flow rate decrement

 Register Blogs Members List Search Today's Posts Mark Forums Read

 August 5, 2008, 10:12 flow rate decrement #1 d.h. Guest   Posts: n/a Hi! I have one question for 3d channel flow simulation. That is the flow rate decrement. I checked the flow rate for each time step. As followed, that decrease continuously. if nstep = 1, flow rate = 0.839421 if nstep = 5000, flow rate = 0.776719 if nstep = 10000, flow rate = 0.743256 For periodic b.c. I use constant pressure gredient((density*friction velocity^2)/(the half width of channel)) in u-momentum NSE. I don't know that I look wrong part.........

 August 5, 2008, 12:05 Re: flow rate decrement #2 pankaj saha Guest   Posts: n/a hi, i am also working on the same , facing some problem. Would you like to discuss with me. Please, tell me the following---- Domain size, grid, Re, time step, and initial condition. i hope you are running a incompressible code. what is the exact value of presure gradient you are using. i mean the numerical value. If you are using non-dimensional code--and then the value of pressure gradient shoud be 1.0 and it will be added to the x-momentum equation . Also, in initial condition , are you using random perderbation?

 August 5, 2008, 20:05 Re: flow rate decrement #3 d.h. Guest   Posts: n/a hi, pankaj saha my code's details are domain size : 2(streamwise)*1(normal)*2(spanwise) staggered grid, 16(streamwise)*64(normal)*32(spanwise) Retau=395(for friction velocity and the channel half width) i.c. : u=-1.5*y(y-2) (i don't know generating perturbation....) i don't use non-dimentionalized NSE. constant pressure gradient = friction vel.^2/half width since NSE is devided by density, density is not. friction velocity used is obtained by averaging all value near the wall(including north and south wall) i think flow rate decrement may be induced by no slip b.c. but i don't find my faults near the wall.

 August 6, 2008, 05:18 Re: flow rate decrement #4 pankaj saha Guest   Posts: n/a hi, you are telling that, you are calulating friction velocity by averaging at near wall. that means at every timestep it will vary and your pressure gradient and Retau will vary. i hope this is wrong. you are using dimensional code.then you are supplying seting viscosity not the retau in your code. i think your viscosity=utau*0.5/395. here you need Utau. normally, people take Utau=1 and set viscosity. from that your pressure gradient will have a constant value=2.0 in your case. just check the formulation of getting mean pressure gradient. also, your x-dimension is very less compare to eddy size. take 2pi=6.3 minimum, othwerwise you will not get anything..also, grid should be large 50x66x60 try...