FieldView
Reference Manual

SOFTWARE RELEASE
VERSION 2025

The GRID PROCESSING section on the Data Input panels allows users to interactively balance performance versus read time and memory. Settings can be easily changed, depending on usage, to maximize productivity.

When FieldView reads data, it performs grid processing after reading the grid. Grid processing increases FieldView's speed in such operations as creating and sweeping Coordinate Surfaces, computing accurate streamlines, and Dataset Sampling. The costs of grid processing are read time and memory. Previous versions of FieldView required setting environment variables to balance performance versus read time and memory. For a full list of operations benefiting from this initial grid processing, please refer to list of operations which will have performance impacted by grid processing in the FieldView User’s Guide.

GRID PROCESSING is shown on the PLOT3D and OVERFLOW-2 panels in Figure 2 and the FV-UNS panel in Figure 3. The slider defaults to a "Balanced" position between faster data input using less memory and faster performance using more memory. If you wish to post-process your data using many Coordinate Surfaces or streamlines, or if you’re going to use Dataset Sampling for comparing results, setting the slider to "More" will improve FieldView's performance. If performance of these features is less critical, setting the slider to "Less" will minimize read time and memory.

This Grid Processing setting is saved in, and read from, the Data Input restart file or FVX.

Reading results data in Replace or Append mode using most readers will automatically bring up the Function Subset Selection panel containing a list of the variable names from the results file.

All Function Names will be initially selected (highlighted). Clicking on a selected Function Name will deselect (unhighlight) it and vice versa.

For a transient dataset given as a series of files, FieldView will prompt you to read the whole series of files and treat them as a transient dataset.

A dataset will automatically be recognized as transient as long as a particular file naming convention is used. The file naming convention embeds the integer time step number (but not the solution time) in the filename. This is represented as ##### below. Note that you should use as many digits as required to represent your time steps. There is no requirement that five be used. FieldView looks for the embedded time step value (#####) to the left of the first 'dot' ( . ), if there is one in the file name, otherwise it will search from the right. The required format can be one of either: prefix#####.extension prefix.#####

Example: pipe04020.extension, pipe04021.extension, …​, pipe04073.extension

The actual extension used is not important.

For the readers that produce a single file containing transient data (AcuSolve, OpenFOAM, FLOW-3D® and FIDAP), FieldView will present you with the Time Step Selection panel (see Figure 1) when the data is read. This will show you the time steps found in the file(s) and allow you to choose which one you wish to initially read into memory. By default, FieldView selects the last time step to read in. You will have to explicitly choose a different time step if that is desired.

When turned on, the 'Read as Steady State' toggle button allows the selection of a specific time step of a multiple time step per file dataset.

This capability is particularly of interest for quickly animating particles computed for a transient dataset but with a fixed geometry over time. This can be done by:

Since only particles will be updated, it will be many times faster than a full transient animation, in which the volume data would be read and updated at each step.

The transient panel will be disabled and transient restrictions on particle path display options will be lifted. The Particle Path Animate button will also be enabled in this mode.

Note that in order for time step or solution time information to be available in a FieldView session (for example, to be included in an annotation using the escape sequence %%T or %%N), the dataset needs to be read as transient with the 'Read as Steady State' option turned off.

For all transient data, any time step can be accessed through the Transient Data Controls panel (see Transient Data Controls section of Working with FieldView for more information).

Changing of time step for a transient dataset will not cause all grid and results files to be re-read. Only the file or files containing the results of the current dataset will be re-read when the time step number is changed. The transient sweep reading of grid and results files (PLOT3D, FV-UNS-split) is optimized so that only the files that need to be changed for the new time step will be read. An invariant grid for the current dataset and all files for other datasets will remain in memory.

PLOT3D and OVERFLOW-2 data files can be saved using different file formats and attributes. A previous limitation of the FieldView data input panels for both types of data was that the file formats and attributes needed to be explicitly and correctly set before the file could be read. By default, FieldView will attempt to automatically detect the values for the FILE FORMAT, COORDS and DATA FORMAT. You still have the option to disable this default behavior by turning the Auto-Detect Format option OFF - in this case, explicit setting of the file format and attributes will be required.

If FieldView is unable to correctly detect the file format, the Auto-Detect Format button will be turned off, and the following pop-up message will appear:

Note that if Auto-Detect Format switch has been turned OFF, it will remain OFF unless you turn it back ON - it does not get reset back to the original default during the session.

Auto-Detect can only be used interactively or with FVX Data Input. The syntax for the FVX read_dataset command is:

Auto-Detect is NOT read or written to the data file restart.

Auto-Detect is enabled whether using Direct or Server, and supports non-parallel servers, parallel servers, and partitioned-file parallel (PFPR).

Often it is useful to review only the boundary surfaces for a dataset. By limiting the dataset read operation to read in just the boundary data, a significant reduction in the time required to read can be realized. Reading of boundary data only is available for all unstructured data readers, including user-defined (toolkit plugin) unstructured data readers.

If an unstructured data reader always has boundary data without any volume data, such as XDB Import, or STL [Direct Reader], then Read Boundary Data Only is always enabled and the check button is grayed out.

The time savings associated with Read Boundary Data Only will depend on the ratio of boundary to volume data, and will typically be in the range of 15x to 20x faster than reading the entire volume data for a given dataset.

Read Boundary Data Only is enabled by turning on the check button in the Data Input panel (see Figure 3). This setting is off by default. Read Boundary Data Only is fully compatible with FieldView Parallel. Partitioned File Parallel formats are also fully supported. This feature will also work correctly to examine extruded 2D boundary data, created using the FV_2D_TO_3D environment variable.

Recognize that all surface and rake types which rely on volume data cannot be created following a read boundary only data option.

Complete support has been provided for RESTARTS. However, it is important to keep in mind that it will not be possible to create any rakes or surfaces which need volume data. So, consider the case where a Complete RESTART has been saved for a full volume data read. If we attempt to apply this RESTART (ALL, NO_DATA_READ) following a Boundary Data only read, any surfaces based on volume data will not be present.

Regarding transient streaklines however, It is possible however to do the following:

In the above scenario, the time savings by not having to read in the volume data at each time will be considerable. In general, the time required to read Boundary Data only is dramatically reduced relative to the time needed to read the volume data for the same case.

This feature is currently limited to work with unstructured datasets only. Writing DataGuide^TM files is disabled when the read boundary only switch is on.

You can read multiple datasets into FieldView. This is done with the Append button that appears on all of the data input panels. This will allow you to read in, display and visualize several datasets at the same time.

The datasets that you read into memory do not have to be of the same type or format or have the same number of variables. One can be a structured dataset and another can be an unstructured dataset. However, a Results file Append will always try to read results for the highest-numbered dataset. If it succeeds, the current dataset will be changed to the highest-numbered dataset. A Results file Replace will (as in previous versions of FieldView) always try to read results for the first dataset. If it succeeds, the current dataset will be set to 1. Therefore, reading grid files and results files in the following order: Grid1, Grid2, Result1, Result2 will not yield the desired result in FieldView memory. The order Grid1, Result1, Grid2, Result2 will need to be used.

Switching between datasets can be accomplished on the Dataset Controls panel or the Main Toolbar (see the Tools section of Working with FieldView for more information). You can also switch between datasets by "quick-picking" Double-Click on any surface, rake, legend, etc. that belongs to that dataset.

Each dataset loaded into memory will have a separate set of functions (results) which will be shown when the Functions button is pressed. Variables of one dataset cannot be used in formulas for a different dataset (with the exception of Dataset Comparison mode).

For PLOT3D and split FieldView Unstructured files the grid will be automatically re-used when appending results that are based on the same grid file. We expect users to save significant disk space, and to see a substantial improvement in the handling of datasets where the grid is invariant with time, or across several datasets.

All surfaces and rakes created will belong to a specific dataset. If you have two datasets in memory and each dataset has a computational surface created in it, then both computational surfaces will be surface number 1, specific to the given dataset. A surface cannot be switched to a different dataset. Control of the current dataset is through the Dataset Controls panel or the Main Toolbar.

FieldView will let you make numerical comparisons between datasets. If the datasets are based on the same underlying grid, Dataset Comparison is directly possible and formulas spanning datasets can be created using the Function Specification Panel (see Function Specification Panel). If the underlying meshes differ, or if the datasets are based on different file formats, then the Dataset Sampling Tool can be used to create a sampled dataset which can then be used for numerical comparisons (see Dataset Sampling). Note that both Dataset Comparison and Dataset Sampling requires that when your input data is not read serially with the FieldView Client and remote or local servers are used, the Append Server button File.. → Data Input.. → Server Append option must be checked ON. This is the default. Note also that Dataset Sampling is not supported on parallel servers.

If more than one transient dataset is read into FieldView, it is possible to animate all time steps using merged solution times for all datasets during the transient sweep. This feature simplifies the task of creating an animation of multiple transient cases (see Use Merged Times).

This feature permits you to use the same SERVER process to read more than one dataset. To turn this feature on, first read a dataset into FieldView using the Client-Server feature. Then, to read another dataset to that same SERVER process, in the File menu Data Input pulldown, turn the Server Append button ON. With a minimum of two datasets read into FieldView in this way, the Dataset Sampling and Dataset Comparison features will be fully enabled, except that Dataset Sampling is not supported when using Parallel Servers. These features will be discussed in more detail in sections describing the Client-Server operation of FieldView.

It is known that files created on LINUX or Windows systems have different byte ordering compared to their UNIX counterparts. It is also often the case that datasets are created on one system, and post-processed on another. The FieldView Unstructured (FV-UNS) and PLOT3D readers permit reading of non-native byte ordered FV-UNS (all types) and PLOT3D (all types) files.

Arbitrary elements are now widely used by commercial CFD solvers. A common discretization strategy is to mesh the internal volume with regular hexahedral cells, the near-surface volume with extruded cell layers, and generate arbitrary polyhedra to fill the gap in between the two. By default, FieldView employs interpolation and gradient calculations natively for arbitrary polyhedra, consistent with other commercial solvers such as STAR, STAR-CCM+ and FLUENT. Tetrahedralization on-the-fly is only used in the following two cases:

The importance of native handling for arbitrary polyhedra is highlighted in the illustration below:

Holes in the tetrahedralized mesh (on the left) are clearly seen.

Special handling is implemented in FieldView to calculate derivatives (gradient, divergence and curl) for the case of split planar faces. These types of faces are commonly encountered in hexcore style meshes, where mesh resolution within the core regions change from coarser to finer cells.