Problem with implementation of SUPG in FEM for flow problems
 

Dear Guys,
I'm trying to implement SUPG stabilization to FEM code for flow problems. I tried it in 1D and 2D for scalar variable and it is OK. Then I tried to use it for flow analysis where problems started. I tried many strategies proposed but I couldn't make it. I think I'm madding a logical mistake in implementing SUPG. I'm attaching the code to get your opinions. The code works well for low Re but when it comes to higher Re, it fails. SUPG was implemented in int_KF2.m function. The code is based on mixed formulation and 9-point quadrilateral Taylor-Hood element was used. Just rum main.m and see the results written in fem.dat file in TECPLOT. The SUPG parameter was calculated as proposed by Tezduyar. I would appreciated any comment on the code...
Tnx
PS: It is better to say that I implemented SU.