tem_tracking_module Module

Tracking enables the simulation to extract values of variables from within arbitrary parts of the mesh.

There maybe multiple tracking objects defined and each one needs to define the location to track (by a shape subtable), the variables to track (variable subtable). The points in time when the values should be evaluated need to be defined by a time_control object, see tem_timeControl_module for details. And finally the format of the output needs to be given in the output subtable. Each tracking object also gets a label to uniquely identify it, and a folder acts as a prefix, that gets prepended to all file names. Further it is possible to define some reduction operation, for example to build the average for the tracked variables over the complete shape, see tem_reduction_spatial_module.

The tracking table might describe a single tracking object, or it may be a table of tracking objects. If there are multiple tracking objects these must be simply listed, without keywords in the form of:

  tracking = {
    {label = 'atrack', ...},
    {label = 'btrack', ...},
    {label = 'ctrack', ...},
  }

If the table does not have this form, it will be assumed, that a single tracking object is defined.

Overview on the settings for the tracking object:

• label: A string to identify the tracking object, will be used to name files. Defaults to 'unnamed_track'.
• variable: A bable of variable names that should be tracked. Needs to be one of the available variables and has to be provided, ther is no default.
• folder: Actually a prefix that will be prepended to filenames. If it is to be a directory, it has to end in the path seperator and the directory has to exist already! Default: empty string, which results in the tracking files being written into the current working directory.
• reduction: A spatial reduction that is to be applied across the complete shape to arrive at a single value, see tem_reduction_spatial_module for details. If reductions are defined, there needs to be one for each variable that is tracked by this tracking object. Default is to not do any reduction.
• output: The description of how the data is to be written. There are three main options to format the output: ascii, vtk and harvester. The ascii format is useful for timeseries at a single point, that provides the point in time along with the respective variable values in an ASCII text file, easily processed by plotting tools. The vtk format is useful for larger shapes, like a slice through the domain. It provides the mesh information along with the values of the tracked variables. The harvester format writes the data of the subsection of the mesh that corresponds to the tree in the same binary format as used for restarting. The resulting data can then be processed by the harvesting tool of the solver to create visualizations. See the hvs_output_module for details on this subtable and other format options. This option has to be defined for each tracking object, there is no default.
• shape: Defines the part of the domain that is to be tracked by the object. There are various shapes available to define parts of the domain, but the most basic and common ones are points, lines and boxes, which we subsumize under the name canoND. Please see tem_shape_module for details.

A simple, single tracking object definition without reduction looks like this:

  tracking = {
    label = 'track_pointA',
    folder = './',
    variable = {'momentum','density','energy'},
    shape = {
      kind = 'canoND',
      object= {
        origin ={0.01, 0, 0}
      }
    },
    time_control = {
      min = 0,
      max = sim_control.time_control.max,
      interval = {iter = 10}
    },
    output = {
      format = 'ascii'
    }
  }

This tracks the variables momentum, density and energy at a single point over time and writes them every 10 iterations to disk with one line per point in time.

See the dedicated tracking page for more examples and further hints.