aliiab
Forum Replies Created
-
AuthorPosts
-
aliiabMember
Ok, thank you. I hoped that the dependence is not linear:). Thank you, Robin.
aliiabMemberRobin, thanks. The last question. Can you tell me the approximation dependence OpenLB simulation time on the L_lattice value. In fact, I just want to keep values of the tau, u_lattice, u_char constant. And I want to reduce char_nu by 3 orders. So I have to reduce L_lattice by 3 orders too. So I want to estimate the simulation time increase after theese changes.
aliiabMemberMathias, thanks.rnOne more question, please:) I have detected, that for low kinematic viscosity values (~10^(-7) – 10^(-9)) one needs to make very-very small grids than for biger viscosity (~10^(-3) – 10^(-5)). I want to know is there a exact relation between viscosity value and latticeL value. My suggestion is that If one reduces viscosity value by 3 orders one has to reduce latticeL value by 3 orders too…
aliiabMemberOne more question. Are there any criterias for best space and time resolutions?rnrnI know that one can use CharU and LatticeU to compute the time step size in SI units (seconds) as:rn
Quote:physTime = latticeU/charU * latticeL(http://optilb.org/openlb/forum?mingleforumaction=viewtopic&t=78)rnrnI suppose that it’s possible to choose appropriate latticeU and latticeL values to get good time step. But what value time step has to be equal to in concrete physical conditions (Re, Rho, charU)?rn
aliiabMemberMathias, thanks. But it’s still not clear: SuperLatticePhysDrag3D (like it’s used in cylinder3d example) presents drag force or drag coefficient?
aliiabMemberThanks, Mathias. Yes, I know than my Re is not used in the code. But I expected that calculated Re = CharRo*CharU*CharL/CharNu. And now I see that OpenLB calculates Re correctly using this formula. Sorry, it was my mistake in calculations (probably I used another value of sphereDiam (and charL) when calculate Re value manually).rnrnThank you for you advice about grid and turbulent model. I will experiment.rnrnThanks,rnAlina.
aliiabMemberRobin, thank you for detailed explanation.
aliiabMemberHi, Liliana, Pedro!rnrnI am intrested in drag force too. I simulate the fluid flow in a rectangular channel with sphere in its center. So if
Quote:drag given by the code is already the drag coefficient, it should be equal to the 0.47 for Re = 10 000 (https://en.wikipedia.org/wiki/Drag_coefficient). But in my simulation it’s about 5.0, but it is not a constant value (fluid velocity and density are constant in this simulation). Can you explain the reason of such difference?rnAlso can you tell me, how OpenLB calculates the Reynolds number?rnrnThanks,rnAlina.
aliiabMemberThanks, Robin! Now it works correctly.
-
AuthorPosts