Optimization of Wind Turbine Blade Fatigue Test Design

The present thesis presents a methodology for efficiently obtaining an optimal blade fatigue test setup that achieves a best fit to a target bending moment distribution. The methodology combines a custom finite element method-based, time domain simulation tool with a genetic algorithm optimizer to obtain a configuration of test parameters (e.g. added mass amounts or locations) that minimizes the error between the simulated and target bending moments subject to testing constraints (e.g. limit on total added mass). A wind turbine blade is fabricated from composite materials and designed with a circular shape at the root, which is blended into airfoil shapes towards the tip. As blade failures are very costly, comprehensive testing is employed to minimize the risk of structural failures. Full-scale structural testing is the primary method used to validate a blade’s performance according to the international standard (IEC-61400-23, 2014). The goal of testing is to validate the structural parts and materials of the blade, which must withstand the ultimate loads in extreme conditions, and show high reliability under long- term fatigue loads experienced in normal service conditions. Full-scale wind blade testing is very time-consuming and costly. This scenario, coupled with the continuous growth in the size of wind blades, has created new challenges for existing wind blade facilities as they attempt to perform efficient, economic tests of large, complex wind blades. In this work, a simulation of a wind blade test is combined with a genetic algorithm optimizer to select test parameters, such as added mass amounts and locations, which produce the best fit to the target bending moment distribution for a fatigue test. The tool is applied to a test fatigue case being conducted at the University of Maine’s Advanced Structures and Composites Center (ASCC), an ISO 17025 accredited laboratory. Test results partially validate the simulation tool and show the promise of the proposed optimization methodology