ODB ➡️ VTK utility

💡 Info - this is an utility to read an Abaqus odb and convert the mesh and field data into a vtk format. This is useful to convert data directly on the HPC where the odb is generated for remote visualization through Paraview.

Usage

• Python 2 : for older versions of pvpython (e.g. 5.2.0). Check your paraview installation.
• Python 3 : works in more recent versions of pvpython (e.g. 5.7.0+).
• The different versions differ only on the vtk side of things, specifically on vtkTools and how printing to console is handled. the odb side of things needs to run in the abaqus python environment which is python 2.
• Note: The odbExtractor script will attempt to run two pvpython commands in the end. In some systems this fails. Steps 2 and 3 below are only needed in those systems where the command execution from python is blocked.

abaqus cae noGui=odbExtractor.py -- -o <odbName>.odb -r <requestName>.json -f all -d true

[optional - see note above]

pvpython vtkFromOdb.py <odbName>.odb

[optional - see note above]

pvpython vtkTimeSeriesBuilder.py <odbName>.odb

Environment and dependencies

• Abaqus installation: tested on Abaqus 2019;
• Paraview installation with pvpython: tested on paraview 5.2.0, 5.8.0 and 5.9.1;

Execution options

• -o (--odb) : odbName.odb

  • specify the name of the odb to extract data from

• -r (--request): requestName.json

  • specify the name of the json file outlining the frames to be extracted. This is specified in the frameIdList as a list of integer pairs [step number, frame number]. Index -1 can be used to specify last entry.

  {
    "frameIdList": [
      [1, 1],
      [1, 2],
      [2, -1],
      [-1, 3]
    ]
  }

• -f (--fields): disp or all (default: all)

  • specify the fields to be extracted. Currently supported options are displacement only ("disp") or every supported field ("all" or anyother string value). Currently supported fields are:

• -d (--duplicate): true or false (default: true)

  • specify whether to duplicate the odb leaving a copy intact to avoid data loss in case of odb corruption.

Features available

Current feature set is limited, but expanding:

• supported fields:

  scalar  = ['SDV', 'FV', 'UVARM']
  vector  = ['U']
  tensor  = ['S', 'LE']
  contact = ['CSDMG']

• supported elements: ['SC8R', 'SC6R', 'CSS8', 'S4R']

• supported stress state: plane stress with the following invariants and components ['11', '22', '33', '12', 'MAX_INPLANE_PRINCIPAL']

• filters / post-processing : currently only envelopes are computed for all tensor fields.

• vector fields are stored with the respective components as vtkFloatArray with 3 components - Paraview can make use of this for deformed representations and so on.

• frame times are saved and a seperate PVD file is created with the time series data containing the vtkMultiBlockDataSets;

Expansion

• Post processing steps can be added in odbFieldExtractor;
• Element types can be added in odbMeshExtractor - note that for different stress states changes need to be made in odbFieldExtractor and for different element geometries (other than hex, wedge and quad) changes need to be made to vtkGridBuilder