Multiobjective wind turbine gearbox design optimization to reduce component damage risk during grid faults

Abstract Cost-intensive downtimes of wind turbines (WT) are often caused by damage to gearbox components. Particularly components of the high-speed-shaft (HSS) are prone to faults. One reason for gearbox component damages can be electromagnetic generator torque excitations caused by grid faults. These torque excitations are transferred from the generator to HSS components and lead to transient load changes. Thus, grid faults may increase the damage risk of the HSS-IMS gear stage and HSS bearings. As shown in previous work, adjusting the micro-geometry of the HSS components can reduce the damage risk during grid faults. In this paper, a method to optimize the reduction of damage risk of the HSS gear wheel and cylindrical roller bearing during grid faults is presented. The method is based on a multiobjective optimization of the micro-geometry of the HSS gear wheel and roller bearing. Only the micro-geometry is optimized as the design of the macro-geometry is dominated by the nominal operation of the WT. The adjusted design parameters are radial bearing clearance, gear stage profile shift and profile modifications. Considered damage types are scuffing, micropitting, flank fracture, tooth breakage (gear wheel) and smearing (cylindrical roller bearing). First, a test plan is defined using Latin Hypercube Sampling to cover the huge parameter space efficiently. Second, the resulting damage risks are determined using multibody simulation of a WT drivetrain during a grid fault. Third, a surrogate model is created for each damage type using regression methods. Fourth, a multiobjective optimization is performed using the surrogate models. Finally, the optimized design adjustments regarding grid faults are derived through analysis of the pareto front. For the investigated research drivetrain the method reduces the risk of each considered damage type of the HSS components during grid faults by at least 20%.