codice matlab mesh staggered harlow welsh

[marvelreveal]

Hello, everyone. I am about to write a matlab code for the resolution of incompressible 2d ship stokes, with a mesh staggered to the harlow welsh in a square domain. in my request we require a bc to the dirichlet on what I believe to be the unord wall, function of x, therefore a real function. insertion of this bc generates a row array of nx points.
the problem is that in the mesh staggered u matrix has size (nx x ny-1).
I therefore do not know how to solve, because the size of my bc is incompatible with the matrix in which it should be inserted. Does anyone have any idea?this is what generates me is a problem from the dimensional point of view, and in the calculation of diffuse terms.

 

[Stan9411]

Well the demand is extremely technical to be placed in a forum where we speak mainly of mechanical design.
in this regard a small note: do not assume that when you turn to someone he shares your own vocabulary... "bc." even the acronym of an Englishman... "condition to the contour" cost so much? :

I don't understand anything about fluid dynamics. but when I was at university I made several dynamic models to partial drifts and I was still passionate about it.
And in my opinion, your problem is not fluidodynamic but mathematical.

In particular, I think you're misinterpreting the concept of dirichlet boundary condition.
in practice when you solve a differential equation, you do not get numbers as a solution, but an equation, a behavior of something within a domain.
the dirichlet condition imposes what is the value of your function on the edge of the domain.

if your differential equation is defined on a 1d domain:
= 0
the condition of dirichlet would be:
= b
where x0 and xf are the extremes of your domain 1d (eg: time)

that your function is defined on a surface, it is flashing that the contour of your domain is no longer a pair of points, but it is a closed line (eg: the domain is a circle, the contour is its outer circumference).

so it is right that your condition generates a 1xn array. if the domain is 2d, the boundary condition is 1d. point.
What doesn't come back is that you only have it on one side of your square. In theory, you should have it for all four sides. but maybe this is a peculiarity of the fluid dynamic problem and here I can't help you.