio of the order of 88% to 92%’ (DNV, 2019) Combining, in quadrature, FLS uncertainty and Base AEP uncertainty using (JCGM,2008; IEC 61400-12-2:2022) AEP Uncertainty: 7.95% (Fixed), 8.14% (Corrected), 11.60% (Raw) •Using FLS TI in AEP estimations, then the –Raw TI contributes >3% compared to fixed –Corrected TI contributes 0.2% compared to fixed •Interannual variability contributes 6% (Pryor, 2018) Fixed Lidar Corrected Raw P50 change <1% <2% <12% P90/P50 >0.98 >0.97 >0.89 𝜎𝐹𝐿𝑆 1.56% 2.24% 8.58% Slide 19 Summary: Is The Correction Good Enough? •Corrected TI measurements bias AEP P50 by 0.5-2%, slightly higher than DNV-0661 •AEP uncertainty from corrected TI increases by 0.2% compared to fixed lidar •Where should we focus the efforts of further correction? –Bias means more accurate corrections required –Machine learning and better access to data •Metmast equivalent TI –Characterising sensitivities and models of FLS TI errors (motion, wind speed, stability) •Apply 50-4 method to better model TI measurement errors •Better measurements do lead to better modelling outputs (Klemmer et al., 2024) •Opportunity to use in-situ TI measurements from FLS; needs evidence of maturity •Presented Preliminary methodology to assess AEP based TI KPIs •Recommend the DNV AEP method should be update to review FLS and offshore sites Precision or Accuracy Slide 20 References Alexandra St Pé, A., Weyer, E., & Arntsen, A. E. (2021). CFARS Site Suitability Initiative: An Open Source Approach to Evaluate the Performance of Remote Sensing Device (RSD) Turbulence Intensity Measurements & Accelerate Industry Adoption of RSDs for Turbine Suitability Assessment (Vol. 27). Bodini, N., Optis, M., Perr-Sauer, J., Simley, E., & Fields, M. J. (2022). Lowering post-construction yield assessment uncertainty through better wind plant power curves. Wind Energy, 25(1), 5–22. https://doi.org/10.1002/we.2645 Brower, M., Robinson, N. M., & Vila, S. (2015). WIND FLOW MODELING UNCERTAINTY: Theory and Application to Monitoring Strategies and Project Design. www.ul.com/openwind Crease, J. (2019). CFD Modelling of Mast Effects on Anemometer Readings. WindEurope 2019, WindEurope, Bilbao, Spain,. DNV. (2019). ENERGY PRODUCTION ASSESSMENT VALIDATION Great Britain, Ireland, North Europe offshore and South Africa. www.dnvgl.com DNV. (2023). RP-0661 Lidar-measured turbulence intensity for wind turbines. IEC. (2022). 61400-1:2019 Wind energy generation systems. IEC 12-1. (2022). IEC 61400-12-1:2022 Part 12-1: Power performance measurements of electricity producing wind turbines. www.gov.uk. IEC 12. (2022). BS EN IEC 61400 12 Wind energy generation systems. Part 12, Power performance measurements of electricity producing wind turbines. Overview. British Standards Institution. IEC 12-2. (2022). BS EN IEC 61400 12-2: Power performance of electricity producing wind turbines based on nacelle anemometry. IEC 50-2. (2022). BS EN IEC 61400 50-2: Wind measurement — Application of groundmounted remote sensing technology. www.gov.uk. IEC 50-4. (2024). Use of floating lidar systems for wind measurements. http://www.dke.de JCGM. (2008). Evaluation of measurement data — Guide to the expression of uncertainty in measurement. www.bipm.org Jensen, N. O. (1983). A note on wind generator interaction. In Downloaded from orbit.dtu.dk on. Katic, I. ;, Højstrup, J. ;, & Jensen, N. O. (1987). A Simple Model for Cluster Efficiency. In Citation (Vol. 1). APA. Kelberlau, F., Neshaug, V., Lønseth, L., Bracchi, T., & Mann, J. (2020). Taking the motion out of floating lidar: Turbulence intensity estimates with a continuous-wave wind lidar. Remote Sensing, 12(5). https://doi.org/10.3390/rs12050898 Kline J. (2019). Detecting and Correcting for Bias in Long-Term Wind Speed Estimate. AWEA Wind Resource and Project Energy Assessment Workshop 2019, AWEA, Renton, WA. Klemmer, K. S., Condon, E. P., & Howland, M. F. (2024). Evaluation of wind resource uncertainty on energy production estimates for offshore wind farms. Journal of Renewable and Sustainable Energy, 16(1). https://doi.org/10.1063/5.0166830 Kosović, B., Basu, S., Berg, J., Berg, L. K., Haupt, S. E., Larsén, X. G., Peinke, J., Stevens, R. J. A. M., Veers, P., & Watson, S. (2025). Impact of atmospheric turbulence on performance and loads of wind turbines: Knowledge gaps and research challenges. https://doi.org/10.5194/wes-2025-42 Lee, J. C. Y., & Jason Fields, M. (2021). An overview of wind-energy-production prediction bias, losses, and uncertainties. Wind Energy Science, 6(2), 311–365. https://doi.org/10.5194/wes-6-311-2021 Liu, S., Li, Q., Lu, B., & He, J. (2024). Impact of incoming turbulence intensity and turbine spacing on output power density: A study with two 5MW offshore wind turbines. Applied Energy, 371. https://doi.org/10.1016/j.apenergy.2024.123648 Nygaard, N. G., Steen, S. T., Poulsen, L., & Pedersen, J. G. (2020). Modelling cluster wakes and wind farm blockage. Journal of Physics: Conference Series, 1618(6). https://doi.org/10.1088/1742-6596/1618/6/062072 Pryor, S. C., Shepherd, T. J., & Barthelmie, R. J. (2018). Interannual variability of wind climates and wind turbine annual energy production. Wind Energy Science, 3(2), 651–665. https://doi.org/10.5194/wes-3-651-2018 Ren, G., Liu, J., Wan, J., Li, F., Guo, Y., & Yu, D. (2018). The analysis of turbulence intensity based on wind speed data in onshore wind farms. Renewable Energy, 123, 756–766. https://doi.org/10.1016/j.renene.2018.02.080 Richter, P., Wolters, J., & Frank, M. (2022). Uncertainty quantification of offshore wind farms using Monte Carlo and sparse grid. Energy Sources, Part B: Economics, Planning and Policy, 17(1). https://doi.org/10.1080/15567249.2021.2000520 Saint-Drenan, Y. M., Besseau, R., Jansen, M., Staffell, I., Troccoli, A., Dubus, L., Schmidt, J., Gruber, K., Simões, S. G., & Heier, S. (2020). A parametric model for wind turbine power curves incorporating environmental conditions. Renewable Energy, 157, 754–768. https://doi.org/10.1016/j.renene.2020.04.123 The Carbon Trust. (2025). Offshore Wind Accelerator roadmap for the commercial acceptance of floating LiDAR technology. DNV-GL. (2016). STUDY ON UK OFFSHORE WIND VARIABILITY Study on UK Offshore Wind Variability. www.dnvgl.com Watson, W., Wolken-Möhlmann, G., & Gottschall, J. (2025). Evaluating the Impact of Motion Compensation on Turbulence Intensity Measurements from Continuous-Wave and Pulsed Floating Lidars. https://doi.org/10.5194/wes2025-45 Thank You Questions? Sam Cressall [email protected].uk Let’s Connect! Slide 22 Studies A number of case studies are run to assess: •Effect of farm deign and site conditions on the uncertainty requirements •If correction impact is related to site specific conditions Case number Base TI Turbine Size Base Shape mean Spacing Type 1.1 0.1 10MW 10.16 7 BASE 2.1-2.10 0.04-0.2 10MW 10.16 7 Sensitivity 3.1-3.10 0.1 5-15MW 10.16 7 4.1-4.10 0.1 10MW 8-12 7 5.15.10 0.1 10MW 10.16 5-10 Slide 23 Sensitivity Analysis P50 % Error P90 / P50 Wind Resource Base TI Spacing Turbine Size AEP P50 % Errors Corrected: 0.5% < P50 < 2% Raw Lidars: 2% < P50 < 12% P90 / P50 Confidence Corrected: 0.97 < P90 / P50 < 0.985 Raw Lidars: 0.89 < P90 / P50 < 0.98 All lidar cases increase the AEP P50 compared to a Fixed TI value In all cases corrected lidar offers improvements over raw lidar Stronger relationship with site resource than the farm design, binned TI KPIs are suitable Fixed lidar confidence is equivalent to Fixed TI confidence Slide 24 TI measurement uncertainty DNV (DNV-RP-0661) Set the AEP threshold for TI to be an MRBE of +/-10% Based on wake affects from experiment ‘Not for floating lidar’ Metric HWS slope HWS R2 Stage 2 0.97-1.03 >0.97 Best Practice 0.98 – 1.02 >0.98 CT OWA Roadmap V3 (The Carbon Trust, 2025) No agreed upon threshold but defines 7 parameters to consider for performance Recommend a single variant regression to fit TI IEC (IEC 61400-50-4:2025) Wind speed interval binned comparison Sensitivity study following type and unit qualification CFARS (St. Pé, A., et al., 2021) 90th percentile thresholds Slide 25 AEP Module PyWake from DTU is used to model the AEPp For each sample the AEP is deterministically calculated based on the input parameters