Difference between revisions of "Tutorials"

From FEAP Wiki
Jump to navigation Jump to search
 
(2 intermediate revisions by 2 users not shown)
Line 8: Line 8:
=== Batch commands needing user input ===
=== Batch commands needing user input ===
If you want to use a command in BATCh mode that requires user data,
If you want to use a command in BATCh mode that requires user data,
you need to place the [[data after the END|data after the end]].
you need to place the [[data after the end|data after the END]].


=== Nodal Stresses ===
=== Nodal Stresses ===
Line 16: Line 16:
FEAP allows one to solve problems using operator splitting solution strategies as are common with
FEAP allows one to solve problems using operator splitting solution strategies as are common with
multi-physics problems.  In FEAP the operator splitting is defined using [[partitions]].
multi-physics problems.  In FEAP the operator splitting is defined using [[partitions]].
=== Initial FEAP control record ===
The first line of the input files controls the number of nodes etc. in your problem.
There are two useful ways of setting the [[control values]].


== Time History Plots ==
== Time History Plots ==
Line 24: Line 28:
=== Collection files for paraview ===
=== Collection files for paraview ===
When making parallel runs, it is desirable to merge the results into a single view for visualization.  If one uses the <code>PVIEw,time</code> Marco command at the end of each time step, this will output paraview files per processor (with time stamps).  These can then be viewed if one makes a paraview [[collection file|Collection File]].
When making parallel runs, it is desirable to merge the results into a single view for visualization.  If one uses the <code>PVIEw,time</code> Marco command at the end of each time step, this will output paraview files per processor (with time stamps).  These can then be viewed if one makes a paraview [[collection file|Collection File]].
=== Partitioning of very large systems ===
In general, for the preparation of parallel runs, a partitioning of the whole domain is necessary. On that, the whole domain is loaded in a serial run of the parallel FEAP program (parfeap). For very large systems this can run in to segmentation errors or memory allocation errors. Thus, the 'largemem' option of FEAP must be utilized. A suitable workflow on that is shown in this tutorial.
Details coming soon


== FE2 Computations ==
== FE2 Computations ==

Latest revision as of 23:41, 13 December 2023

The following tutorials are designed for new FEAP users to gain a deeper understanding of some important capabilities/features of FEAP.

Introduction

FEAP Basics

Batch commands needing user input

If you want to use a command in BATCh mode that requires user data, you need to place the data after the END.

Nodal Stresses

FEAP can compute projections of Gauss point stresses to Nodal Stresses.

Operator Splitting

FEAP allows one to solve problems using operator splitting solution strategies as are common with multi-physics problems. In FEAP the operator splitting is defined using partitions.

Initial FEAP control record

The first line of the input files controls the number of nodes etc. in your problem. There are two useful ways of setting the control values.

Time History Plots

FEAP has a time history output mechanism called tplot.

Parallel Runs

Collection files for paraview

When making parallel runs, it is desirable to merge the results into a single view for visualization. If one uses the PVIEw,time Marco command at the end of each time step, this will output paraview files per processor (with time stamps). These can then be viewed if one makes a paraview Collection File.

Partitioning of very large systems

In general, for the preparation of parallel runs, a partitioning of the whole domain is necessary. On that, the whole domain is loaded in a serial run of the parallel FEAP program (parfeap). For very large systems this can run in to segmentation errors or memory allocation errors. Thus, the 'largemem' option of FEAP must be utilized. A suitable workflow on that is shown in this tutorial.

Details coming soon

FE2 Computations

Isogeometric Computations

Paraview Files

Compressing VTU files

By default, the PVIEw command in FEAP creates non-compressed ASCII VTU files. For large models, this may lead to very large VTU files. In these cases, compressing the ASCII VTU files into Binary VTU files using a proper compressor, results in around 80-90% of reduction in the final size of the post-processing data. Here, a Python script is provided which helps the user with this task. PythonCompressVTUFiles