Fatigue analysis of floating wind turbine support structure applying modified stress transfer function by artificial neural network

Abstract The frequency-domain approach has been studied as a potential replacement modality for the time-domain method in fatigue analysis of offshore wind turbine structures. It is assumed that in the frequency-domain approach, the stress response spectra induced by wind and wave loads can be expressed by a stress transfer function. To obtain the stress transfer function, coupled analysis should be performed in advance. However, since the response of a wind turbine to different average wind speeds is non-linear, the stress transfer function is bound to change with wind speed. This means that repeated simulation is needed in order to calculate the stress transfer function according to wind speed change. The problem, though, is that if the number of simulations is large, prohibitively high computational and time costs probably will be incurred. In this study, to reduce the number of simulations and, at the same time, increase the accuracy of results, a correction factor of the stress transfer function induced by wind load was artificial-neural-network-approximated as a function of mean wind speed and frequency. Sensitivity analysis was conducted to determine how many sample points are required and how to select them. Also, a superposition model is proposed to improve the accuracy of the ANN model. This model is designed so that the peaks in the stress spectrum have a dominant influence on fatigue damage. In order to better simulate the correction factor around the peak, the model considering only the data of the periphery of the peaks and the model reflecting the whole data are superimposed. The total stress spectrum were calculated by summing stress spectrum induced by wind load from the ANN model and induced by inertia load from motion analysis based on linear wave theory. Numerical analysis for a 10 MW class wave and offshore wind hybrid power generation (WWHybrid) system, which is a kind of semi-submersible wind turbine platform, was performed to verify the performance of the proposed model. It was confirmed that the superposition model improves the accuracy by 20–50% compared with the single ANN model.

[1]  Karl Otto Merz A linear State-Space Model of an Offshore Wind Turbine, Implemented in the STAS Wind Power Plant Analysis Program , 2015 .

[2]  Hashim Abdul Razak,et al.  Application of artificial neural network in fixed offshore structures , 2015 .

[3]  Gaurang Panchal,et al.  Searching Most Efficient Neural Network Architecture Using Akaike's Information Criterion (AIC) , 2010 .

[4]  H. Akaike A new look at the statistical model identification , 1974 .

[5]  J. van der Tempel,et al.  Design of support structures for offshore wind turbines , 2006 .

[6]  Moo-Hyun Kim,et al.  The Dynamic Coupling Effects of a MUFOWT (Multiple Unit Floating Offshore Wind Turbine) with Partially Broken Blade , 2015 .

[7]  H. Y. Kang,et al.  Time-Domain Hydroelastic Simulations and Structural Analyses , 2013 .

[8]  Mazaheri Saeid THE USAGE OF ARTIFICIAL NEURAL NETWORKS IN HYDRODYNAMIC ANALYSIS OF FLOATING OFFSHORE PLATFORMS , 2006 .

[9]  Torgeir Moan,et al.  Linear And Nonlinear Analysis of Tension Leg Platform Wind Turbines , 2012 .

[10]  P. Ferrant,et al.  Comparison of Time and Frequency Domain Simulations of an Offshore Floating Wind Turbine , 2011 .

[11]  Michael Muskulus,et al.  Simplified fatigue load assessment in offshore wind turbine structural analysis , 2016 .

[12]  T. Moan,et al.  Frequency Versus Time Domain Fatigue Analysis of a Semisubmersible Wind Turbine Tower , 2015 .

[13]  Michael Muskulus,et al.  Mooring System Optimization for Floating Wind Turbines using Frequency Domain Analysis , 2012 .

[14]  Gaurang Panchal,et al.  Behaviour Analysis of Multilayer Perceptrons with Multiple Hidden Neurons and Hidden Layers , 2011 .

[15]  Lance Manuel,et al.  Comparing Estimates of Wind Turbine Fatigue Loads Using Time-Domain and Spectral Methods , 2007 .