underworld.swarm.layouts module¶

This module contains classes for populating swarms with particles across the domain.

Module Summary¶

classes:¶

`underworld.swarm.layouts.ParticleLayoutAbstract`	Abstract class.
`underworld.swarm.layouts.PerCellSpaceFillerLayout`	This layout fills the domain with particles in a quasi-random pattern.
`underworld.swarm.layouts.GlobalSpaceFillerLayout`	This layout fills the domain with particles in a quasi-random pattern.
`underworld.swarm.layouts.PerCellGaussLayout`	This layout populates the domain with particles located at gauss locations within each element of the swarm’s associated finite element mesh.
`underworld.swarm.layouts.PerCellRandomLayout`	This layout fills the domain with particles in a random (per element) pattern.

Module Details¶

classes:¶

class underworld.swarm.layouts.ParticleLayoutAbstract(swarm, **kwargs)[source]¶

Bases: underworld._stgermain.StgCompoundComponent

Abstract class. Children classes are responsible for populating swarms with particles, generally across the entire domain.

Parameters:	swarm (underworld.swarm.Swarm) – The swarm this layout will act upon

swarm¶

Returns:	Swarm this layout will act to fill with particlces.
Return type:	underworld.swarm.Swarm

class underworld.swarm.layouts.PerCellSpaceFillerLayout(swarm, particlesPerCell, **kwargs)[source]¶

Bases: underworld.swarm.layouts._PerCellMeshParticleLayout

This layout fills the domain with particles in a quasi-random pattern. It utilises sobol sequences to generate per element particle locations which are more uniform than that achieved by a purely random generator.

Parameters:	swarm (underworld.swarm.Swarm) – The swarm this layout will act upon particlesPerCell (int) – The number of particles per element that this layout will generate.

Example

>>> import underworld as uw
>>> mesh = uw.mesh.FeMesh_Cartesian('Q1/dQ0', (1,1), (0.,0.), (1.,1.))
>>> swarm = uw.swarm.Swarm(mesh)
>>> layout = uw.swarm.layouts.PerCellSpaceFillerLayout(swarm,particlesPerCell=4)
>>> swarm.populate_using_layout(layout)
>>> swarm.particleLocalCount
4
>>> swarm.particleCoordinates.data
array([[ 0.5  ,  0.5  ],
       [ 0.25 ,  0.75 ],
       [ 0.75 ,  0.25 ],
       [ 0.375,  0.625]])

class underworld.swarm.layouts.GlobalSpaceFillerLayout(swarm, particlesPerCell, **kwargs)[source]¶

Bases: underworld.swarm.layouts.ParticleLayoutAbstract

This layout fills the domain with particles in a quasi-random pattern. It utilises sobol sequences to generate global particle locations which are more uniform than that achieved by a purely random generator. This layout is mostly useful where populating particles across a rectangular domain.

Parameters:	swarm (underworld.swarm.Swarm) – The swarm this layout will act upon particlesPerCell (float) – The average number of particles per element that this layout will generate.

Example

>>> import underworld as uw
>>> mesh = uw.mesh.FeMesh_Cartesian('Q1/dQ0', (1,1), (0.,0.), (1.,1.))
>>> swarm = uw.swarm.Swarm(mesh)
>>> layout = uw.swarm.layouts.GlobalSpaceFillerLayout(swarm,particlesPerCell=4)
>>> swarm.populate_using_layout(layout)
>>> swarm.particleLocalCount
4
>>> swarm.particleCoordinates.data
array([[ 0.5  ,  0.5  ],
       [ 0.25 ,  0.75 ],
       [ 0.75 ,  0.25 ],
       [ 0.375,  0.625]])

class underworld.swarm.layouts.PerCellGaussLayout(swarm, gaussPointCount, **kwargs)[source]¶

Bases: underworld.swarm.layouts.ParticleLayoutAbstract

This layout populates the domain with particles located at gauss locations within each element of the swarm’s associated finite element mesh.

Parameters:	swarm (underworld.swarm.Swarm) – The swarm this layout will act upon gaussPointCount (int) – Per cell, the number of gauss points in each dimensional direction. Must take an int value between 1 and 5 inclusive.

Example

>>> import underworld as uw
>>> # choose mesh to coincide with global element
>>> mesh = uw.mesh.FeMesh_Cartesian('Q1/dQ0', (1,1), (-1.,-1.), (1.,1.))
>>> swarm = uw.swarm.Swarm(mesh)
>>> layout = uw.swarm.layouts.PerCellGaussLayout(swarm,gaussPointCount=2)
>>> swarm.populate_using_layout(layout)
>>> swarm.particleLocalCount
4
>>> swarm.particleCoordinates.data
array([[-0.57735027, -0.57735027],
       [ 0.57735027, -0.57735027],
       [-0.57735027,  0.57735027],
       [ 0.57735027,  0.57735027]])
>>> import math
>>> # lets check one of these gauss points
>>> ( swarm.particleCoordinates.data[3][0] - math.sqrt(1./3.) ) < 1.e-10
True

class underworld.swarm.layouts.PerCellRandomLayout(swarm, particlesPerCell, seed=13, **kwargs)[source]¶

Bases: underworld.swarm.layouts._PerCellMeshParticleLayout

This layout fills the domain with particles in a random (per element) pattern.

Parameters:	swarm (underworld.swarm.Swarm) – The swarm this layout will act upon particlesPerCell (int) – The number of particles per element that this layout will generate. seed (int) – Seed for random generator. Default is 13.

Example

>>> import underworld as uw
>>> mesh = uw.mesh.FeMesh_Cartesian('Q1/dQ0', (1,1), (0.,0.), (1.,1.))
>>> swarm = uw.swarm.Swarm(mesh)
>>> layout = uw.swarm.layouts.PerCellRandomLayout(swarm,particlesPerCell=4)
>>> swarm.populate_using_layout(layout)
>>> swarm.particleLocalCount
4
>>> swarm.particleCoordinates.data
array([[ 0.24261743,  0.67115852],
       [ 0.16116546,  0.70790335],
       [ 0.73160516,  0.08792286],
       [ 0.71953113,  0.15966135]])