# Utilities¶

Various utility routines are listed below.

## REGISTER_CALLBACK¶

This routine registers a Python callback later used by Solfec. All callback and data pairs must be registered in the same order on all MPI ranks, when Solfec is run in parallel. This function is ignored in sequential mode.

REGISTER_CALLBACK (callback | data)

- callback – callback function of form:
*value = callback (data)* - data – data passed to the callback function

## VIEWER¶

This routine tests whether the viewer is enabled.

obj = VIEWER ()

- obj – True or False depending on whether the viewer (–v command line option) was enabled

## SUBDIR¶

This routine returns the optional output subdirectory.

obj = SUBDIR ()

- obj – None object or subdirectory string depending on whether the –s command line option was used

## BODY_CHARS¶

This routine overwrites referential characteristics of a body.

BODY_CHARS (body, mass, volume, center, tensor)

- body – BODY object
- mass – body mass
- volume – body volume
- center – (x, y, z) mass center
- tensor – \(\left(t_{11},t_{21},\,...\,,t_{33}\right)\) column-wise inertia tensor for a rigid body or Euler tensor otherwise

## DELETE¶

This routine deletes a BODY object or a CONSTRAINT object from a SOLFEC object.

DELETE (solfec, object)

- solfec – SOLFEC object
- object (emptied) – BODY or CONSTRAINT object

## SCALE¶

This routine scales a geometrical object or a collection of such objects.

obj = SCALE (shape, coefs)

- obj – when shape is not (x, y, z) tuple: same as shape, returned for convenience. Otherwise the (x \(\cdot\) coefs[0], y \(\cdot\) coefs[1], z \(\cdot\) coefs[2]) tuple.
- shape – object, collection of objects, or a list [a, b, c, …] of objects of type CONVEX, MESH, SPHERE, ELLIP. Alternately this can be a single (x, y, z) tuple, but then one must use point = SCALE (point, coefs) in order to modify the point (Python tuples are immutable – they cannot be modified “in place” after creation).
- coefs – (cx, cy, cz) tuple of scaling factors along each axis

## TRANSLATE¶

This routine translates a geometrical object or a collection of such objects.

obj = TRANSLATE (shape, vector)

- obj – when shape is not (x, y, z) tuple: same as shape, returned for convenience. Otherwise the (x+vector[0], y+vector[1], z+vector[2]) tuple.
- shape – object, collection of objects, or a list [a, b, c, …] of objects of type CONVEX, MESH, SPHERE, ELLIP. Alternately this can be a single (x, y, z) tuple, but then one must use point = TRANSLATE (point, vector) in order to modify the point (Python tuples are immutable – they cannot be modified “in place” after creation).
- vector – (vx, vy, vz) tuple defining the translation

## ROTATE¶

This routine rotates a geometrical object or a collection of such objects.

obj = ROTATE (shape, point, vector, angle)

- obj – when shape is not (x, y, z) tuple: same as shape, returned for convenience. Otherwise the rotated (x1, y1, z1) image of (x, y, z).
- shape – object, collection of objects, or a list [a, b, c, …] of objects of type CONVEX, MESH, SPHERE, ELLIP. Alternately this can be a single (x, y, z) tuple, but then one must use point1 = ROTATE (point1, point2, vector, angle) in order to modify point1 (Python tuples are immutable – they cannot be modified “in place” after creation).
- point – (px, py, pz) tuple defining a point passed by the rotation axis
- vector – (vx, vy, vz) tuple defining a direction of the rotation axis
- angle – rotation angle in degrees

## SPLIT¶

This routine splits a geometrical object (or a collection of objects) by a plane passing by a point. Depending on the topological properties of the body shape and plane position this may or may not result in splitting of the body in two parts.

(one, two) = SPLIT (shape, point, normal | surfid, topoadj, remesh) (Experimental)

- one – objects placed below the splitting plane (
*None*if no objects were placed below the plane) - two – objects placed above the splitting plane (
*None*if no objects were placed above the plane, or if the initial shape has not been fragmented in two parts) - shape (emptied) – object, collection of objects, or a list [a, b, c, …] of objects of type CONVEX, SPHERE, ELLIP or MESH
- point – (px, py, pz) tuple defining a point passed by the splitting plane
- normal – (nx, ny, nz) tuple defining the splitting plane normal
- surfid – (surf1, surf2) tuple defining a pair of surface identifier for the two newly created surfaces (default: 0,0). Surface surf1 has the outward normal (nx, ny, nz).
- topoadj – ‘ON’ or ‘OFF’ (default: ‘OFF’); when ‘OFF’ the splitting will always propagate across the whole body and result in two body fragments; when ‘ON’ the splitting will propagate from the input point through the topologically adjacent elements, which may not produce fragmentation;
- remesh – ‘ON’ or ‘OFF’ (default: ‘ON’) flag used only for MESH based shapes; when ‘ON’ mesh splitting away from inter–element boundaries will lead to tetrahedral re–meshing; when ‘OFF’ it will raise an error.

*WARNING:* Mesh splitting generates tetrahedral mesh in place of the input one if the splitting plane
is not aligned with element boundaries. The meshing is randomized and it may generate different results
for the same input. Use TETRAHEDRALIZE in order to refine and save the generated mesh parts. Otherwise
may encounter input/output errors.

## MESH_SPLIT¶

This routine splits a mesh object along internal element boundaries whose nodes belong to the given node or face set. Depending on the topological properties of the mesh this may or may not result in splitting of the mesh in multiple parts.

[out1, out2, …] = MESH_SPLIT (mesh | nodeset, faceset, surfid1, surfid2) (Experimental)

- [out1, out2, …] – a list of output meshes (
*None*if no internal element boundaries in the input mesh were split) - mesh – input MESH object (the input mesh is not modified by this routine)
- nodeset – a list of nodes [n0, n1, n2, …] defining the splitting surface (zero based indexing); ignored when
**faceset**is passed - faceset – a list of lists face nodes [[n0, n1, n2], [n3, n4, n5, n6], …] defining the splitting surface (zero based indexing)
- surfid1 – surface identifier for the newly created surfaces (default: 0); used with
**nodeset**or outward–counter–clockwise–normal aligned with**faceset** - surfid2 – surface identifier for the newly created surfaces (default: 0); inward–counter–clockwise–normal aligned with
**faceset**

## COPY¶

This routine makes a copy of input objects.

obj = COPY (shape)

- obj – created collection of copied objects
- shape – object, collection of objects, or a list [a, b, c, …] of objects of type CONVEX, MESH, SPHERE, ELLIP

## BEND¶

This routine bends a shape around an axis. The bending is performed from the section of the shape closest to the axis onward. The orientation of the axis direction determines the orientation of the bending according to the right hand rule. Let \(\mathbf{q}\) be the closest to the axis mesh node. Then \(\mathbf{v}=\mathbf{d}\times\left(\mathbf{q}-\mbox{proj}\left(\mathbf{q}\right)\right)\), where \(\mathbf{d}\) is the axis direction and \(\mbox{proj\ensuremath{\left[\cdot\right]}}\) projects a point onto the axis. Bending starts from the section containing \(\mathbf{q}\) and proceeds in the direction of \(\mathbf{v}\).

obj = BEND (shape, point, direction, angle)

- obj – same as shape
- shape – object of type MESH
- point – axis point
- direction – axis direction
- angle – positive bending angle in degrees

## BYLABEL¶

This routine finds a labelled object inside of a SOLFEC object.

obj = BYLABEL (solfec, kind, label)

- obj – returned object (
*None*if a labelled object was not found) - solfec – SOLFEC object
- kind – labelled object: ‘SURFACE_MATERIAL’, ‘BULK_MATERIAL’, ‘BODY’, ‘FIELD’
- label – the label string

## MASS_CENTER¶

This routine calculates the mass center of a geometrical object or a collection of such objects.

obj = MASS_CENTER (shape)

- obj – (x, y, z) tuple storing the mass center
- shape – object, collection of objects, or a list [a, b, c, …] of objects of type CONVEX, MESH, SPHERE, ELLIP. Alternately this can be a single BODY object.

## LOCDYN_DUMP¶

This routine dumps into a file the most recent state of local dynamics. It is meant for debugging and test purposes, e.g. comparing local dynamics between runs on various processor counts.

LOCDYN_DUMP (solfec, path)

- solfec – SOLFEC object
- path – file path

## OVERLAPPING¶

This routine looks for shapes (not) overlapping the obstacles.

obj = OVERLAPPING (obstacles, shapes | not, gap)

- obj – list of shapes (not) ovrelapping the obstacles
- obstacles – object, collection of objects, or a list [a, b, c, …] of objects of type CONVEX, MESH, SPHERE, ELLIP
- shapes (emptied) – object, collection of objects, or a list [a, b, c, …] of objects of type CONVEX, MESH, SPHERE, ELLIP
- not – ‘NOT’ string
- gap – maximal negative gap

## MBFCP_EXPORT¶

This routine exports Solfec model into the MBFCP problem definition format. See this link for details.

MBFCP_EXPORT (solfec, path)

- solfec – SOLFEC object
- path – output path

## NON_SOLFEC_ARGV¶

This routine returns all command line arguments (in the form of a list of strings) that have been passed to ‘solfec’ or ‘solfec–mpi’ application and has not been identified as valid Solfec arguments. This way the user can pass some arguments to the input scripts.

argv = NON_SOLFEC_ARGV ()

- argv – list of non–solfec specific arguments passed to ‘solfec’ or ‘solfec–mpi’ command

## MODAL_ANALYSIS¶

This routine performs modal analysis of FEM bodies. The modal analysis results are stored with bodies and can be viewed.

obj = MODAL_ANALYSIS (body, num, path | abstol, maxiter, verbose)

- obj = (val, vec) – the returned tuple of: val = obj[0] list of eigenvalues and vec = obj[1] list of eigen vectors (stored contiguously one after another)
- body – input FEM body; the model analysis results are stored with this body
- num – number of lowest modes to extract
- path – path to file where the results will be stored (to avoid recomputing if possible).
Note, that if previous modal analysis results are found they are used rather then recomputed
if the number of modes and num are the same. If num is different from the previous modes count,
then new num modes is computed from scratch.
*Note:*“.h5” extension is automatically added to the stored path; - abstol – residual tolerance for the eigenvalue solver (default: 1E-11)
- maxiter – iterations bound for the eigenvalue solver (default: 100)
- verbose – ‘ON’ or ‘OFF’ verbosity flag for the eigenvalue solver (default: ‘OFF’)

## COROTATED_DISPLACEMENTS¶

This routine extracts snapshots of co–rotated displacements of FEM bodies. Co–rotated displacements factor out
the rigid body rotation, only including deformational motion about the initial configuration of the body. Multiple
calls to this command need to be used to extract multiple snapshot sets for distinct subsets of bodies. *Note 1:*
identical sequence of calls to this routine must be executed on all MPI ranks; *Note 2:* this routine is relevant
in the context of the ‘BC–RO’ FEM formulation, see BODY page and Table 4,
and Python’s modred package which can be used to calculate a reduced base;
before being passed to the modred package the outputted displacements
snapshots need to be complemented by the six rigid displacements generated by the RIGID_DISPLACEMENTS command (defined below);

obj = COROTATED_DISPLACEMENTS (solfec, subset | sampling) (Experimental)

- obj – upon termination of all RUN commands, a list of lists of displacement snapshots; this works both in ‘WRITE’ and ‘READ’ modes; MPI–parallel extraction of co–rotated displacement snapshots is enabled in the ‘WRITE’ mode: in this case only MPI rank 0 process will store a valid output list (None is returned for ranks > 0); in ‘READ’ mode enable the corotated_displacements flag of the FORWARD command in order to sample displacements while skipping forward through results;
- solfec – SOLFEC object
- subset – specification of a subset of bodies for which co-rotated displacements are to be extracted; a string can be used to define a POSIX regular expression [1] that will be matched against body labels; a list of body objects or integer body identifiers can be used [body1, body2, id3, id4, body5, …] mixed up in an arbitrary manner; or a tuple specifying extents of a bounding box can be used (xlow, ylow, zlow, xhigh, yhigh, zhigh), which the bounding boxes of exported bodies will overlapped at time t=0; also a list of an arbitrary combination of those can be used, e.g. [‘BOD*A’, 123, body1, body2, 256, (0, 0, 0, 1, 1, 1), ‘KEY??7’, (3, 3, 3, 4, 4, 4)] defines two labels, two integer body ids, two body objects, and two bounding boxes, that together define a subset of bodies that will be used during snapshot extraction; Note:* meshes of all bodies in the subset must have the same number of nodes;
- sampling – optional collection of time instants, e.g. [t0, t1, t2, …, tN], or a time interval, e.g. dt0, at which the displacement snapshots are to be sampled; default: same as the output interval, see OUTPUT

## RIGID_DISPLACEMENTS¶

This routine outputs six unit vectors of rigid displacements of a FEM body (three translations and three rotations).
*Note:* this routine is relevant in the context of the ‘BC–RO’ FEM formulation, see BODY page
and Table 4, and Python’s modred package which can be used to
calculate a reduced base; see also the COROTATED_DISPLACEMENTS command (defined above);

obj = RIGID_DISPLACEMENTS (body) (Experimental)

- obj – a list comprising six lists representing the unit displacement vectors
- body – a finite element BODY object

## BODY_MM_EXPORT¶

Export body matrices in the MatrixMarket sparse format.

BODY_MM_EXPORT (body, pathM, pathK | spdM, spdK)

- body – BODY object of ‘FINITE_ELEMENT’ kind
- pathM – output path for mass matrix M
- pathK – output path for stiffness matrix K
- spdM – symmetric positive definite flag M; ‘ON’ or ‘OFF’ (default: ‘ON’); only lower triangle is exported when ‘ON’
- spdK – symmetric positive definite flag K; ‘ON’ or ‘OFF’ (default: ‘ON’); only lower triangle is exported when ‘ON’

## DISPLAY_POINT¶

Attach a display point to a body. Display points are defined in reference configuration and convected with bodies. Display points can be visualised by selecting ‘display points on/off’ in the ‘tools’ viewer menu. They serve purely auxiliary purpose, for example allowing to make sure that the results are read from correct locations.

DISPLAY_POINT (body, point | label)

- body – BODY object
- point – referential (x, y, z) point
- label – optional label

## RENDER¶

Render selected bodies in the Viewer. *Note:* This *cannot* be used from within a normal analysis script,
but only from a Viewer script by selecting ‘run python script’ in the ‘tools’ viewer menu.

RENDER(solfec, object) (Experimental)

- solfec – SOLFEC object
- object – BODY object or a list of BODY objects

## REGISTER_BASE¶

Register ‘BC–RO’ or ‘BC–MODAL’ base for the finite element BODY formulation. Registering a reduced order or modal base saves memory when multiple instances of bodies employing the same base are used.

REGISTER_BASE (solfec, base, label) (Experimental)

- solfec – SOLFEC object
- base – base definition:
*(val, vec)*where*val*is a list of eigenvalues and*vec*is a list of eigenvectors (stored contiguously one after another) - label – unique string label

[1] | POSIX regular expressions |