PODIUM

Format Conversion:

​

PODIUM works natively in the Gmsh, Triangle and TetGen formats, and converts to and from the formats around them. Unstructured mesh files can be read and written as node, element, polygon and neighbour files, and exported as VTU for viewing in VISIUM (or ParaView). UBC-GIF meshes and models convert in both directions, as do a range of industry file formats and simple column-based data such as XYZ and CSV. Several files of the same format can be combined into one.

 

Data Processing:

​

Observation locations can be created on a grid, split into survey lines by separation distance and deviation angle, and given noise and standard deviations for synthetic studies. Data can be decimated by distance, against a mesh so that only one observation sits above each column of cells, or against a second set of points. Duplicates, nodes in a given range, and polynomial trends can all be removed, and coordinates transformed between systems, including longitude and latitude to UTM.

 

For understanding a dataset, PODIUM calculates correlations, linear regressions, cluster analyses and histograms, reports coordinate and attribute ranges, and produces difference statistics between two datasets. Terrain correction is available for vertical component gravity data, alongside data interpolation and geomagnetic field projection. Drillcore intervals can be converted from CSV into node and mesh files, and there are dedicated utilities for seismic and muon tomography sources, receivers and pierce points.

 

Model Processing:

​

Tetrahedral meshes can be built to typical forward and inverse modeling requirements, populated from point values, and extended with rectangular prisms, spheres and discs. A mesh can be made to conform to a topography surface and split into the volumes above and below it, then padded, coarsened, smoothed or subdivided as required.

 

Surfaces receive particular attention, with utilities for cleaning, sewing, snapping, ordering facet normals and finding holes, along with thresholding by attribute or edge length to extract the bounding surface of a body. Cells and nodes can be removed, regions reordered, and boundary outlines extracted.

 

For interrogating a model, PODIUM calculates cross-gradient, fuzzy c-means, total variation and spatial gradient measures, along with bulk magnetization, cell dimensions, surface volumes and cell counts. It will check whether a recovered model is sitting on its bounds, detect intersecting facets, identify physically connected groups of cells, produce difference statistics between two models, and interpolate model values at specified points. Values can be interpolated from one mesh to another, whether unstructured or rectilinear.

 

Inversion Input Files:

 

Cell-based and face-based weights can be set from the values in a model, and weights calculated for the robust modeling iterative re-weighting scheme. A set of builders writes the input files read by the MAGNUM programs FOGO, VIDI and DYNO, covering the mesh, physical property, data, optimization and joint coupling specifications.

 

Working in ZAMINEX:

 

Within ZAMINEX, every PODIUM utility has its own dialog with labelled fields and a help button on each parameter, so no command line syntax needs to be remembered. Programs are placed on a workflow canvas and connected into processing chains, with the status of each program shown on the canvas as the chain runs.

 

A complete workflow, with every parameter, saves to a project file, so a processing chain can be re-run months later or handed to a colleague. Notes can be recorded against each program and travel with the project. Many of the utilities are parallelised, and thread and MPI settings are configured in one place.

​

​