jwallwork23 / nn_adapt

Accelerating goal-oriented error estimation using neural networks
1 stars 1 forks source link

Add a simple Burgers example #19

Closed jwallwork23 closed 2 years ago

jwallwork23 commented 2 years ago

Setup based on this notebook. At present, we only consider one timestep and the initial condition is always a scaling of the same sinusoid. The QoI is always the same integral over the right hand boundary.

For simplicity, the same uniform mesh is used initially in each case. There were some minor changes required to the code to support the case with a non-mixed vector function space. Coincidentally, the number of DoFs per element is 12, as with the existing steady-state shallow water example.

To test out this new example, change the MODEL variable to burgers in the makefile.