diff --git a/doc/content/media/news/mar2019/Picture1.png b/doc/content/media/news/mar2019/Picture1.png new file mode 100644 index 0000000000..3bf2a6fb20 Binary files /dev/null and b/doc/content/media/news/mar2019/Picture1.png differ diff --git a/doc/content/media/news/mar2019/Picture2.png b/doc/content/media/news/mar2019/Picture2.png new file mode 100644 index 0000000000..cb6ea663b3 Binary files /dev/null and b/doc/content/media/news/mar2019/Picture2.png differ diff --git a/doc/content/media/news/mar2019/Picture3.png b/doc/content/media/news/mar2019/Picture3.png new file mode 100644 index 0000000000..ee9e1ee6f5 Binary files /dev/null and b/doc/content/media/news/mar2019/Picture3.png differ diff --git a/doc/content/media/news/mar2019/Picture4.png b/doc/content/media/news/mar2019/Picture4.png new file mode 100644 index 0000000000..2195a6705c Binary files /dev/null and b/doc/content/media/news/mar2019/Picture4.png differ diff --git a/doc/content/media/news/mar2019/Picture5.png b/doc/content/media/news/mar2019/Picture5.png new file mode 100644 index 0000000000..ee61399de1 Binary files /dev/null and b/doc/content/media/news/mar2019/Picture5.png differ diff --git a/doc/content/news/2019-03.md b/doc/content/news/2019-03.md new file mode 100644 index 0000000000..4f75b613e8 --- /dev/null +++ b/doc/content/news/2019-03.md @@ -0,0 +1,64 @@ +# MASTODON Newsletter (Spring 2019) + +## Explicit Integrator in MOOSE + +!row! + +!col! small=12 medium=6 large=6 +Made progress in implementing an explicit solver that will significantly reduce runtime. The explicit solver is benchmarked by comparing results of simple problems with the corresponding implcit, Newmark-Beta solution. The explicit solver is expected to be merged into the MOOSE framework in summer 2019 after thorough testing and will include both consistent and lumped-mass matrix options. + +The implementation comprises changes to the MOOSE framework and includes a new, `CentralDifference` TimeIntegrator. The implementation includes several other changes that make the usage of the explicit solver extremely easy to use. The user is simply required to change the TimeIntegrator in the Executioner and add the input 'implicit = false' in the strain kernels. Sample results for a single solid element problem are presented in [image001]. +!col-end! + +!col! small=12 medium=6 large=6 +!media news/mar2019/Picture1.png style=width:400px;margin-right:50px id=image001 + caption=Comparison of the results of the implicit and explicit integration for a single solid element problem. +!col-end! + +!row-end! + +## Beam Mesh Generator + +!row! + +!col! small=12 medium=6 large=6 +A beam mesh generator has been created and is currently under testing. The mesh generator will input nodal co-ordinates and beam connectivity information through csv files and will create a mesh automatically. This circumvents the usage of a meshing software and provides a much easier way to import meshes from traditional structural engineering software like SAP2000 or GTStrudl. +[image002] shows an L-shaped mesh and the corresponding input csv files created for testing. More tests and examples will follow. This will be implemented in summer 2019. +!col-end! + +!col! small=12 medium=6 large=6 +!media news/mar2019/Picture2.png style=width:300px;margin-left:50px id=image002 + caption=Illustration of a beam mesh generator example. +!col-end! + +!row-end! + +## Shell Element Implementation + +!row! + +!col! small=12 medium=6 large=6 +Implementing strain calculator for shell implements that computes the strain tensor in the shell local coordinate system. Modified the strain calculator so that the strains are calculated in the element local coordinate system. This will enable use of the shell element with the existing material models in the long run, through the use of the weak plane stress kernel. Added the stress calculators and linear elastic isotropic elasticity tensor class for shell elements that has weak plane stress assumption built in it. Also, added stress divergence kernel for shell elements using the stresses and test function gradients calculated in the shell local coordinate system. Began testing it on bending of a 4 element cantilever beam bending problem from Dvorkin and Bathe paper. The residual is diverging, mostly likely due to a bug in the linear elastic isotropic elasticity tensor for the shell. +!col-end! + +!col! small=12 medium=6 large=6 +!media news/mar2019/Picture3.png style=width:300px;margin-left:50px id=image003 + caption=??? +!col-end! + +!row-end! + +## MASTODON VectorPostProcessors + +!row! + +!col! small=12 medium=6 large=6 +Extended the Response History Builder, Response Spectra Calculator and HSI Calculator vector post processors to accept sets of nodes, or boundaries at input. Users do not have to create an individual VPP block for history, or spectrum at each node anymore. This makes the MASTODON `ResponseHistoryBuilder`, `ResponseSpectraCalculator`, and the `HSICalculator` more user friendly. Illustrations and examples are provided in the corresponding documentation pages. +!col-end! + +!col! small=12 medium=6 large=6 +!media news/mar2019/Picture5.png style=width:300px;margin-left:50px id=image004 + caption=Response histories of two nodes calculated from a single ResponseHistoryBuilder block. +!col-end! + +!row-end! diff --git a/doc/content/news/index.md b/doc/content/news/index.md index e76889fb3f..970e766c8d 100644 --- a/doc/content/news/index.md +++ b/doc/content/news/index.md @@ -1,3 +1,5 @@ # MASTODON Newsletter -[October News](2018-10.md) \ No newline at end of file +[Fall 2018 News](2018-10.md) + +[Spring 2019 News](2019-03.md)