- Frans Basson
GEM4D Version 18.104.22.168 available for download
Updated: Dec 15, 2019
Version 22.214.171.124 contains smaller but significant updates to the markers and iso-surface capabilities, and adds the ability to do structure mapping with 1, 2, 3 and N mouse pick points. I also made some small interface changes to ensure the "Information" text box is always displayed.
Many thanks for the useful feedback I received from many users.
Mapped structural planes can be added with 1, 2, 3 or N points and their orientations are displayed on a polar plot (equal angle stereo-net). The created discs can be saved as a DXF-file, or the disc plane parameters can be saved as a CSV text file, which can then be loaded again as a "Mapping" file. This functionality is further discussed in the "Mapping with mouse picking" section below.
Meshes can be split on any picked polygon, where the selected polygon plane is used as the splitting plane. This functionality works well in conjunction with the discs discussed in Point 1, as a disc can be created with three pick points, and the surface then used to split a mesh triangulation into separate layers.
More options were added to the iso-surface creation parameters that determines the outcome: Resolution: This selection determines the number of sub-division "blocks", which in turn regulates the resolution of the created surface. A higher resolution will take longer to solve. Search radius: This selection determines the search distance from the centre of each "block" when interpolating an inverse distance value. Artefacts result if the search distance is too short, and the surface could be over-smoothed when the distance is too long. Longer search distances results in substantially longer calculation times.This functionality is further discussed in the "Iso-surface search radius" section below.
When loading files, a "Coordinates" pop-up box is always displayed where the coordinates system can be changed. When a loaded file is saved again, the file can now be saved in the transformed or original coordinate systems. The saved coordinate system is selected with "Save => Save in original coordinate system"and "Save => Save in new coordinate system".
The information text box is now placed separately and visible in all modes (file meshes, markers and mapping). Measurement and picking information is thus now visible in all "modes".
Determining the boundaries for iso-surface creation were previously done from the scene boundaries. If a mesh triangulation was loaded, the iso-surface boundaries would thus have included the triangulation and resulted in poor resolution around the actual data points. The new version calculates the boundaries for the iso-surfaces from the data boundaries only - as intuitively expected.
Interaction with the buttons around the iso-surface functionality were changed after use in various studies. These are small changes that determine when buttons become available and when values are reset. These changes should further improve the flow of iso-surface creation.
When loading "Marker" data, the default column "0" is now the Z-coordinate. Loaded values are thus coloured on the Z-coordinate by default, and other values follows as columns "1", "2", etc.
GEM4D attempts to automatically accommodate any coordinate system. This is reasonably involved to manage correctly under any circumstance, and coding changes were made to the logic for the latest version. These changes could have introduced bugs, so please let me know when anything suspect occurs.
Iso-surface search radius
The input data is typically in the form of an irregular grid, such as stress results, mapping results, logging data, observations etc. When the check box of (3) is pressed, a regular grid is generated, with the resolution based on the resolution value in the drop-down text-box below the slider, and values interpolated to the regular grid with the Shepard Method (inverse distance). These interpolated values are based on the selected search radius in the search radius drop-down text-box below the slider. An iso-surface is then created based on the interpolated values of the regular grid.
If the search radius is very short, only regular grid values close to actual data points will pick up values and the other points left blank. Visually, this results in a blocky iso-surface close to the irregular data points and looks artificial. When the search radius is too long, values are smoothed over too many values, and the visual outcome could be less accurate than for a shorter search radius.
Selecting the correct search radius is thus very important, and GEM4D attempts to select the appropriate value based on the data set. It is still valuable to choose different values, observed the impact on the iso-surface, and decide on the most appropriate value for the data set.
Mapping with mouse picking
Mapping discs in three dimensions are an valuable way to understand the potential impact of discontinuities on underground excavations and slope walls. As creating these discs in mine design software are typically a bit lengthy, I added 4 options to GEM4D to do this in a few mouse clicks.
Pick with 1 point
Only the location is picked in the scene; the dip, dip direction and trace length are thus requested in a pop-up form.
Pick with 2 points
Two points are picked and they determine the trace length and strike (thus also dip direction), and only the dip is requested by the pop-up form. The trace value now determines how far the disc will extend beyond the picked points.
Pick with 3 points
Three points are picked and they fully define the plane of the disc. The disc is centered around the average value of the picked coordinates. The disc radius is the distance from the average value and the picked point furthest away, plus the value selected in the trace text-box.
Pick with N points
Picked by N-points allows the gradual building of a plane by many points. One way to solve the plane is to build an average plane through all the selected points, but I found that does not work well in practice. The system I decided to use is simple and works well for my applications, let me know if you feel otherwise.
The disc orientation is determined by the first two and last selected points only, but the other selected points do determine the disc size. The way I normally use this, is to selected the plane strike with the first two points, next click at the extremities of the plane without worrying much about the orientation, and use the last pick to fix the dip direction. This system also allows many "last picks" until the plane visually fit the available data well.