Presentations

Abstract

Turbulence Intensity has received a lot of interest of late, especially with respect to Offshore Wind. Turbulence Intensity (TI) is used by structural engineering to help achieve the optimal structural design for offshore wind. It is also used commonly for wake modelling, and has an impact on the predicted power curve for a given wind turbine being assessed for the offshore wind project. However, with the transition from met masts offshore to floating lidar systems, there remains uncertainty on the best method to acquire site specific TI data. This has led to approximations being used and the term "conservative approach" being put forward by the relevant engineering departments. However, TI defies a unified conservative approach. Conservatism in structural design wants a higher-than-expected value. In wind resource it is lower. So how do we solve this? There is little normative guidance which aligns both wind resource and structural requirements. At a pre-normative level there have been recent publications issued by DNV (DNV RP-0661 2023) and NEDO (NEDO March 2023) which promote an acceptable level of bias when looking to acquire and present TI data when seeking to use lidar systems as the campaign primary measurement device. DNV suggest acceptable bias error levels depending on use case which range from -6% to +10% applied in a windspeed bin-wise manner. NEDO suggest a tighter mean bias error of less than 5% when applied to the standard deviation comparison between the campaign measurement device and a cup anemometer reference. It is implied that this is a bin-wise metric in that the standard deviation error is calculated for each wind speed bin. Structurally, papers such as that presented by Wood Thilsted (Berkeley et al. May 2023) have shown a clear impact of TI error on structural modelling, but not on energy predictions. The Wood Thilsted work linked a 1% change in TI to a 1% change steel weight required. This sensitivity when compared to the pre-normative error levels gives quite a sobering range of possible costs when looking at a 500MW wind farm development. But what about the impact on energy yield? This paper applies the suggested error proposed by NEDO and DNV to a standard 500MW wind turbine layout. The work presents the impact on the predicted level of wake losses when using engineering models, and the impact on a predicted power curve. Engineering wake models commonly use TI to tune the wake decay coefficient according to (e.g. ) Pena et al 2016. Whereas this work does not say this is the best method of determining the predicted annual energy production, these models are still commonly used as they provide a quick sensitivity to layout changes. Further work is required to reduce the uncertainty in the wake loss values, but this is not assessed here. The impact of the proposed errors is evaluated and two questions are asked: 1.Are the error levels sufficient given the sensitivity of projects to the cost of energy? 2.Are TI measurements justified in terms of cost?