laminar flow

Log Out | Topics | Search Moderators | Register | Edit Profile

Author Message   Emin kebeli (dextermania) Member Username: dextermania Post Number: 9 Registered: 05-2009 Posted on Sunday, June 28, 2009 - 06:16 pm:    is there a chance that we can move the input and output hole to the side walls ? title '3D flow through a plenum' coordinates cartesian3 variables vx(1e-6) vy(1e-6) vz(1e-6) p select ngrid=20 stages=2 regrid=off definitions long = 2 wide = 1 high = 1/2 xin = 0 zin = 0 xout = 0 yout = 0 rc = 0.2 duct = 0.2 dens=staged(0,1) { fluid density } visc= 0.01 { fluid viscosity } v=vector(vx,vy,vz) vm=magnitude(v) div_v = dx(vx) + dy(vy) + dz(vz) PENALTY = 1e4*visc/high^2 Pin = 0.05 Pout = 0 initial values vx=0 vy=0 vz=0 p=Pin+(Pout-Pin)*(z+high+duct)/(2*high+2*duct) equations vx: dens*(vx*dx(vx) + vy*dy(vx) + vz*dz(vx)) + dx(p) -visc*div(grad(vx)) = 0 vy: dens*(vx*dx(vy) + vy*dy(vy) + vz*dz(vy)) + dy(p) -visc*div(grad(vy)) = 0 vz: dens*(vx*dx(vz) + vy*dy(vz) + vz*dz(vz)) + dz(p) -visc*div(grad(vz)) = 0 p: div(grad(p)) = PENALTY*div_v extrusion z = -high-duct,-high,high,high+duct boundaries Region 1 { plenum box } surface 4 value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 surface 3 value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 layer 3 void layer 1 void start(-long,-wide) value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 { fix all side values } line to (long,-wide) to (long,wide) to (-long,wide) to close limited Region 2 { input hole } layer 3 surface 3 natural(vx)=0 natural(vy)=0 natural(vz)=0 value(p)=Pin { input duct opening } start(xin,zin-rc) layer 3 value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 { duct sidewall drag } arc(center=xin,zin) angle=360 limited Region 3 { exit hole } layer 1 surface 1 natural(vx)=0 natural(vy)=0 natural(vz)=0 value(p)=Pout { output duct opening } start(xout,yout-rc) layer 1 value(vx)=0 value(vy)=0 value(vz)=0 natural(p)=0 { duct sidewall drag } arc(center=xout,yout) angle=360 monitors contour(vx) on x=0 report dens report pin contour(vx) on y=0 report dens report pin contour(vz) on y=0 report dens report pin vector(vx,vz)on y=0 report dens report pin contour(vx) on z=0 report dens report pin contour(vy) on z=0 report dens report pin contour(vz) on z=0 report dens report pin vector(vx,vy)on z=0 report dens report pin contour(p) on y=0 report dens report pin plots contour(vx) on x=0 report dens report pin contour(vx) on y=0 report dens report pin contour(vz) on y=0 report dens report pin vector(vx,vz)on y=0 report dens report pin contour(vx) on z=0 report dens report pin contour(vy) on z=0 report dens report pin contour(vz) on z=0 report dens report pin vector(vx,vy)on z=0 report dens report pin contour(p) on y=0 report dens report pin end   Marek Nelson (mgnelson) Moderator Username: mgnelson Post Number: 133 Registered: 07-2007 Posted on Tuesday, June 30, 2009 - 10:39 pm:    You could move the input/output holes to the sidewalls, but you would then have to define the surfaces of the cylinders. Whereas when the holes are on the top/bottom, you can get the cylinders by circles in the projection plane. i.e. it is simpler to model the cylinders in the extrusion direction. Take the model you have in mind and turn it on it's end. Then it is the same as the example. It would be much easier to adjust the dimensions of the existing problem to match the shape and size of your model turned on its side.