Artificial neural network-based one-equation model for simulation of laminar-turbulent transitional flow

A mapping function between the Reynolds-averaged Navier-Stokes mean flow variables and transition intermittency factor is constructed by fully connected artificial neural network(ANN),which replaces the governing equation of the intermittency factor in transition-predictive Spalart-Allmaras(SA)-γmodel.By taking SA-γmodel as the benchmark,the present ANN model is trained at two airfoils with various angles of attack,Mach numbers and Reynolds numbers,and tested with unseen airfoils in different flow states.The a posteriori tests manifest that the mean pressure coefficient,skin friction coefficient,size of laminar separation bubble,mean streamwise velocity,Reynolds shear stress and lift/drag/moment coefficient from the present two-way coupling ANN model almost coincide with those from the benchmark SA-γmodel.Furthermore,the ANN model proves to exhibit a higher calculation efficiency and better convergence quality than traditional SA-γmodel.

Basic Study on Perfect Push-Pull Local Ventilation

In this study, the authors experimentally investigated the changes of the mean velocity component profiles, half-widths （b12）, turbulence intensities, Reynolds shear stress and intermittency of turbulence of a transient plane turbulent jet developing from a jet exit into a hood opening. The values of maximum mean-velocity and half-widths of the axial velocity profile along the center-line of the jet are greater than those for a fully developed two-dimensional jet. Turbulence intensity in the axial direction is not affected by the flow rate ratio. At the same time, turbulence intensity in the lateral direction becomes greater as the hood is approached and the flow rate ratio Q3/Q1 becomes larger （QI is jet flow rate from nozzle and Q3 is suction flow rate produced by the hood）. These experimental results are in accord with the distributions of production terms in the axial and lateral directions. Reynolds shear stress becomes smaller as the flow rate ratio becomes larger near the hood. Dimensionless distance y1/br2, from the center axis of the flow to the point where intermittency factor y becomes a constant value, narrows as the flow rate ratio becomes larger near the hood.

Research of the parameters of SL scaling law

As a universal principle of turbulence cascade,scaling law is important in the research on statistics of turbulence.In the thesis PIV system is used to measure the streamwise velocity vector(u') and the component of vorti-city (dv'/dx) in order to analyze the parameters of SL scaling law in the boundary layer at Reθ=2694 (Reθ is obtained by the momentum thickness and free flow velocity).The method of wavelet and statistics are used in this paper.Then,the parameters of SL scaling law are fitted with the scaling law of extended self similarity,extended of refined similarity hypothesis and a new form of refined similarity.At different distances from the wall and different scale,the results show that the fitted curves of u' is in accordance with experimental curves.For dv'/dx,the fitted curves which are fitted with extended of refined similarity hypothesis and a new form of refined similarity is in accordance with experimental curves.

Validation of a RANS transition model using a high-order weighted compact nonlinear scheme

A modified transition model is given based on the shear stress transport (SST) turbulence model and an intermittency transport equation. The energy gradient term in the original model is replaced by flow strain rate to saving computational costs. The model employs local variables only, and then it can be conveniently implemented in modern computational fluid dynamics codes. The fifth-order weighted compact nonlinear scheme and the fourth-order staggered scheme are applied to discrete the governing equations for the purpose of minimizing discretization errors, so as to mitigate the confusion between numerical errors and transition model errors. The high-order package is compared with a second-order TVD method on simulating the transitional flow of a flat plate. Numerical results indicate that the high-order package give better grid convergence property than that of the second-order method. Validation of the transition model is performed for transitional flows ranging from low speed to hypersonic speed.