Neural network modeling for near wall turbulent flow

A neural network methodology is developed in order to reconstruct the near wall field in a turbulent flow by exploiting flow fields provided by direct numerical simulations. The results obtained from the neural network methodology are compared with the results obtained from prediction and reconstruction using proper orthogonal decomposition (POD). Using the property that the POD is equivalent to a specific linear neural network, a nonlinear neural network extension is presented. It is shown that for a relatively small additional computational cost nonlinear neural networks provide us with improved reconstruction and prediction capabilities for the near wall velocity fields. Based on these results advantages and drawbacks of both approaches are discussed with an outlook toward the development of near wall models for turbulence modeling and control.

[1]  Stephen J. Kline,et al.  The production of turbulence near a smooth wall in a turbulent boundary layer , 1971, Journal of Fluid Mechanics.

[2]  Nicol N. Schraudolph,et al.  Local Gain Adaptation in Stochastic Gradient Descent , 1999 .

[3]  Lawrence Sirovich,et al.  LOW DIMENSIONAL DESCRIPTION OF COMPLICATED PHENOMENA , 1988 .

[4]  Ramesh K. Agarwal,et al.  NEURAL NETWORK APPROACHES TO SOME MODEL FLOW CONTROL PROBLEMS , 1997 .

[5]  P. Holmes,et al.  The Proper Orthogonal Decomposition in the Analysis of Turbulent Flows , 1993 .

[6]  Nadine Aubry,et al.  The dynamics of coherent structures in the wall region of a turbulent boundary layer , 1988, Journal of Fluid Mechanics.

[7]  John L. Lumley,et al.  RECONSTRUCTING THE FLOW IN THE WALL REGION FROM WALL SENSORS , 1998 .

[8]  William E. Faller,et al.  Real-time prediction and control of three-dimensional unsteady separated flow fields using neural networks , 1994 .

[9]  Parviz Moin,et al.  Active turbulence control for drag reduction in wall-bounded flows , 1994, Journal of Fluid Mechanics.

[10]  Lawrence Sirovich,et al.  The Karhunen–Loéve decomposition of minimal channel flow , 1997 .

[11]  F. Takens,et al.  On the nature of turbulence , 1971 .

[12]  S. K. Robinson,et al.  Coherent Motions in the Turbulent Boundary Layer , 1991 .

[13]  R. Goodman,et al.  Application of neural networks to turbulence control for drag reduction , 1997 .

[14]  L. Sirovich.  Turbulence and the dynamics of coherent structures. III. Dynamics and scaling , 1987 .

[15]  L. Sirovich.  Turbulence and the dynamics of coherent structures. I. Coherent structures , 1987 .

[16]  Vladimir Cherkassky,et al.  Learning from data , 1998 .

[17]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[18]  P. Moin,et al.  Turbulence statistics in fully developed channel flow at low Reynolds number , 1987, Journal of Fluid Mechanics.

[19]  D. Chambers,et al.  Karhunen-Loeve expansion of Burgers' model of turbulence , 1988 .

[20]  John L. Lumley,et al.  A low-dimensional approach for the minimal flow unit of a turbulent channel flow , 1996 .

[21]  William E. Faller,et al.  Neural network prediction of three-dimensional unsteady separated flowfields , 1993 .

[22]  S. Shvartsman,et al.  Nonlinear model reduction for control of distributed systems: A computer-assisted study , 1998 .

[23]  P. Moin,et al.  The minimal flow unit in near-wall turbulence , 1991, Journal of Fluid Mechanics.

[24]  P. R. Bevington,et al.  Data Reduction and Error Analysis for the Physical Sciences , 1969 .

[25]  Kurt Hornik,et al.  Neural networks and principal component analysis: Learning from examples without local minima , 1989, Neural Networks.