ICCMEEE

Christian Masson Professor
B.Eng., M.Sc.A.
(Ecole polytechnique, Montréal),
Ph.D. (McGill University) Mechanical Engineering Department
Ecole de Technolgie Superieure de Montréal, Canada

On the use of the k − ε turbulence model for atmospheric
boundary layer ﬂow and wind turbine wakes (1/2)

The RANS/k−ε approach has established itself as the method of choice for carrying out flow simulations of the atmospheric boundary layer. Its popularity is primarily due to its ability to provide reasonable estimates of turbulence properties under a wide range of conditions, from wall-bounded to free-shear flows, at relatively low computational cost. The research group of Prof. Christian Masson has worked specifically on improving the accuracy of the k −ε model and on adapting it for use in forested areas and in the wake of operating turbines.
Concerning the flow in forested regions, Dalpé and Masson have proposed appropriate boundary conditions and carried out 1D simulations of fully-developed ﬂow within and above forests. In analogy with flow through porous media, source terms are added to the momentum and k−ε transport equations that are functions of characteristic aerodynamic properties (ie leaf area density and drag coefficient) to account for loss of momentum and increased turbulence. Although k is slightly underestimated within the canopy, predicted velocity proﬁles within and above the forest agree with measurements.
For flow in the near wake of a turbine, El Kasmi and Masson have proposed to use the Chen and Kim modified k − ε model with the stability-dependent buoyancy terms proposed by Alinot and Masson. The action of the wind turbine on the flow is approximated using the actuator disc concept; the rotor is represented by a permeable disc that performs the same average work on the flow and imparts the same angular momentum. Simulations of the wake flow behind the Danwin 180-kW turbine for various thermal stratiﬁcations are in excellent agreement with measurements and much better than the standard k − ε model which typically underestimates the velocity defect.

1 | 2

Brief Bio | Abstract