フィルターのクリア

Solving elliptic pde for G and Q components with different b.c.s

1 回表示 (過去 30 日間)
S. Rae
S. Rae 2017 年 9 月 26 日

Please refer to the matlab documentation on elliptic pdes. https://www.mathworks.com/help/pde/ug/elliptic-pdes.html

So, I am ONLY trying to solve the components Q and G for multiple finite segments on the boundary.

$Q(i,j) = \int_{El} \phi_i \phi_j dS$ and $G(i) = \int_{El} \phi_i dS$,

where El, l = 1,...,N is some finite arc length of an object... say, .005m.... that occurs N times along the boundary.

How do I properly specify the boundary conditions matrices g and q? I attached an image of problem I am describing. The stars indicate the nodes I have to integrate over.

Here is what I know. We solve for G (and Q similarly by referring to the assemb function) as G = G + sparse(e(1,:),g,nodes,1); G = G + sparse(e(2,:),g,nodes,1), where it calculates the surface area between two nodes and {g = |P1-P2|/2 on El OR 0 off El, l=1,...,N} for all nodes on the boundary.

However, for nodes with indices n1,n2, do I need to find x1_n1 = find(e(1,:)==n1) and x2_n1 = find(e(2,:)==n1) => g(x1_n1) = 1 and g(x2_n1) = 1 AND x1_n2 = find(e(1,:)==n2) and x2_n2 = find(e(2,:)==n2) => g(x1_n2) = 1 and g(x2_n2) = 1?

I cannot figure out how to properly form the boundary conditions matrices for g and q because e(1,:) covers all the nodes on the boundary and so does e(2,:). Do I set g to its nnz value where e(1,x1*),e(2,x2*)=n1, etc.?

If anyone knows another simpler way to solve Q and G, please let me know as well! Please message me if you have further questions. I need help desperately because I have been stuck on this problem for a long time.

Thank you for the help!

回答 (0 件)

カテゴリ

Help Center および File ExchangeBoundary Conditions についてさらに検索

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by