A simplified method for structural and fatigue analyses of wind turbine support structures

Abstract

Long-term structural reliability of wind turbines relies on massive time-history structural analyses to estimate adequately the probability of failure. This is commonly performed on specialized software, such as FAST or HAWC2, which allows flexibility for a wide variety of analyses. Nevertheless, as the number of analyses needed to estimate the various possible outcomes from diverse environmental or operational conditions is large, these might still be computationally expensive. This work presents a simplified approach to estimate reliable results in a manner computationally economical. Two cases are covered: one for structural response computation, and one for fatigue damage evaluation. The approach relies basically on translation models for non-Gaussian process simulation, and few analyses executed on specialized software. A numerical example is carried out from the proposed approach, and the results are compared with quantities measured from analyses performed in FAST. Adequate resemblances are observed from the results, which makes the method attractive for its application in practice to be used in preliminary structural analyses and prompt reliability analyses. Comments on assumptions and methods employed for the analyses are made, and appendices with useful algorithms are included.