A Sparse ANN-based Fatigue Estimation for Wind Turbine Control based on NMPC

Paper i proceeding, 2018

In this paper, an Artificial Neural Network (ANN) methodology for estimating the tower fatigue of a wind turbine (WT) is developed. Specifically, the fatigue is estimated using an ANN that receives frequency-domain measurements of the tower base position and velocity. The frequencies are selected using an algorithm that detects the most relevant values and generates a sampling grid. The complexity of the ANN-based fatigue estimation is analysed to study the viability of its deployment in a real-time Nonlinear Model Predictive Control (NMPC) formulation.