![]() |
OptFEM2DP1 Toolbox
V1.2
Matlab/Octave Optimized P1-Lagrange Finite Element Method in 2D
|
00001 function R=StiffAssembling2DP1OptV2(nq,nme,q,me,areas) 00002 % function R=StiffAssembling2DP1OptV2(nq,nme,q,me,areas) 00003 % Assembly of the Stiffness Matrix using `P_1`-Lagrange finite elements 00004 % - OptV2 version (see report). 00005 % 00006 % The Stiffness Matrix `\Stiff` is given by 00007 % ``\Stiff_{i,j}=\int_\DOMH \DOT{\GRAD\FoncBase_i(\q)}{\GRAD\FoncBase_j(\q)}d\q,\ \forall (i,j)\in{\ENS{1}{\nq}}^2`` 00008 % where `\FoncBase_i` are `P_1`-Lagrange basis functions. 00009 % 00010 % Parameters: 00011 % nq: total number of nodes of the mesh, also denoted by `\nq`. 00012 % nme: total number of triangles, also denoted by `\nme`. 00013 % q: Array of vertices coordinates, `2\times\nq` array. <br/> 00014 % `{\q}(\il,j)` is the 00015 % `\il`-th coordinate of the `j`-th vertex, `\il\in\{1,2\}` and 00016 % `j\in\ENS{1}{\nq}` 00017 % me: Connectivity array, `3\times\nme` array.<br/> 00018 % `\me(\jl,k)` is the storage index of the 00019 % `\jl`-th vertex of the `k`-th triangle in the array `\q` of vertices coordinates, `\jl\in\{1,2,3\}` and 00020 % `k\in{\ENS{1}{\nme}}`. 00021 % areas: Array of areas, `1\times\nme` array. areas(k) is the area of the `k`-th triangle. 00022 % 00023 % Return values: 00024 % R: Stiffness Matrix, `\nq\times\nq` sparse matrix. 00025 % 00026 % Example: 00027 % @verbatim 00028 % Th=SquareMesh(10); 00029 % R=StiffAssembling2DP1OptV2(Th.nq,Th.nme,Th.me,Th.q,Th.areas); 00030 % @endverbatim 00031 % Copyright: 00032 % See \ref license 00033 Ig = me([1 2 3 1 2 3 1 2 3],:); 00034 Jg = me([1 1 1 2 2 2 3 3 3],:); 00035 00036 q1 =q(:,me(1,:)); q2 =q(:,me(2,:)); q3 =q(:,me(3,:)); 00037 u = q2-q3; v=q3-q1; w=q1-q2; 00038 clear q1 q2 q3 00039 areas4=4*areas; 00040 Kg=zeros(9,nme); 00041 Kg(1,:)=sum(u.*u,1)./areas4; % K1 ou G11 00042 Kg(2,:)=sum(v.*u,1)./areas4; % K2 ou G12 00043 Kg(3,:)=sum(w.*u,1)./areas4; % K3 ou G13 00044 Kg(5,:)=sum(v.*v,1)./areas4; % K5 ou G22 00045 Kg(6,:)=sum(w.*v,1)./areas4; % K6 ou G23 00046 Kg(9,:)=sum(w.*w,1)./areas4; % K9 ou G33 00047 Kg([4, 7, 8],:)=Kg([2, 3, 6],:); 00048 R = sparse(Ig,Jg,Kg,nq,nq);