Unsteady Simulations for Industrial Flows: LES, DES, hybrid LES-RANS and URANS

Unsteady Simulations:
LES, DES, hybrid LES/RANS and Machine Learning
2, 4, 6 December 2024

course material

course language

registration & fee

OTHER COURSES:

Large-Eddy Simulation & Detached-Eddy Simulations - How to use an in-House Fortran source code

	A three-day online course on Unsteady Simulations: LES, DES, hybrid LES/RANS and Machine Learning
LES is suitable for bluff-body flows or flows at low Reynolds numbers. To extend LES to cover industrial flows at high Reynolds numbers, new approaches (hybrid LES-RANS, DES, URANS, SAS, PANS, PITM) must be used. They are all based on a mix of LES and RANS. The course will give an introduction to LES and these new methods. Day 3, an introduction will be given on how to use Machine Learning for improving underlying RANS turbulence models [41, 42] and wall-functions [38-40]. LES, or any of the new approaches, is the first step when performing accurate CAA (Computational Aero-Acoustics) The lectures will be given on-line (Live) using Zoom. During the workshops, the participants will get supervision in a joint Zoom room which will enable participants to learn from each others questions. Part of the supervision may also be given in individual break-out Zoom rooms. Zoom Lectures will be given in the mornings; in the afternoons there will be workshops using Python (recommended) , Matlab or Octave. The participants must have a PC/Mac/Desktop with one of these software installed. In the workshops, the participants will use Python/Matlab/Octave for analyzing SGS models, SAS, PANS, DES and DDES. Scripts will be use for generating isotropic and anistropic (non-isotropic) synthetic turbulent fluctuations for inlet boundary conditions and embedded LES. These will be the topics Day 1 and 2. Day 3 is devoted to Machine Learning. Neural network and KDTRee in Python's Pytorch will be used to inmprove RANS turbulence models and wall functions. The number of participants is limited to 16.
BACKGROUND The development of computers and Computational Fluid Dynamics (CFD) has made the numerical simulation of complex fluid flow, combustion, aero-acoustics and heat transfer problems possible. Turbulent flow in three-dimensional, complex geometries -- unsteady or steady -- can be dealt with. Presently CFD methods can replace, or complement, many experimental methods; we can use a numerical wind tunnel instead of an experimental one. Today, most CFD simulations are carried out with traditional RANS (Reynolds-Averaged Navier-Stokes). In RANS, we split the flow variables into one time-averaged (mean) part and one turbulent part. The latter is modelled with a turbulence model such as k-eps or Reynolds Stress Model. For many flows it is not appropriate to use RANS, since the turbulent part can be very large and of the same order as the mean. Examples are unsteady flow in general, wake flows or flows with large separation. For this type of flows, it is more appropriate to use Large Eddy Simulation (LES). In order to extend LES to high Reynolds number flows new methods have recently been developed. These are called DES (Detached Eddy Simulation), URANS (Unsteady RANS), PITM (Partially Integrated Transport Model), PANS (Partially Averaged Navier-Stokes) or Hybrid LES-RANS. They are all unsteady methods and they are a mixture of LES and RANS.. In aero-acoustics the noise is generated by turbulence. The best way to accurately predict large-scale turbulence is to carry out an unsteady simulation of the flow field (i.e. LES, DES, hybrid LES-RANS or URANS). After that the noise is predicted separately in CAA (Computational Aero-Acoustics) in which the large-scale turbulence is used in analogy methods based on Lighthill, Kirchhoff or Ffowcs Williams. In LES, DES, URANS and Hybrid LES-RANS the large-scale part of the turbulence is solved for by the discretized equations whereas the small-scale turbulence is modelled. The definition of ''large-scale'' varies in the different methods. Furthermore, the limit between ''large-scale'' and ''small-scale' is often not well defined. Since turbulence is three-dimensional and unsteady, it means that in all the methods the simulations must always be carried out as 3D, unsteady simulations. THE COURSE The course will give an introduction to LES, DES, hybrid LES-RANS and unsteady RANS. During the lectures we will discuss the theory and during the workshops We will use simple Python (recommended) , Matlab or Octave. programs to gain detailed insight in various numerical and modelling aspects. The participants must have access to a PC/Mac/Desktop with on of these three programs installed. The number of participant is limited to 16 We will address questions like: how should I make my mesh? why should I in LES use a dissipative discretization scheme? is it necessary to used central differencing in DES and URANS? what is the difference between LES and unsteady RANS? what turbulence models can I use in DES and unsteady RANS? what f_k values are suitable in PANS? to enhance numerical stability, can a turbulence model with high dissipation be used? how do I prescribe inlet boundary conditions? inlet boundary conditions: can I use steady conditions? which is best, synthesized turbulence or a pre-cursor DNS? Download Python/Matlab/Octave files for generating synhetic inlet fluctuations How can Machine Learning be used for improving turbulence models and wall functions [38-42}? In the workshop, we will learn how to interpretate results from an unsteady simulation. We will evaluate and compare the two types of turbulent stresses, i.e. the resolved stresses and the modelled stresses (more detail at the bottom). When doing LES-URANS/DES, you have to ask yourself similar questions as when doing measurements: when is the flow fully developed so that I can start time-averaging? for how long time do I need to time-average? is it enough if I get accurate mean flow or do I also need accurate resolved turbulent stresses? how do I estimate the quality of my LES or hybrid LES-RANS? Spectra? 2-point correlations? SGS dissipation? For more info, see QLES 2009 and references when and how should I use the different Machine Learning tools availalle in Python (SVR, kNN, pytorch)? The most important drawback/bottleneck of LES is the requirement to use very fine grid near walls. The grid must be fine in all directions, not only the wall-normal direction. Much of the research on LES is today focused in getting around this bottleneck. One approach is hybrid LES-RANS. In this method RANS is used near walls and LES is used in the remaining part of the domain. In the afternoon of Day 2 some hybrid LES-RANS methods (including the SAS model) will be presented and discussed. Inlet boundary conditions are much more difficult in LES and hybrid LES-RANS than in RANS. In RANS it is sufficient to prescribe time-averaged profiles. In LES and hybrid LES-RANS unsteady, turbulent fluctuations must be supplied. One alternative is to do a pre-cursor DNS and store data at a cross-sectional plane on disc which can be read in the subsequent LES or hybrid LES-RANS simulation. Another alternative is to use synthesized turbulence. The participants will during the last workshop (Day 3) have the opportunity to learn how to to create synthesized turbulence using Python/Matlab/Octave. As an alternative, the participants can choose to instead learn how to use Machine Learning for improving RANS models and wall functions [38-41]. The participants will use Support Vector Regression (SVR), Nearest Neighbour (kNN) and Neural Network (pytorch), all available in Pyhton. Synthetic turbulent fluctuations are also important in Embedded LES in which an LES/DES region is embedded in a steady or unsteady RANS region. Usually the upstream region is a RANS region and the downstream region is LES or DES. Turbulent fluctuations must be added at the interface between RANS and LES to ensure a rapid transition from steady RANS where all turbulence is modelled to LES/DES where most of the turbulence is resolved. OBJECT The participants will be given an introduction to LES, DES, hybrid LES-RANS and unsteady RANS. We expect many participants to be first-year PhD students or users of in-house CFD codes or commercial CFD packages for traditional RANS simulations. This course will give the required knowledge to do CFD predictions using also unsteady methods. PARTiCiPANTS We believe that the course will be useful for engineers and PhD students working with problems including pure fluid flow, aero-acoustics, combustion and heat transfer in industry as well as at universities. The participants must have access to a PC/Mac/Desktop with Python (recommended), Matlab or Octave installed. The number of participants is limited to 16. LECTURER The lecturer at the course (both during lectures and workshops) will be Prof. Lars Davidson, Chalmers University of Technology. homepage COURSE MATERiAL L. Davidson, eBook (Opens a PDF file, Chapters 18 - 27) COURSE LANGUAGE The course material is in English and the lectures will be given in English. LOCATiON The course will be held 2, 4, 6 December 2024 on Zoom. and is organized by Flowsim AB. REGiSTRATiON Registration form should be submitted no later than November 8, 2024. The price is 14,700 SEK (excl. VAT) which includes course material. No refunding after November 8. The number of participants is limited to 16. registration form PROGRAM DAY 1, Monday (13.00 -- 21.00, Swedish time) Introduction to Large Eddy Simulation Filtering of the equations; discretization convection schemes for LES, SGS models. Workshop: interpretation of results from a LES and unsteady RANS. Time-averaging, evaluation of various forms of turbulent stresses etc. Tueday (no teaching). Participants can work on assignments DAY 2, Wednesay (13.00 -- 21.00, Swedish time) Dynamic SGS models, scale-similarity models, transport equations for SGS stresses Introduction to DES, URANS, PANS and SAS Introduction to hybrid LES-RANS Workshop continued: explicit filtering, SGS models, spectra, two-point correlations, viscous and SGS dissipation, scale-similarity models, DES, DDES and SAS Thursday (no teaching). Participants can work on assignments DAY 3, Friday (13.00 -- 21.00, Swedish time) Introduction to Machine Learning for improving turbulence models [40-42] Workshop: Machine Learning for turbulence modeling (Neural Network in Pytorch) using Python Introduction to Machine Learning for improving wall functions [38-40] Workshop: Machine Learning for turbulence modeling (Neural Network and KDTree in Pytorch) using Python PUBLiCATiONS L. Davidson, "How to generate synthetic turbulent inlet fluctuations" Download Python/Matlab scripts L. Davidson and S.-H. Peng "A Hybrid LES-RANS Model Based on a One-Equation SGS Model and a Two-Equation k-omega Model", The Second International Symp. on Turbulence and Shear Flow Phenomena, Eds: E. Lindborg, A. Johansson, J. Eaton, J. Humphrey, N. Kasagi, M. Leschziner, M. Sommerfeld, Vol. 2, pp. 175-180, Stockholm, 2001. View PDF file L. DAVIDSON Hybrid LES-RANS: A Combination of a One-Equation SGS Model and a k-omega Model for Predicting Recirculating Flows" ECCOMAS Computational Fluid Dynamics 2001 Conference, Swansea, UK, 2001. View PDF file L. Davidson and S.-H. Peng "Hybrid LES-RANS: A one-equation SGS Model combined with a k-omega model for predicting recirculating flows", Int. J. Num. Meth. in Fluids, Vol. 43, pp. 1003-1018, 2003. S. Dahlström and L. Davidson "Hybrid RANS/LES employing Interface Condition with Turbulent Structure", Dept. of Thermo and Fluid Dynamics, Chalmers University of Technology, Report, Göteborg, Sweden, 2003 View PDF file S. Dahlström and L. Davidson "Hybrid RANS-LES with Additional Conditions at the Matching Region", Turbulence Heat and Mass Transfer 4, pp. 689-696, K. Hanjalic, Y. Nagano and M.J. Tummers (eds.), begell house, inc., New York, Wallingford (UK), 2003. View PDF file L. Davidson and S. Dahlström "Hybrid RANS-LES: an Approach to make LES Applicable at High Reynolds Number", CHT-04: Advances in Computational Heat Transfer III, Keynote Lecture, G. de Vahl Davis and E. Leonardi (eds.), Norway, April 2004 (updated version in International Journal of Computational Fluid Dynamics, Vol. 19, No. 6, pp 415-427 2005, see below). L. Davidson and S. Dahlström "Hybrid RANS-LES: an Approach to make LES Applicable at High Reynolds Number", Int. J. of Comp. Fluid Dynamics Vol. 19, No. 6, pp 415-427, 2005. L. Davidson and M. Billson, "Hybrid LES/RANS Using Synthesized Turbulence for Forcing at the Interface", ECCOMAS 2004, P. Neittaanmaki, T. Rossi, S. Korotov, E. Onate, J. Periaux, and D. Knorzer (eds.), July 24-28, Finland. View PDF file L. Davidson and S. Dahlström "Hybrid LES-RANS: Computation of the Flow Around a Three-Dimensional Hill", Engineering Turbulence Modelling and Measurements - ETMM6, Sardinia, Italy, May 23-25, 2005. View PDF file C. Wollblad and L. Davidson "POD based reconstruction of subgrid stresses for wall bounded flows using neural networks", 5th International Symposium on Turbulence, Heat and Mass Transfer, Dubrovnik, Croatia, September 25-29, 2006. View PDF file C. Wollblad and L. Davidson "POD based reconstruction of subgrid stresses for wall bounded flows using neural networks", Flow, Turbulence and Combustion, Vol. 81, No. 1-2, pp. 77-96, 2008. Go to journal L. Davidson "Evaluation of the SST-SAS Model: Channel Flow, Asymmetric Diffuser and Axi-symmetric Hill", ECCOMAS CFD 2006, September 5-8, 2006, Egmond aan Zee, The Netherlands, 2006. View PDF file L. Davidson "Transport Equations in Incompressible URANS and LES", Rept. 2006/01, Division of Fluid Dynamics, Dept. of Applied Mechanics, Dynamics, Chalmers University of Technology Göteborg, 2006. View PDF file L. Davidson "Using Isotropic Synthetic Fluctuations as Inlet Boundary Conditions for Unsteady Simulations" Advances and Applications in Fluid Mechanics, Vol. 1(1), pp. 1-35, 2007. L. Davidson " Hybrid LES-RANS: Estimating Resolution Requirements Using Two-Point Correlations and Spectra", ERCOFTAC Bullentin, Special Issue on Wall modelling in LES, pp. 19--24, March, 2007. (corrected) View PDF file L. Davidson A dissipative scale-similarity model, DLES7: Direct and Large-Eddy Simulations 7, 8-10 Sept 2008, Trieste, 2008. View PDF file L. Davidson "Hybrid LES-RANS: back scatter from a scale-similarity model used as forcing", Phil. Trans. of the Royal Society A, Vol. 367, Issue 1899, pp. 2905-2915, 2009. J. Ask and L. Davidson "A Numerical Investigation of the Flow Past a Generic Mirror and its Impact on Sound Generation", Journal of Fluids Engineering, vol.131, number 061102, 2009. L. Davidson "Large Eddy Simulations: how to evaluate resolution", International Journal of Heat and Fluid Flow, Vol. 30(5), pp. 1016-1025, 2009. Get article at publisher's www page View PDF file of manuscript L. Davidson Fluid mechanics, turbulent flow and turbulence modeling, course material in MSc courses, Division of Fluid Dynamics, Dept. of Applied Mechanics, Chalmers University of Technology, Göteborg, 2010 View PDF file J. Ma, S.-H. Peng, L. Davidson and F. Wang A Low Reynolds Number Partially-Averaged Navier-Stokes Model for Turbulence, 8th International ERCOFTAC Symposium on Engineering Turbulence, Modeling and Measurements, Marseille, France, 9-11 June, 2010. View PDF file J. Ask, L. Davidson Flow and Dipole source evaluation of a generic SUV,J. Fluids Eng., Vol. 132, No. 051111, 2010. J. Ask, L. Davidson A Numerical Investigation of the Flow Past a Generic Side Mirror and its Impact on Sound Generation,J. Fluids Eng., Vol. 131, No. 061102, 2009. L. Davidson "How to estimate the resolution of an LES of recirculating flow", Quality and Reliability of Large-Eddy Simulations II, Ercoftac Series, Springer, 2010. View PDF file Ma, J.M, S.-H. Peng, L. Davidson and F.J. Wang A low Reynolds number variant of partially-averaged Navier-Stokes model for turbulence, Int. J. Heat Fluid Flow, Vol. 32, pp. 652-669, 2011. Get article at publisher's www page View PDF file of manuscript L. Davidson and S.-H. Peng "Embedded LES Using PANS", I6th AIAA Theoretical Fluid Mechanics Conference, AIAA paper 2011-3108, 27 - 30 Jun 2011, Honolulu, Hawaii. View PDF file L. Davidson "A New Approach of Zonal Hybrid RANS-LES Based on a Two-equation k-eps Model", ETMM9: International ERCOFTAC Symposium on Turbulence Modelling and Measurements, Thessaloniki, Greece, 2012 View PDF file L. Davidson "Large Eddy Simulation of Heat Transfer in Boundary Layer and Backstep Flow Using PANS", Turbulence, Heat and Mass Transfer 7 Hanjalic, Y. Nagano, D. Borello and S. Jakirlic (Editors), Begell House, Inc., 2012 View PDF file L. Davidson and S.-H. Peng "Embedded Large-Eddy Simulation Using the Partially Averaged Navier-Stokes Model", AIAA J, Vol. 51(5), pp. 1066-1079, 2013. View PDF file L. Davidson "Backscatter from a scale-similarity model: embedded LES of channel flow, developing boundary layer flow and backstep flow", 8th International Symposium on turbulence and shear flow phenomena (TSFP8), Poitiers, France, 28-30 August 2013 View PDF file L. Davidson "The PANS k-eps model in a zonal hybrid RANS-LES formulation", International Journal of Heat and Fluid Flow, pp. 112-126, vol. 46, 2014. View PDF file of manusctip Get PDF file from publisher L. Davidson and C. Friess, The PANS and PITM model: a new formulation of f_k, Proceedings of 12th International ERCOFTAC Symposium on Engineering Turbulence Modelling and Measurements (ETMM12), Montpelier, France 26-28 September, 2018 View PDF file L. Davidson, Zonal Detached Eddy Simulation coupled with steady RANS in the wall region, ECCOMAS MSF 2019 Thematic Conference, 18-20 September 2019, Sarajevo, Bosnia-Herzegovina. View PDF file L. Davidson, "Non-Zonal Detached Eddy Simulation coupled with a steady RANS solver in the wall region", ERCOFTAC Bullentin 120, Special Issue on Current trends in RANS-based scale-resolving simulation methods, pp. 43-48, 2019. View PDF file L. Davidson and Ch. Friess, "Detached Eddy Simulations: Analysis of a limit on the dissipation term for reducing spectral energy transfer at cut-off", ETMM13: The 13th International ERCOFTAC symposium on engineering, turbulence, modelling Rhodes, Greece, 15-17 September, 2021 View PDF file L. Davidson "Detached Eddy Simulation coupled with steady RANS in the wall region", ETMM13: The 13th International ERCOFTAC symposium on engineering, turbulence, modelling Rhodes, Greece, 15-17 September, 2021 View PDF file L. Davidson "Using Machine Learning for formulating new wall functions for Large Eddy Simulation: A First Attempt", Div. of Fluid Dynamics, Mechanics and Maritime Sciences, Chalmers University of Technology, 2022. View PDF file L. Davidson "Using Machine Learning for formulating new wall functions for Large Eddy Simulation: A Second Attempt", Div. of Fluid Dynamics, Mechanics and Maritime Sciences, Chalmers University of Technology, 2022. View PDF file L. Davidson "Using Machine Learning for formulating new wall functions for Detached Eddy Simulation", ERCOFTAC symposium on Engineering, Turbulence, Modelling and Measurements (ETMM14), in Mini-Symposium: Machine learning for turbulence, Barcelona, Spain 6th - 8th September 2023; Chalmers University of Technology, 2022. View PDF file L. Davidson "Using Machine Learning for Improving a Non-Linear k-eps Model: A First Attempt", Div. of Fluid Dynamics, Mechanics and Maritime Sciences, Chalmers University of Technology, 2023. View PDF file L. Davidson "Using Neural Network for Improving an Explicit Algebraic Stress Model in 2D Flow", CUSF 2024, Proceedings of the Cambridge Unsteady Flow Symposium", Springer, Editors: J. C. Tyacke and N. R. Vadlamani, 2024 (to appear) View presentation View PDF file Proceedings Download code QUESTiONS & FURTHER iNFORMATiON Please contact Lars Davidson tel. 46 (0) 730-791 161, E-mail: lada@flowsim.se, lada@chalmers.se www.cfd-sweden
Page by I Ljungström ilj//flowsim.se www.flowsim.se