Publications

A wind turbine rotor should capture as much wind power as possible. However, the more energy extracted from the wind, the larger the aerodynamic forces induced on the rotor blades, which thus requires a stronger blade structure. Consequently, one of the challenges during the design of an optimal rotor blade is to find the right balance between its aerodynamic performance and its structural behavior. This contribution presents a coupled multi-objective shape and topology optimization framework for the aero-structural design of Horizontal-Axis Wind Turbine (HAWT) rotor blades. It aims at optimizing simultaneously the blade outer shape and the interior layout from aerodynamic and structural requirements. Accordingly, a multi-objective optimization model is developed, based on the weighted sum method, in which the rotor power coefficient and the blade structural compliance are considered as two sub-objectives. In order to limit the computational demand, the power coefficient and structural compliance are efficiently evaluated via the Blade Element Momentum (BEM) and beam Finite Element Method (FEM) models, respectively. The shape design variables are characterized by the control point locations of Non-Uniform Rational B-Splines (NURBS) that govern the blade geometry. The topology design variables are represented by the relative densities assigned to the finite elements modeling the blade cross-sections, as common for the Simplified Isotropic Material with Penalization (SIMP) method. The coupled optimization model is solved by sequentially performing shape and topology optimizations with different weighting factors for the sub-objectives. In order to further improve the computational efficiency of the optimization model, the sensitivities for shape and topology optimization are derived in closed form, and incorporated within a gradient-based optimization algorithm. The presented optimization framework is used to optimize the National Renewable Energy Laboratory (NREL) 5MW wind turbine rotor blades. The optimized design of the rotor blades achieves a higher power coefficient and a lower structural compliance than the reference design, which clearly indicates the practical applicability of the coupled optimization framework.

Door de nog altijd toenemende rekenkracht van computers komen op vele werkterreinen steeds geavanceerdere modellen beschikbaar. Voor de bepaling van windbelastingen op gebouwen wordt echter nog altijd gebruikt gemaakt van oude methoden: normbladen en windtunnelonderzoek. Uiteraard zijn die doorontwikkeld maar een doorbraak van een rekenkundig alternatief - met Computational Fluid Dynamics (CFD) - is uitgebleven. Een belangrijke reden is dat CFD nog altijd zeer veel rekenkracht eist. En dan nog is het niet zonder meer bruikbaar

In the past decades, numerous studies were performed to investigate the effects of mechanical ventilation system (MV) on aerosol concentration in indoor environments. However, few attention was given to the prediction of aerosol concentration in multi-zone indoor spaces, where the distribution and migration of aerosol particles are significantly influenced by airflow patterns, room dimensions and geometries, supply and exhaust locations, internal partitions, and so on. This study aims at analyzing the aerosol concentration and airflow pattern in a players' dressing room (composed of three-connected spaces) of a stadium by means of Computational Fluid Dynamcs (CFD) simulations. The simulations are performed with Reynolds-averaged Navier-Stokes (RANS) approach by means of passive scalar (PS), drift flux (DF), and Eulerian-Lagrangian (EL) methods. The aerosol concentration (c) from field measurements, previously performed in the same indoor environment, is then compared to simulated (CFD) data. The comparison shows that the deviation between the measured and simulated data is of about 37% for the PS method, 24.3% for the DF method, and 18% for the EL method.

Numerically reproducing wind tunnel experimental tests of flow behavior around a pitching airfoil is a challenge, especially under the occurrence of dynamic stall at relatively high angles of attack. This not only requires the application of advanced turbulence models, but also asks for examining the influence of the most relevant model parameters in detail. This contribution presents the results of an extensive computational study of the unsteady flow around a pitching NACA0012 airfoil at a Reynolds number of 1.35 × 105, as obtained by first performing 2D Unsteady Reynolds-Averaged Navier-Stokes (URANS) simulations whereby the flow characteristics are simulated by the Transition SST turbulence model. The influence of a large number of parameters on the numerical results is investigated, namely the blockage ratio, computational grid resolution and y+ value, time step size, inlet freestream turbulence properties, airfoil trailing edge detailing and turbulence model. Integral aerodynamic force coefficients and detailed flow patterns are analyzed and compared with measurements from wind tunnel experiments presented in the literature. For the best-performing parameters, an adequate agreement with the experimental tests is obtained for the upstroke phase, while for the downstroke phase some differences appear. These differences are investigated in more detail by considering the results from a 2.5D Large Eddy Simulation (LES), which provides deep and complementary insights into the flow behavior during dynamic stall at the selected Reynolds number.

Topology optimization is a widely-used technique for finding the most favorable, internal structural lay-out with a minimal weight under the specific loading and boundary conditions applied[1,2]. Accordingly, within a finite element setting this technique searches for the optimal relative density of a fixed, discretized spatial domain representing the actual structure. To enable more diversity within the design domain and to enlarge the search space of optimal structural configurations, in the present work a coupled method for topology optimization and shape optimization is proposed. The method incorporates the shape design variables into a SIMP (Simplified Isotropic Material with Penalization) topology optimization formulation, whereby the shape and topology optimization steps are performed in a sequential manner. The computational efficiency of the method is warranted by using Non-Uniform Rational B-Splines (NURBS) for describing the outer shape of the design domain, and by combining gradient-based optimization solvers with analytically derived shape and topology sensitivities. The coupled method has been implemented in a finite element framework to analyze 2D, 2.5D, and 3D structural design problems. The results of representative case studies clearly show that the features of the design domain can have a large influence on the final topology calculated. Additionally, the optimization sequence in the coupled method may affect the path followed within the design space; however, this typically only has a minor effect on the final computational result. [1] Bendsoe, M.P., Sigmund, O. (2004). Topology optimization by distribution of isotropic Material. In Topology Optimization (pp. 1-69). Springer, Berlin, Heidelberg. [2] Sigmund, O., Maute, K. (2013). Topology optimization approaches. Structural and Multidisciplinary Optimization, 48(6), 1031-1055.