This study demonstrates lifetime assessments of a wind turbine structural component as the tower and rotating components as the main bearings using the controller data, measurements, and no blade design information, representing a realistic scenario for operating turbines. A tower bottom virtual load sensor framework based on neural networks was proposed using different input combinations to replace the tower sensor. The estimated lifetime was considerably longer than the design lifetime.

Wind turbine extreme response estimation based on statistical extrapolation necessitates using a small number of simulations to calculate a low exceedance probability. This is a challenging task especially if we require small prediction error. We propose the use of a Gaussian mixture model as it is capable of estimating a low exceedance probability with minor bias error, even with limited simulation data, having flexibility in modeling the distributions of varying response variables.