Reliability-Based Design Optimization of Wind Turbine Systems

Abstract

Wind turbine reliability is considered as one of the most critical factors that affects the wind energy cost and is investigated tremendously nowadays. Design optimization techniques are implemented in wind energy industry to obtain reliable wind turbine systems, e.g., hub, blade, gearbox, and bearing, while satisfying design constraints. However, traditional deterministic design optimization (DDO) methods do not consider realistic uncertain situations during the life span of wind turbines potentially causing premature failures of wind turbine systems. In contrast, reliability-based design optimization (RBDO) is aimed to design reliable and cost-effective wind turbine systems considering various uncertainties, e.g., load uncertainty, manufacturing uncertainty, and material uncertainty, at early design stage. This chapter elaborates the RBDO procedure for wind turbine systems considering uncertainties focusing on wind load uncertainty. Using measured wind data, a dynamic wind load uncertainty model has been developed to involve both the annual wind load variation and the wind load variation in a large spatiotemporal range, for example, in different years and at different locations. Reliability analysis under wind load uncertainty applies the Monte Carlo simulation (MCS) method to simulate uncertain wind load using the developed dynamic wind load uncertainty model. This method can estimate the probability that a wind turbine system could survive 20 years of target lifespan. Finally, a sampling-based RBDO method is explained for wind turbine design problems in which the design sensitivity is often complicated, nonlinear, and implicit. One of the most important advantages of the sampling-based RBDO method is that the design sensitivity of the probabilistic constraint of complex wind turbine design problems is derived using the score function and the MCS method which avoid calculating the sensitivity of the turbine performance measure. Thus, a typical wind turbine system can be treated as a black box smoothly integrating into the RBDO procedure. One case study, RBDO of a composite wind turbine blade, is demonstrated in this chapter.