Add OP nonconforming interface in +scheme/Elastic2dCurvilinearAnisotropicIncompatible.m
Change default tuning for displacement BC and interfaces to 1.2 in Elastic2dCurvilinearAnisotropic.m
Add OP interpolation in Elastic2dVariableAnisotropicIncompatible
Merge heads
Support using both boundary optimized and standard operator sets for gluing
Bugfix u->v in impedance-weighted interface treatment. Add 1e-14 to avoid division by zero in dummy discrtization with zero material parameters.
Add impedance weighting in normal-tangential interface treatment in 2DElasticAnisotropicCurve. Add the same interface treatment in the upwind (dual-pair) scheme.
Set default penalty strength for Curvilinear Incompatible to 1.2 istead of 1.
Add periodicUpwind class, copy from D2_boundary_opt
mergin branch?
Bugfixes: 1. Borrow C_boundary from C_interior instead of the other way around. 2. Fix off-by-1 error in shifting. Also set default tuning parameter to 1.0 istead of 1.2.
Add scheme for incompatible curvilinear
Bugfix with borrowing matrix accidentally being sparse Cartesian anisotropic incompatible.
Bugfix for non-fully compatible when Nx \neq Ny.
Add interface, not fully compatible.
Add scheme for Elastic2dVariable, not fully compatible.
Refactor computation of borrowing from C. Update to correct number of boundary points for borrowing from R.
Add boundary optimized upwind operators (copy from default). In Elastic2dCurvilinearAnisotropicUpwind, adapt metric derivative computation to use the same Dp operator.
Bugfix zeros in norm for 6th order optimal upwind.
In Elastic2dCurvilinearAnisotropicUpwind, if periodic upwind ops are used, use central D1 of (at least) same order to compute metric derivatives.
Add periodic updwind ops (copy from default)
Add class for periodic ops.
Added periodic Upwind ops
clean up
Added Optimal Upwind ops
Add Upwind Optimal, maybe
Added optimal Upwind
Add support for higher orders than 6 in Elastic2dCurvilinearAnisotropicUpwind. Include possibility for periodic ops.
Added optimal Upwind
Euler RV feature
Change remote to sourceforge
Use non-stiff SATs for dirichlet and interface conditions
Update boundary borrowing constants
Merge heads for feature/D2_boundary_opt
Add comments on passing grid lines to Rectangle.show
Add glue interpolation operators
Add normal and tangental vectors to scheme
Add support for Erickson2022 non-stiff BC in LaplaceCurvilinearNew.
Add method getBoundaryField in multiblock.DiffOp, which returns a square matrix of a field that lives on the boundary.
Add function multiblock.joinBoundaryGroups
In EmbeddedRungeKutta: add check for NaNs/Infs in rates K. Separate errors dt=NaN and dt < 1e-30.
Remove Erickson2022 SAT closures for frictional fault from Laplace1DVariable.interface. The closures are the same as for BC and can easily be handled in a Discr class.
Add working (but ugly) implementation of Erickson2022.
Add support for boundary optimized D2
Use properties instead of calling tableu function
Add 3rd and 5th order time steppers based on the adaptive schemes
Add embedded RK timestepper
Commit pending changes. Quite messy. Not clear what to keep.
Merge latest changes from default
Merge with upstream
Allow for non-linear continuous and discrete source
Merge feature/laplace1d_variable
Closing branch before merge to default
Add support for different opsets and clean up bc/interface methods
Closing this head
Closing Martins bad branch to the tune of 20 whip lashes.
Played around with Ti show and plot settings to create pretty pics for thesis.
Merge with feature/frids
Add quadrature stencils and operator assembly. Add tests
Add D2 files
Fixed bug in tests
add testfor D2 and sign in assemble_op
rename d2_2 to D1_standard_2
fix bug in assemble_D1
rename boundart_stencil and and boundary_block to boundary_stencils
Add files for parsing operator data stencil file
Add assemble_boundary_op
Stupid merge makes this branch dangerous. Merging with other branches will result in unexpected changes. Never merge this branch with anything. Closing branch.
Merge with feature/poroelastic. Use only the changes made to multiblock.DiffOp and scheme.Elastic2dVariable. DiffOp.getBoundaryOperator/Quadrature now use scheme methods instead of propeties.
Remove round off zeros from D2Nonequidistant operators
Pass logic grid along to diracDiscr
Allow passing number of gridlines for each block
Fix comment
Add periodic D1 and D2 operators for orders 8,10,12
Add methods for creating grids with different grid point distributions for each coordinate direction, and also supports constructing periodic grids
Calculate width for all orders
Use the logic grid associated with a CurvilinearGrid instead of creating a new one
Add support for RK6
Clarify comments
Merge with feature/poroelastic for Elastic schemes
Add Zt to output args for Elastic2dCurvilinearAnisotropic.interfaceNormalTangential
Remove warning about normal-tangential BC, because the BC decouple in normal-tangential coordinates
Add some interface forcing penalties in ViscoElastic2d
Bugfix interface forcing in ElasticCurveAnisotropic
Implement normal-tangential interface treatment in viscoElastic2d
Add some interface forcing penalties in ElasticCurvilinearAnisotropic
Add interfaceForcing method in multiblock.DiffOp to facilitate mms forcing
Add new interface method interfaceNormalTangential in Elastic2dCurvilinearAnisotropic.
Add traction operators that include the viscous terms and use them to clean up the imposition of traction BC.
Add viscoelastic fault interface
Add viscoelastic regular interfaces
Move common code for displacement and traction BC outside of switch statement
Add viscoelastic displacement BC for normal-tangential. Bugfix traction BC for normal-tangential
Add displacement BC to viscoelastic
Add viscoElastic traction bc for normal-tangential
Add diffOp for viscoElastic
Clean up hollow D2 generation by using precomputed coefficients stored in mat files
Improve performance for the hollow case
Make the hollow scheme generation more efficient by introducing the D2VariableHollow opSet
Add the hollow option to the anisotropic diffops