Surrogate Model Monte Carlo simulation for stochastic flutter analysis of wind turbine blades

Abstract With the larger size of wind turbine blades, coupled-mode flutter of long, flexible rotating blades has become an issue that cannot be neglected. Recent studies have shown that flutter speed can be predicted accurately using a three-dimensional, finite-element model by deterministic aeroelastic analysis. However, modeling uncertainty is an important, non-negligible issue as it influences probability of flutter failure. The stochastic flutter problem is investigated herein to understand the influence of uncertainty in selected input characteristics: (i) aerodynamic loads and (ii) blade structural properties. Physical Model Monte Carlo simulation is usually utilized for stochastic flutter analysis. Surrogate Model Monte Carlo simulation is proposed to obtain the solution more efficiently in terms of computing time. Surrogate models include: least squares fit and collocation methods. A nonlinear relationship between the input and output random variables is incorporated for least squares fit. For collocation methods, Lagrange polynomials are employed to perform random simulations. Results indicate that surrogate models efficiently facilitate the stochastic flutter analysis with acceptable errors.