Hello everyone, I encountered a weird situation when calculating the error order of Time-dependent Ginzburg-Landau(TDGL) equation.
In order to calculate this equation, I use BDF1 discretization in time, P1 element in space for Psi, and BDM1Ortho element for A which means A belongs to hcurl space.
Here is my weak form, and due to the influence of the boundary term, there is no boundary integral in the weak form
In calculation, because of the discretization method I selected, theoretically, the time error order should be 1 and the space error order should be 2. And I set the relationship between dt and h as dt=h^2.
And the weird part came. When I calculated T=1 or T=1/8, the result was wrong, but when T=10, the error order was right. I don’t know what caused it. What I hope is that the correct result will appear when T is small. Here is my code :
GL(bdf1).edp (4.1 KB)
Because of the computer, I didn’t set h to be small, but I don’t think it’s the key of the problem
If you have any suggestions, I would be very grateful! 