Abstract
This paper presents analytical and computational predictions of the performance of an operating 2.5 m long composite wind turbine blade and compares these predictions with the results of detailed measurements. Part 1 of this paper describes in detail the important aspects of the 5 kW wind turbine, the experimental equipment and data acquisition procedures and presents and discusses some of the results from the experimental data. Here the analytical methodology of Eggleston and Stoddard (1987) and the solutions from the computational wind turbine software package Bladed, were used to predict the performance of the blade for a particular set of experimental conditions. The solutions of Eggleston and Stoddard, where only the first order dynamic equations of a simplified blade model are solved, underestimate the root flapwise moment, streamwise blade tip deflection and lead-lag tip deflection, but gave fairly accurate predictions for the blade lead lag moment. The turbine's structural dynamics within Bladed used a more accurate model of the blade and solved the structural dynamic equations by implementing modal analysis. The results gave root flapwise moment of the same order as those determined from the measurements, close agreement with the measured lead-lag moment, a slight underprediction of the flapwise tip deflection and large under prediction of the lead-lag tip deflection. The under prediction of the lead-lag deflection by both methods is likely to be due to the uncoupling of the lead-lag and flapping motions, and the unusual shape of the blade close to its root connection.