The check program computes and reports a variety of characteristics of the objects specified from the opened database. The characteristics which can be computed include mass, centroid, moments of inertia, volume, overlaps, surface area, exposed air, gaps/voids, unconfined air and adjacent air. Only the objects from the database specified on the command line are analyzed.
It works by shooting grids of rays from the three axis-aligned directions (sometimes called views) or a single view when azimuth/elevation angles are mentioned.
For volume/mass/surface area calculations the resulting calculations for each view are compared to each other. The grid of rays is progressively refined until the results from all three views agree within a user-specifiable tolerance, or a limit on grid refinement is reached.
For mass/centroid/moments of inertia calculations it is suggested for better accuracy to opt for three axis-aligned grid by not mentioning azimuth/elevation angles.
For surface area calculations, the rays are fired from three grids at random azimuth/elevation angles and then taken mean of the values reported by the three grids at the end.
Detects air volumes which are next to each other but have different air_code values applied to the region. This would typically indicate that the regions are different types of air, such as crew_air (which fills the crew compartment of a vehicle) and engine_air (which surrounds the engine). When these different types of air adjoin each other, it is generally considered a modeling error.
Computes the centroid of the objects specified.
This causes checks to be made to see if the ray encounters air regions before (or after all) solid objects. It also checks to see if the ray moves from a void to an air region. Typically, only the air inside a building or vehicle is modeled if the purpose of the model is to support analysis of that single structure/vehicle. There are exceptions, such as when modeling larger environments for more extended analysis purposes.
This reports when there is more than overlap_tol_dist (see the -t option below) between objects on the ray path. Note that not all gaps are errors. For example, gaps between a wheel and a fender are expected (unless external air is modeled). Typically, users should perform gap analysis on contained subsets of a model (such as passenger compartments) rather than on whole vehicles.
Computes the mass of the objects specified.
Computes the moments and products of inertia of the objects specified.
This reports overlaps, when two regions occupy the same space. In the real world, two objects may not occupy the same space. This check is sometimes also known as interference checking. Two objects must overlap by at least overlap_tol_dist (see the -t option below) to be considered to overlap.
Computes the surface area of the objects specified.
This reports when a partition with nonzero air code follows or precedes another partition and the space between them is more than overlap_tol_dist (see the -t option). The data reported are the names of the two regions (and solids) involved, the length of the gap along the ray, and the model coordinates of the ray's exiting the first partition and entering the second
Computes the volume of the objects specified.
If mass calculation is selected, a value is calculated and reported for each object specified on the command line. Note that if there are overlaps or other errors in the geometry, the values reported will be invalid.
For mass computation, the density of every region must be specified. Densities are specified as an index in a table of density values. This index is stored in the GIFTmater attribute of each region (typically set with the edcodes or adjust commands in MGED).
The density table consists of three columns:
This is the value to which the GIFTmater attribute will be set to select this material for the region.
This is the density for the material, and is specified in grams/cc.
This is a name or description the material.
An example file might look like the following:
#******************************************************************************
# sample .density file for check, gqa and rtweight
#******************************************************************************
#
# *** IMPORTANT: This file may change over time and should not be
# *** relied upon for production analysis work.
#
# All densities are in units of g/cm^3.
#
#******************************************************************************
#
# To use the file with rtweight, copy the file to your working directory:
# cp GQA_SAMPLE_DENSITIES .density
#
# To use the file with check/gqa, import the file data as a binary object:
# mged file.g bo -i u c _DENSITIES GQA_SAMPLE_DENSITIES
#
#******************************************************************************
2 7.82 Carbon Tool Steel
3 2.7 Aluminum, 6061-T6
4 2.74 Aluminum, 7079-T6
5 8.9 Copper, pure
6 19.32 Gold, pure
7 8.03 Stainless, 18Cr-8Ni
8 7.47 Stainless 27Cr
9 7.715 Steel, tool
10 7.84 Carbon Steel
12 3.00 Gunner
14 10.00 Fuel
The table is typically created in an external file using a text editor.
The geometry editor MGED automatically assigns an index value of
1
to a newly created region. While this default can be handy
when a vast majority of objects are made from the same material, it can lead to
surprising errors when objects which are supposed to have a
certain mass are computed to have different mass because one or two regions
were not set to the correct, non-default index value. As a result, it is advised that
the index value 1
never be used. If this practice is followed,
then an error message will be generated for any regions which have not had their
material index set to something other than the default.
The user will typically want to run gqa and verify the results using the
-f
option (see below) before importing the table into the database.
For example, if a material index is left out of the table, it is easier to rectify
the situation using the external file. Once the table has been verified as correct
and complete, it is imported to the database as the binary object
_DENSITIES .
To import the text file into the database, the following command is used:
mged>
bo -i u c _DENSITIES
filename
All of these calculations run until the grid refinement limit is reached.
For each pair of regions that cause an error, the tool reports the two erroneous regions, the maximum line-of-sight thickness of the error, and the in-hit location of the ray that caused that maximum error thickness.
-a
azimuth_deg [deg|rad]
Sets a rotation (in degrees) of the coordinate system by a given amount about
the Z axis. When mentioned, check shoots only one grid of rays along the
azimuth/elevation angle. The default is 35. See also -e
.
-e
elevation_deg [deg|rad]
Sets a rotation (in degrees) of the coordinate system by a given elevation
from the XY plane (rotation about X axis?). When mentioned, check shoots only one
grid of rays along the azimuth/elevation angle. The default is 25.
See also -a
.
-d
Enables debugging (off by default).
-f
filenameSpecifies that density values should be taken from an external file instead of from the _DENSITIES object in the database. This option can be useful when developing the density table with a text editor, prior to importing it to the geometric database.
-g
[initial_grid_spacing-]grid_spacing_limit or
[initial_grid_spacing,]grid_spacing_limitSpecifies a limit on how far the grid can be refined and optionally the initial spacing between rays in the grids. The first value (if present) indicates the initial spacing between grid rays. The mandatory argument, grid_spacing_limit, indicates a lower bound on how fine the grid spacing may get before computation is terminated. In general, the initial_grid_spacing value should be an integer power of the grid_spacing_limit. So for example, if grid_spacing_limit has the value 1, then any initial_grid_spacing specified should be in the sequence 2, 4, 8, 16, 32... so that the grid will refine to precisely the lower limit. The grid spacing may be specified with units. For example: 5 mm or 10 in . If units are not provided, millimeters are presumed to be the units.
The default values are 50.0 mm and 0.5 mm, which is equivalent to specifying:
-g 50.0mm-0.5mm
or -g 50.0mm,0.5mm
on the command line. This is a hard limit. If other analysis constraints are not
met, the grid spacing will never be refined smaller than the minimum grid size to
satisfy another constraint. The initial grid spacing is divided in half at each
refinement step. As a result, if you desire a lower limit to actually be tested, then
the initial grid size must be a power of 2 greater. For example, specifying -g10mm,1mm
would result in grid spacings of 10, 5, 2.5, 1.25 being used. If the goal was
to exactly end at a 1mm grid, then values such as 8 or 16 should have been
chosen for the initial values. This would result in testing 16, 8, 4, 2, 1
grid spacing values.
-G
[grid_width,]grid_heightsets the grid size, if only grid width is mentioned then a square grid size is set.
-i
Gets 'view information' from the view to setup the eye position of the single grid. Used only for overlaps calculations.
-M
mass_tolerance[units]
This is like the volume tolerance, -V
, but is applied to the mass
computation results, not the volume computation results.
The mass computation tolerance is probably more appropriate when doing whole-vehicle analysis. If mass computation is selected, it is set to a value equal to the mass of an object 1/100 the size of the model, which is made of the most dense material in the table.
-n
num_hitsSpecifies that the grid be refined until each region has at least num_hits ray intersections. It applies only when mass or volume calculations are being performed. This limit is not applied per-view, but rather per-analysis. So, for example, it is accepted that a thin object might not be hit at all from one view, but might be hit when it is shot from other views.
The default is 1. Hence, each region must be intersected by a ray at least once during the analysis.
-N
num_viewsSpecifies that only the first num_views should be computed. This is principally a debugging option.
-o
Specifies to display the overlaps as overlays.
-p
Specifies that check should produce plot files for each of the analyses it performs. These can be overlaid on the geometry in mged with the overlay command to help visualize the analysis results. Each of the different analysis types write to a separate plot file and use different colors for drawing.
-P
ncpuSpecifies that ncpu CPUs should be used for performing the calculation. By default, all local CPUs are utilized. This option exists primarily to reduce the number of computation threads from the machine maximum. Note that specifying more CPUs than are present on the machine does not increase the number of computation threads.
-q
Quiets (suppresses) the "was not hit" reporting.
-r
Indicates that check should print per-region statistics for mass, volume and surface area as well as the values for the objects specified on the command line.
-R
Disables the reporting of overlaps. Used only for overlaps sub-command.
-s
surf_area_toleranceSpecifies a surface area tolerance value that the three view computations must be within for computation to complete. If surface area calculation is selected and this option is not set, then the tolerance is set to 1/1,000 of the estimated surface area of the model bounding box.
-S
samples_per_model_axis
Specifies that the grid spacing will be initially refined so that at least
samples_per_axis_min will be shot along each axis of
the bounding box of the model. For example, if the objects specified have a bounding
box of 0 0 0 -> 4 3 2 and the grid spacing is 1.0, specifying the option
-S 4
will cause the initial grid spacing to be adjusted to 0.5 so that
4 samples will be shot across the Z dimension of the bounding box. The default is to ensure
1 ray per model grid axis.
-t
overlap_tolerance
Sets the tolerance for computing overlaps. The overlap_tolerance
must be a positive number equal to or greater than 0.0. Any overlap smaller than the value
specified will be ignored. The default value is 0.0, which causes any overlap to be
reported/processed. The value may be specified with a unit specifier such as:
-t 1.0mm
or -t 0.25in.
-U
use_air
Specifies the Boolean value (0 or 1) for use_air
which indicates whether regions which are marked as "air" should be retained and included
in the raytrace. Unlike other BRL-CAD raytracing applications,
the default is to retain air in the raytracing. The -U 0
option causes air regions to be discarded prior to raytracing. If you turn off use_air, and
request any analysis that requires it (see -A
above), then the program will
exit with an error message.
-u
distance_units,volume_units,mass_units
Specify the units used when reporting values. Values must be comma delimited and provided
in the order distance_units,volume_units,
mass_units. For example: -u "cm,cu ft,kg"
or -u ,,kg
(The latter example sets only the mass units.)
Note that unit values with spaces in their names such as cu ft
must be contained in quotes for the shell to keep the values together.
The default units are millimeters, cubic millimeters, and grams.
-v
Turns on verbose reporting of computation progress. This is useful for learning how the computation is progressing, and what tolerances are causing further computation to be necessary.
-V
volume_tolerance[units]Specifies a volumetric tolerance value that the three view computations must be within for computation to complete. If volume calculation is selected and this option is not set, then the tolerance is set to 1/1,000 of the volume of the model bounding box. For large, complex objects (such as entire vehicles), this value might need to be set larger to achieve reasonable runtime (or even completion). Given the approximate sampling nature of the algorithm, the three separate view computations will not usually produce identical results.
Example 1. Specifying Grid and Target Objects
The following will check objects hull, turret, and suspension for overlaps. The grid starts at 1 cm and is refined to 1 mm.
check overlaps -g 1cm-1mm hull turret suspension
Example 2. Specifying Using Non-Default Units
The following computes volume of hull, turret, and suspension. Results are reported in cubic centimeters (cc). The grid spacing starts at 5 in. and will not be refined below 0.3 mm spacing.
check volume -g5in-0.3mm -u ft,cc,oz hull turret suspension
For an example of some independent analysis type, consider the following:
%
check overlaps -g50,50 -u m,m^3,kg overlaps
Units:
length: m volume: m^3 weight: kg
grid spacing 50mm 199 x 199 x 199
OVERLAP PAIRS
------------------------------------------
/overlaps/overlap_obj.r and /overlaps/closed_box.r
</overlaps/overlap_obj.r , /overlaps/closed_box.r>: 32039 overlaps detected, maximum depth is 8m
==========================================
SUMMARY
32039 overlaps detected
1 unique overlapping pair (1 ordered pair)
Overlapping objects: /overlaps/overlap_obj.r /overlaps/closed_box.r
2 unique overlapping objects detected
%
check exp_air -u m,m^3,kg exposed_air.g
Units:
length: m volume: m^3 weight: kg
grid spacing 50mm 199 x 199 x 199
list Exposed Air:
/exposed_air.g/exposed_air.r count:25921 dist:9m @ (10000 1000 1000)
%
check unconf_air -u m,m^3,kg unconf_air.g
Units:
length: m volume: m^3 weight: kg
grid spacing 50mm 199 x 199 x 199
list Unconfined Air:
/unconf_air.g/air1.r /unconf_air.g/air2.r count:23921 dist:7m @ (10000 1000 1000)
%
check gap -u m,m^3,kg gap.g
Units:
length: m volume: m^3 weight: kg
grid spacing 50mm 199 x 199 x 199
list Gaps:
/gap.g/closed_box.r /gap.g/closed_box.r count:26082 dist:8m @ (9000 1000 1000)
/gap.g/adj_air2.r /gap.g/closed_box.r count:25921 dist:4m @ (1000 5000 1000)
%
check volume -u m,m^3,kg closed_box.r
Units:
length: m volume: m^3 weight: kg
setting volume tolerance to 1 m^3
grid spacing 50mm 199 x 199 x 199
grid spacing 25mm 399 x 399 x 399
grid spacing 12.5mm 799 x 799 x 799
closed_box.r 484.195 m^3
Average total volume: 488.327 m^3
%
check surf_area -u m,m^3,kg closed_box.r
Units: length: m volume: m^3 mass: kg
Using estimated surface area tolerance 640000 mm^2
grid: (50, 50) mm, (278, 278) pixels
grid: (50, 50) mm, (278, 278) pixels
grid: (50, 50) mm, (278, 278) pixels
Surface Area:
closed_box.r 384.485 m^2
Average total surface area: 384.485 m^2
%
check weight -u m,m^3,kg closed_box.r
Units:
length: m volume: m^3 weight: kg
setting weight tolerance to 768000 kg
grid spacing 50mm 199 x 199 x 199
Weight:
closed_box.r 3.6375e+06 kg
Average total weight: 3.67541e+06 kg
Reports of bugs or problems should be submitted via electronic
mail to <devs@brlcad.org>