Turbulence and wind energy
Credits : 2 ECTS
Duration : 21 hours
Semester : S9
Persons in charge:
Emmanuel Plaut, professor
Michaël Hölling, associate researcher at ForWind, universität Oldenburg (with Erasmus+ support)
Keywords: deterministic models, stochastic models
Prerequisites: fluid dynamics elementary course
Objective: learn advanced turbulence modeling, with a focus on wind energy systems, that always operate under turbulent conditions.
Program and contents:
Advanced turbulence modelling is introduced, with deterministic then stochastic models.
- Reynolds Averaged Navier-Stokes (RANS) models
In order to motivate the need for `sophisticated' RANS models, the weaknesses of the `standard' k – ε model are evidenced, by comparisons with recent direct numerical simulations (DNS). Improved k – ε models or alternative RANS models are then introduced.
Large Eddy Simulations (LES)
The failures of RANS models to describe some turbulent flows with large dynamic eddies is exemplified. LES is then introduced.
Hybrid RANS – LES methodsAn introduction to these methods concludes this first part of the module.
approach and models
We study turbulence modelling for computational fluid dynamics (CFD) by focusing on the RANS approach and models. The RANS approach is indeed closely linked to the statistical theory of turbulence, which is quite relevant. The RANS approach is basically interesting to characterize and gain some understanding on turbulence phenomena. Finally, RANS models are still, today, the preferred choice for engineering studies. After a brief discussion of Reynolds stress equations and models, the focus is on eddy-viscosity 2-equations models, namely, the k – ε and k – ω models.
Using a problem-based learning approach, studies of the Direct Numerical Simulations (DNS) database of Lee & Moser (2015) will be performed with Matlab.
Stochastic models, especially, in the context of wind energy
This second part of the module is focused on the wind energy sciences. Knowing that the wind in the atmospheric boundary layer is always turbulent, the small scale statistics of turbulent flows is reviewed, with a focus on the intermittency phenomena andextreme events like`wind gusts'. The principles of wind energy conversion and the aerodynamicsof wind turbines are also presented. The International Electrotechnical Commission (IEC) standards to evaluate power curves and annual energy production of a given wind turbine is introduced. The limitations of this approach are discussed and more accurate stochastic approaches are introduced, which lead for instance to the notion of `Langevin power curves'.
The students will use the statistical computing software R to study various `turbulence experimental databases'.
Please check the web page of this module on http://emmanuelplaut.perso.univ-lorraine.fr/turbuwe : itsketchesthe planning of this module, gives instructions and files, etc...In particular, the students will use the statistical computing software R on their laptop, to study various `turbulence data bases'.
Description and operational vocabulary
Hybrid RANS - LES methods. Be aware of current research in the domain of CFD of turbulent flows
The RANS approach - the LES approach - Strengths and weaknesses of each approach
The notion of intermittency and extreme events
The RANS approach - the LES approach
The RANS approach - the LES approach
Analyze a velocity time series: be able to extract mean and standard deviation values, a PDF of the velocity and of the velocity increments, etc...
Analyze a wind time series: construct the IEC power curve
Choose a good approach - model to solve numerically a given turbulent flow problem