Abstract
An analysis is made of wind turbines in a row by means of an extension to actuator disk theory and a representation of the turbulent diffusion in the wake by a velocity deficit scale and a single free parameter. Beyond this, no wake model is used. It is shown that when the thrust coefficient is 'high' a maximisation of overall power output leads to a large drop in power after the first turbine, followed by a fairly constant level and a rise at the end of the row; this behaviour is a natural consequence of optimisation, and on this basis a 'deep array effect' is to be expected. A variation of turbine size and the effect of impaired turbine performance are examined. The approach can also be used to calculate the turbine upstream velocity (with respect to a reference) from a distribution of measured power output and to make inferences about wake development. The approach could be useful in the assessment of wake models as well as turbine operation.