[hiter]

I have one simple qestion that is how to use Taylor’s frozen turbulence hypothesis in DNS similation. In many references, they all simply say 'Imposed'.

For instance, we have already obtained the HIT external turbulence.
u_et(t,y,z). t is the time step, and y, z denote vertical and tranvese directions.

At the inlet X=0, the instantoust velocity u(0, y, z) = u_et(0, y, z) + U0(y, z). U0 is the given mean inlet velocity.
At this step, everything is O.K.

In our DNS, in x direction the nonlinear terms are discretized by the fourth-order central difference scheme, and visoucous term by eighth-order central compact scheme.
We assume that we set the caculation domain point in X direction N1, NX.
So the boudary condion reqired velocity at point N-3, N-2, N-1. Howver, we only have the u_et(t,y,z) at one time shot.

Also, another problem, maybe less that puzzing. We use the fractal step method, at every substep, is it reason to remain the u_et unchanged.

Thanks in advance.