| Author | Message | ||
     Augusto (cdfx) Member Username: cdfx Post Number: 9 Registered: 04-2007 |
Hi Mr. Nelson I receiving negative values for this experiment (in theory this is not correctly). The experiment consist in plates parallel bars with a viscou fluid between. Exist an analytical solution for this problem, but I can't obtain. Other question is how I make the scalar value inside of the solid bars (in x y x) equal to zero? Thanks again Mr. Nelson TITLE 'Bars 3D' COORDINATES cartesian3 VARIABLES vx(1e-6) vy(1e-6) vz(1e-6) p SELECT { method controls } ngrid = 10.0 errlim = 0.001 stages = 1 regrid = off DEFINITIONS { parameter definitions } LIx = 0.0 LIy = 0.0 LSx = 3.0 LSy = 3.0 RPT = 2 dens = 1.0 visc = 0.001 penalty = 1.0 Pin = 0.001 INITIAL VALUES vx = 0 vy = 0 vz = 0 p = Pin EQUATIONS vx: div(grad(vx)) - dx(p)/visc = dens*(vx*dx(vx) + vy*dy(vx) + vz*dz(vx))/visc vy: div(grad(vy)) - dy(p)/visc = dens*(vx*dx(vy) + vy*dy(vy) + vz*dz(vy))/visc vz: div(grad(vz)) - dz(p)/visc = dens*(vx*dx(vz) + vy*dy(vz) + vz*dz(vz))/visc p: div(grad(p)) = penalty*(dx(vx)+dy(vy)+dz(vz)) ! CONSTRAINTS { Integral constraints } EXTRUSION SURFACE 'A-INI' z = 0.0 LAYER 'LA' SURFACE 'A-FIN' z = 1.0 LAYER 'LB' SURFACE 'B-FIN' z = 2.0 LAYER 'LC' SURFACE 'C-FIN' z = 3.0 BOUNDARIES REGION 1 START(LIx, LIy) value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 LINE TO (LSx, LIy) natural(vx)=0 natural(vy)=0 natural(vz)=0 value(p)=-Pin LINE TO (LSx, LSy) value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 LINE TO (LIx, LSy) natural(vx)=0 natural(vy)=0 natural(vz)=0 value(p)=Pin LINE TO CLOSE ! FIELD 1 LIMITED REGION 2 LAYER 'LA' VOID LAYER 'LC' VOID REPEAT i = 0 TO RPT REPEAT j = 0 TO RPT START (0.0 + i, 0.3333333333333333 + j) value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 LINE TO (1.0 + i, 0.3333333333333333 + j) TO (1.0 + i, 0.3333333333333333 + 0.3333333333333333 + j) TO (0.0 + i, 0.3333333333333333 + 0.3333333333333333 + j) TO CLOSE ENDREPEAT ENDREPEAT ! FIELD 2 LIMITED REGION 3 LAYER 'LB' VOID REPEAT i = 0 TO RPT REPEAT j = 0 TO RPT START (0.0 + i, 0.3333333333333333 + j) value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 LINE TO (1.0 + i, 0.3333333333333333 + j) TO (1.0 + i, 0.3333333333333333 + 0.3333333333333333 + j) TO (0.0 + i, 0.3333333333333333 + 0.3333333333333333 + j) TO CLOSE ENDREPEAT ENDREPEAT ! FIELD 3 LIMITED REGION 2 LAYER 'LC' VOID REPEAT i = 0 TO RPT REPEAT j = 0 TO RPT START (0.0 + i, 0.3333333333333333 + j) value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 LINE TO (1.0 + i, 0.3333333333333333 + j) TO (1.0 + i, 0.3333333333333333 + 0.3333333333333333 + j) TO (0.0 + i, 0.3333333333333333 + 0.3333333333333333 + j) TO CLOSE ENDREPEAT ENDREPEAT monitors GRID(x,y,z) vector(vx,vy) on z = 1 vector(vx,vy) zoom(1.0, 1.0, 1.0, 1.0) on z = 1 vector(vy,vz) on x = 1 vector(vz,vx) on y = 1 contour(vx) on z = 1 contour(vy) on x = 1 contour(vz) on y = 1 plots GRID(x,y,z) VECTOR(vx,vy) zoom(1.0, 1.0, 1.0, 1.0) ON z = 1 points = 70 export format "#1,#2" file "field_2d.txt" ! VECTOR(k*grad(vy)) zoom(1.0, 1.0, 1.0, 1.0) ON x = 0 points = 70 export format "(#x,#y) = (#1,#2)" file "FIELD_YZ.TXT" as "Vector Field YZ" ! VECTOR(k*grad(vz)) zoom(1.0, 1.0, 1.0, 1.0) ON y = 0 points = 70 export format "(#x,#y) = (#1,#2)" file "FIELD_XZ.TXT" as "Vector Field XZ" ! TABLE(dx(vx), dy(vy), dz(vz)) zoom(2.0, 2.0, 2.0, 1.0, 1.0, 1.0) points = 70 format "(#x, #y, #z) = (#1, #2, #3)" file "FIELD_XYZ.TXT" TABLE(dx(vx), dy(vy), dz(vz)) zoom(1.0, 1.0, 1.0, 1.0, 1.0, 1.0) points = 70 format "#1 #2 #3" file "field_3d.txt" vector(vx,vy) on z = 1 vector(vy,vz) on x = 1 vector(vz,vx) on y = 1 contour(vx) on z = 1 contour(vy) on x = 1 contour(vz) on y = 1 end | ||
| Augusto (cdfx) Member Username: cdfx Post Number: 10 Registered: 04-2007 |
Ops the paragraph "Exist an analytical solution for this problem, but I can't obtain." Is wrong, sorry! I can't generate the numerical values to compare with analytical solution! Thanks again! | ||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 878 Registered: 06-2003 |
1. The only negative numbers I see are in Vy, which is nominally zero. These are simply numerical noise, four orders of magnitude below the dominant flow velocities. This is a numerical approximation, after all, not an analytic solution, and on a very crude mesh at that. 2. Your four slabs are all disjoint. There is no point in running them all together; it just creates a larger mesh and runs longer. You can do each one individually. 3. If "the space between the bars" you refer to is the gaps between slabs, then they are already zero, because you excluded them. If you mean to stop the flow in the mid section of each slab, just exclude that as well. 4. If you are using the student version, it won't work. You need more nodes than the student version allows. Try running the individual slabs.   |