Refactored MAPDL Launcher

These release notes cover a variety of new features and bug fixes that have occured in the past year.

New Features

• Exchange vectors and arrays from numpy to MAPDL through the Mapdl class
• Better cyclic result processing, including cyclic strain, stress, temperature, etc.
• Element result plotting
• Support for newer ANSYS MAPDL releases (i.e. up to 2020R2)
• Support all element types in the result file.
• Interactive geometry plotting (e.g. KPLOT, APLOT, VPLOT) within pyansys using VTK. No reason to use open_gui for geometry scripting.
• Add support for Python 3.8
• Add sparse result file reading (necessary for latest ANSYS releases)

Bug fixes

• Element components are properly supported in result files
• Fix CORBA interface for ANSYS 2019R1
• Fix VTK 9 compatibility issues
• Fix various archive file formatting issues when reading.
• Fix negative/positive cyclic rotation issues.
• Correctly apply euler angles.
• Fix MAPDL executable finder.

API Changes

There's been a variety of API changes since the last release, and it's past time to list those changes and some of the reasoning behind those changes.

Use launch_mapdl instead of Mapdl

This is a bit more straightforward in terms of launching MAPDL since there are two methods for communicating with MAPDL at the moment, and it opens up MAPDL for additional interfaces

Remove uppercase methods from Mapdl

This was a necessary change to make the Mapdl class PEP8 compliant. While annoying to change, it's way better to provide a consistent interface like to other popular modules (numpy, scipy, matplotlib...)

Improved pyansys Interface and Translation

Added ability to translate scripts ANSYS input scripts directly to python scripts. See:
https://pyansys.readthedocs.io/en/latest/ansys_control.html#translating-scripts

Verified compatibility with ANSYS versions 15, 17, and 18.2 for

• Archive files
• Nodal results
• Element results

Bug with inaccurate nodal stress results for mixed nodes has been patched. pyansys now includes unit and functional testing to verify accuracy with ANSYS results.

See https://pyansys.readthedocs.io/en/latest/ for the latest documentation.

Added ANSYS Interface

Now able to spawn and control an instance of ANSYS. See the documentation below for more details:
http://pyansys.readthedocs.io/en/latest/ansys_control.html

Quick example:

import time
import pyansys

ansys = pyansys.ANSYS()

# create a square area using keypoints
ansys.Prep7()
ansys.K(1, 0, 0, 0)
ansys.K(2, 1, 0, 0)
ansys.K(3, 1, 1, 0)
ansys.K(4, 0, 1, 0)
ansys.L(1, 2)
ansys.L(2, 3)
ansys.L(3, 4)
ansys.L(4, 1)
ansys.Al(1, 2, 3, 4)

# Start a non-interactive plotting window
ansys.Show()
ansys.Menu('grph')
ansys.View(1, 1, 1, 1)  # sets the view to "isometric"

# Plot geometry while waiting between plots
ansys.Pnum('kp', 1)  # enable keypoint numbering
ansys.Pnum('line', 1)  # enable line numbering
ansys.Aplot()
time.sleep(1)  # could also use ansys.Wait(1)
ansys.Lplot()
time.sleep(1)
ansys.Kplot()
ansys.Save()
ansys.Exit()

Improved Full Reader

Mass and stiffness matrices from the full reader before 0.24 needed to have their constrained degrees of freedom removed from the arrays if they were constrained within ANSYS. These constrained degrees of freedom are now automatically removed.

The sort option has been added back. Sorting is disabled by default. See the latest release notes:
http://pyansys.readthedocs.io/en/latest/loading_km.html

Added single precision mesh quality check

Prior version needed to convert to double before checking cell quality.

Principal Stress Computation

Added principal stress computation and plotting. Example is in documentation and below:

pstress = result.PrincipalNodalStress(0)
result.PlotPrincipalNodalStress(0, 'SEQV')

See updated documentation at:
http://pyansys.readthedocs.io/en/latest/loading_results.html

Garbage Collection Fix and Updated Documentation

Due to limitations with vtkInterface, object were not being garbage collected when deleted. That's fixed now, along with the outdated documentation and examples.

Archive File Code Streamlining

ParseFEM has been removed entirely from the archive reader and new archive files can be read in and displayed with VTK using:

import pyansys

archive = pyansys.ReadArchive('tiny.cdb')
grid = archive.ParseVTK()
grid.Plot()  # plot

This approach simplifies much of the code and eliminates an unnecessary function. Information about the ansys archive file is now contained within the unstructured grid and can be accessed using the GetPointScalars function. See vtkInterface.

ParseVTK also allows you to specify element types to read in. See the doc_string of ParseVTK below:

Parses raw data from cdb file to VTK format.

Parameters
----------
force_linear : bool, optional
    This parser creates quadradic elements if available.  Set this to
    True to always create linear elements.  Defaults to False.

allowable_types : list, optional
    Allowable element types.  Defaults to:
    ['45', '95', '185', '186', '92', '187']

    Can include:
    ['45', '95', '185', '186', '92', '187', '154']

Returns
-------
uGrid : vtk.vtkUnstructuredGrid
    VTK unstructured grid from archive file.

Added Cyclic Plotting Results and Output to Paraview

See the documentation at https://pyansys.readthedocs.io/en/latest/loading_results.html#results-from-a-cyclic-analysis to plot cyclic results.

Also see https://pyansys.readthedocs.io/en/latest/loading_results.html#exporting-to-paraview for the steps to export and convert rst files to paraview.

Modified Binary Geometry, Matrix, and Result Reader

pyansys can now read component stresses from the result file. The mass and stiffness matrices can now be read using sorted and unsorted indexing. Additionally, the geometry can be directly read from the result file, making it unnecessary to load in a separate archive file.

See the full documentation Loading Results and Loading Full for details on the new API.