Presentations

Abstract

Maximising the energy production of operating wind turbines is of primary concern to asset owners. Changes to wind turbine control, e.g., to fine pitch angle or generator torque-speed relationship, are often observed when monitoring performance, yet their impact on production is not easily determined. Significant hurdles in quantifying the impact are inaccurate measurement and inherent stochastic nature of the wind, and not having access to the structural and aerodynamic design data of the turbine required to perform the theoretical calculations. This work presents a novel algorithm to estimate the effect of commonly experienced controller changes on the annual energy yield (AEY) without relying on measurements of the wind conditions. The algorithm is due to be open-sourced in the spring of 2024. The impact of control path changes on aerodynamic efficiency of the rotor (Cp) can be estimated by means of a representative lookup table that describes Cp as a function of rotor tip-speed ratio and pitch angle (such as the IEA task 37 reference 3.4MW model), as the shape of this function is not expected to show large variation amongst modern pitch regulated turbines. The algorithm uses this relationship to build a reference torque-speed control path using the tip-speed ratio and pitch angle associated with the maximum power coefficient to define control region 2. The user can adjust control regions 1.5 and 2.5, via a small set of normalised parameters. In addition, non-optimal alternative control paths can be defined which differ from the reference case by, for example, having an offset from the fine pitch angle or relative change from the optimal mode gain. A power curve is then built for each control path by finding the wind speed that balances the aerodynamic torque (constrained by the rotor speed) with the generator torque for discrete points along the path. The AEY values are then calculated by integrating the power curves with a user-defined Weibull wind speed distribution, and the resulting relative change between the reference and alternative control paths cases. We present here results for a typical case with an annual mean wind speed of 8m/s. In such a case a 1-degree change in the fine pitch angle results in a 0.4% reduction in AEY. However, the effect on AEY is non-linear as 2-degree and 5-degree increases in the fine pitch angle result in 1.9% and 12.6% reductions, respectively. On the other hand, changes in the mode gain (the ratio between the generator torque and square of speed in region 2) away from the optimum have a weaker effect. A 10% change in mode gain reduces the AEY by only 0.1% and even a 30% change results in only a 0.5% AEY reduction. Once set up the algorithm provides a quick estimate of the AEY (and monetary) impact of various controller changes that we often observe in the field, supporting asset managers in providing the potential impact of (not acting on) a controller change and therefore allows for prioritisation of operational tasks and reducing lost yield.