dumux merge requestshttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests2019-05-29T13:22:23Zhttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1582[newton] Allow to change minimum number of Newton iterations2019-05-29T13:22:23ZTimo Kochtimokoch@math.uio.no[newton] Allow to change minimum number of Newton iterationsFixes #659.
* Introduce a parameter MinSteps for Newton
* Introduce a parameter Verbosity for Newton
* Replace bool verbose by int verbosity for verbosity levels
* Introduce parameter output (default: on) for NewtonFixes #659.
* Introduce a parameter MinSteps for Newton
* Introduce a parameter Verbosity for Newton
* Replace bool verbose by int verbosity for verbosity levels
* Introduce parameter output (default: on) for Newton3.1Martin SchneiderMartin Schneiderhttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1580Feature/pdesolver2019-05-07T11:06:32ZTimo Kochtimokoch@math.uio.noFeature/pdesolver* Implement abstract base class for PDESolvers
* Make NewtonSolver inherit from this new base class
* Implement LinearPDESolver based on this class
* Use the linear pde solver in one of the 1p tests* Implement abstract base class for PDESolvers
* Make NewtonSolver inherit from this new base class
* Implement LinearPDESolver based on this class
* Use the linear pde solver in one of the 1p tests3.1Dennis GläserDennis Gläserhttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1577[timeloop] Feature/checkpointtimeloop with memory2019-05-20T13:16:30ZTimo Kochtimokoch@math.uio.no[timeloop] Feature/checkpointtimeloop with memoryFixes #693.
Several reference solutions (dumux and dumux-lecture) have to be updates because this change might change some time step sizes after check points significantly which can lead to slightly different results.
* suggests th...Fixes #693.
Several reference solutions (dumux and dumux-lecture) have to be updates because this change might change some time step sizes after check points significantly which can lead to slightly different results.
* suggests the time step size before the last time step before a check point if the time step size had to be reduced to reach that check point
* increases the time step suggestion before a check point by up to 20% if the check point can be reached exactly
* makes floating point comparisons more robust by removing the dependence on the current simulation time
* [x] Update references in dumux
* [x] Update references in dumux-lecture 3.1Simon EmmertSimon Emmerthttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1576[io] Implement grid managers in seperate headers2019-05-07T10:56:04ZTimo Kochtimokoch@math.uio.no[io] Implement grid managers in seperate headers* With this patch you only include what you need
* The header gridmanager.hh includes all specializations
as convenience. This makes this patch also fully backwards-compatible
Fixes #583 * With this patch you only include what you need
* The header gridmanager.hh includes all specializations
as convenience. This makes this patch also fully backwards-compatible
Fixes #583 3.1Dennis GläserDennis Gläserhttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1573Feature/move gravity to spatialparams2019-05-13T14:48:33ZTimo Kochtimokoch@math.uio.noFeature/move gravity to spatialparamsFixes #696Fixes #6963.1Dennis GläserDennis Gläserhttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1555Improve flux variables cache filler handling2019-04-24T09:59:46ZTimo Kochtimokoch@math.uio.noImprove flux variables cache filler handling* Make the filler alias private in the caches (it's just an implementation policy)
* Make filler a property
* Extend interface of empty cache / filler
* If you want an empty cache (like for staggered-navierstokes) just set the cache a...* Make the filler alias private in the caches (it's just an implementation policy)
* Make filler a property
* Extend interface of empty cache / filler
* If you want an empty cache (like for staggered-navierstokes) just set the cache and filler to the empty placeholders
* Removes the cache implementation for freeflow (replaced by empty cache)
- [x] Get the Scalar type from somewhere else
Fixes #6873.1Kilian WeishauptKilian Weishaupthttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1553Feature/container variable precision2019-04-05T15:28:50ZTimo Kochtimokoch@math.uio.noFeature/container variable precision3.1Dennis GläserDennis Gläserhttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1550[facet][mpfa][fix] fix additional derivatives for numEq > 12019-04-04T11:35:22ZDennis Gläser[facet][mpfa][fix] fix additional derivatives for numEq > 13.1Martin SchneiderMartin Schneiderhttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1549Feature/improve facet grid manager2019-04-11T09:54:56ZDennis GläserFeature/improve facet grid managerThe facet coupling gmshreader/gridmanager did not work for grids where vertices of a lower-dimensional grid were not contained in the bulk grid. Also, if the mesh file only contained lower-dimensional elements grid creation failed. Both ...The facet coupling gmshreader/gridmanager did not work for grids where vertices of a lower-dimensional grid were not contained in the bulk grid. Also, if the mesh file only contained lower-dimensional elements grid creation failed. Both of these issues are adressed in this commit.
Is WIP because it has to wait for !1545.3.1Simon EmmertSimon Emmerthttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1545Some fixes for coupling moving meshes.2019-04-05T14:07:14ZSamuel Burbullasamuel.burbulla@mathematik.uni-stuttgart.deSome fixes for coupling moving meshes.These are the changes I needed so far to use the multidomain coupling with my moving mesh implementation.These are the changes I needed so far to use the multidomain coupling with my moving mesh implementation.3.1Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.nohttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1542Feature/improve facetcoupling on surface grids2019-04-23T14:19:22ZDennis GläserFeature/improve facetcoupling on surface gridsImplements support of the box-facet coupling scheme for surface grids. Furthermore removes support for xi != 1.0 on surface grids for cell-centered schemes as this condition does not generalize for surface grids.
This is WIP because it ...Implements support of the box-facet coupling scheme for surface grids. Furthermore removes support for xi != 1.0 on surface grids for cell-centered schemes as this condition does not generalize for surface grids.
This is WIP because it depends on !1541, where the segment intersections of segments with polygons embedded in 3d space will be implemented. After this is done, it remains to do:
- [x] Add test using box for three-domain test3.1Martin SchneiderMartin Schneiderhttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1541[geometryisection] introduce policy classes for intersections2019-04-05T13:52:45ZDennis Gläser[geometryisection] introduce policy classes for intersectionsIntroduces policy classes for geometry intersections. With the policy one can select what type of intersection is seeked (i.e. polygon, segment, point), each for which a different algorithm can be provided.Introduces policy classes for geometry intersections. With the policy one can select what type of intersection is seeked (i.e. polygon, segment, point), each for which a different algorithm can be provided.3.1Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.nohttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1538Improve tabulation for components and other functions2019-12-20T12:43:31ZTimo Kochtimokoch@math.uio.noImprove tabulation for components and other functions* [x] Make tabularized component work with any component (needs to check for availability of functions)
Fixes #676
Fixes #682 * [x] Make tabularized component work with any component (needs to check for availability of functions)
Fixes #676
Fixes #682 3.2Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.nohttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1537[Newton] Added retry time step reduction factor2019-03-27T18:13:37ZTimo Kochtimokoch@math.uio.no[Newton] Added retry time step reduction factor3.1Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.nohttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1536Feature/box prism grids2019-03-27T18:48:39ZMartin SchneiderFeature/box prism grids3.1Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.nohttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1534Improve TimeloopBase2019-04-03T05:17:41ZKilian WeishauptImprove TimeloopBase* Extends the interface of the abstract base class by the methods
`virtual void advanceTimeStep()`, `virtual void setTimeStepSize(Scalar dt)`,
`virtual Scalar maxTimeStepSize()`, and `virtual bool finished()`.
* Changes the internal...* Extends the interface of the abstract base class by the methods
`virtual void advanceTimeStep()`, `virtual void setTimeStepSize(Scalar dt)`,
`virtual Scalar maxTimeStepSize()`, and `virtual bool finished()`.
* Changes the internal logic of the time loop step size computation. The private variable maxTimeStepSize is no longer needed. Instead the maxTimeStepSize is computed in the method with the same name and can be overloaded. The new time step can thus always be determined as the minimum of a given timestep and the maximum time step
* Uses the base class in the assembler
fixes #679 3.1Kilian WeishauptKilian Weishaupthttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1529Improve Newton convergence writer, add implementation for staggered2019-03-26T12:59:04ZKilian WeishauptImprove Newton convergence writer, add implementation for staggered* [x] Adapt changelog
This contains a backward incompatible change: The convergence writer now needs the FVGridGeometry as template parameter instead of the GridView, and resize takes no arguments anymore. As the convergence writer is ...* [x] Adapt changelog
This contains a backward incompatible change: The convergence writer now needs the FVGridGeometry as template parameter instead of the GridView, and resize takes no arguments anymore. As the convergence writer is a rarely used feature this change shouldn't implicate many people.3.1Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.nohttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1526Feature/2p tracer2019-03-26T17:50:39ZSimon EmmertFeature/2p tracerContinued work from !1520 here
- renamed pIdx -> phaseIdx
- introduced saturation with enable_if for multiphase tracer (single phase stays untouched)
- use saturation in local residual
- use phaseName for output
- introduc...Continued work from !1520 here
- renamed pIdx -> phaseIdx
- introduced saturation with enable_if for multiphase tracer (single phase stays untouched)
- use saturation in local residual
- use phaseName for output
- introduced a test case (e.g. 2p lensproblem + tracer in water phase (maybe accumulates at fluid-fluid interface))
Closes #4983.1Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.nohttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1525Feature: Implement a shallow water model2019-05-02T00:47:55ZLeopold StadlerFeature: Implement a shallow water modelCloses #671
This is the reduced version (no friction, complex boundaries,...) for the shallow water model with a dam break as test case.Closes #671
This is the reduced version (no friction, complex boundaries,...) for the shallow water model with a dam break as test case.3.1Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.nohttps://git.iws.uni-stuttgart.de/dumux-repositories/dumux/-/merge_requests/1523Feature/improve box facet fluxes2019-03-19T14:45:23ZDennis GläserFeature/improve box facet fluxesComputation of the support point for flux computations across coupled facets using the box scheme now reuses the geometry intersection algorithms to compute an optimal point within the element. Therefore, 2d-1d intersection algorithms ha...Computation of the support point for flux computations across coupled facets using the box scheme now reuses the geometry intersection algorithms to compute an optimal point within the element. Therefore, 2d-1d intersection algorithms have been implemented.3.1Timo Kochtimokoch@math.uio.noTimo Kochtimokoch@math.uio.no