Effective foundation damping prediction of monopile-supported offshore wind turbines based on integrated fitting equation and PSO–SVM algorithm

Abstract

Foundation damping affects the fatigue life of offshore wind turbines. However, the effective damping ratio determined using a finite element analysis is not always consistent with the input. A series of simplified finite element models were established using realistic structural properties for a monopile-supported wind turbine to analyze the influence of wall thickness, pile embedment length, structure density, and input damping ratio on the calculated effective damping ratio. An empirical fitting equation was proposed using a power function based on the comprehensive laws between various factors and the effective value. An integrated prediction approach combining this equation with a hybrid support vector machine and particle swarm optimization algorithm was subsequently described and its results compared with those of other intelligent algorithms. The results showed that the effective damping ratio was much smaller than the input value, with the former reaching a maximum of only 15% of the latter. The integrated prediction approach was shown to strike an equilibrium between computing efficiency and forecast accuracy. Considering the soil–structure interaction and elastoplastic behavior of soil only increased the effective damping ratio to 35% of the input value, and also obfuscated the roles of the various factors in the effective damping.