how to set the periodic BC in channel?
 

the channel flow is computed with 3d compressible NS code using MacMarco scheme. in stream wise the periodic bc is used, denoted f_x. in aiaa paper 98-3315, it was -2.*u_taw**2. another paper suggest that it is evaluated using f(d_u/dy). which is better? my code is not valid for this problem now. in initial conditions, all flow parameters is uniform. u=1,v=w=0, density r=1, temperature T=1.(non-dimension) and i add a fluctuate to u within [-0.05,0.05] at inlet boundary in certain time-step. the flow can not develop. i think maybe the initial u is wrong. anyone can send me a velocity profile datas with any Mach number any Renoly number, only compressible. thanks. maybe you would give some advice about it. best regards