Multi-Objective Structural Optimization of Wind Turbine Tower and Foundation Systems using Isight: A Process Automation and Design Exploration Software

1. Abstract Currently, wind turbine towers are being built to increasing heights in order to tap into higher and more consistent winds available at these heights. Additionally, increasingly large wind turbines are being deployed atop these towers in order to capture more energy. This trend of larger turbines being deployed at greater heights makes obtaining the most efficient and safe, or optimal, designs of the structures that support them ever more important. Towards this goal, the present work formulates the design of an integral wind turbine tower and foundation system as a multi-objective optimization problem using the process automation and design exploration software Isight. Specifically, a new general metamodeling and optimization methodology for highly non-linear multi-objective problems is applied to the design of a 130-m hybrid pre-cast concrete and tubular steel tower supported by a gravity based foundation by: 1) applying Design of Experiments (DOE) techniques to identify response drivers and collect experimental data 2) developing metamodels of responses from this experimental data using Elliptical Basis Functions (EBF) 3) performing Multi-Objective Genetic Algorithm (MOGA) optimization to generate a Pareto Front of Pareto-Optimal designs 4) performing Multi-Objective Gradient Based (MOGB) optimization from select Pareto-Optimal designs 5) validating the resulting optimal designs by full finite element simulations 6) performing Reliability Based Design Optimization (RBDO) to ensure reliability of the final design. To analyze the support structure and obtain response information, a geometrically non-linear transient finite element analysis is performed at each DOE sample point using Abaqus. This problem is multi-objective in nature with an ideal design being one that both minimizes costs and maximizes structural stiffness to reduce vibrational wear on turbine components. Therefore, a composite objective function is developed that minimizes raw material costs and deflections. The physical dimensions of the wind turbine support structure are taken as design variables. Design requirements such as yielding or fracture of the tower wall, limits on tower top deflection and rotation, limits on the natural frequency of the tower and foundation system, bearing capacity of the foundation, stiffness of the foundation, and foundation overturning moment are evaluated using Isight’s calculator component and converted to optimization constraints by Isight. Results obtained show that the proposed methodology yields significant value to the design engineer in terms of improved designs and enhanced understanding of the problem. Additionally, this work shows how Isight can be used to overcome many of the practical barriers associated with applying advanced optimization methods at the detailed design level.