Working with grids

This tutorial demonstrates basic features of dealing with grids in Bempp. Simple grids can be easily created using built-in commands. More complicated grids can be imported in the Gmsh format.

In order for the commands in this tutorial to work, Gmsh must be installed and the command gmsh must be available in the path.

Bempp currently only supports grid consisting of flat surface triangles.

Creating basic grid objects

We first import Bempp. We will also import Numpy as it will be needed later.

import bempp.api
import numpy as np

Built-in grids

The following command creates the sphere grid.

grid = bempp.api.shapes.regular_sphere(3)

The command regular_sphere creates a sphere by refining a base octahedron. The number of elements in the sphere is given by \(8 * 4^n\), where n is the refinement level.

Another option is to generate a sphere based on a given element diameter. For example, we can create a unit sphere with element diameter \(h=0.1\) with the command

grid = bempp.api.shapes.sphere(h=0.1)

Full documentation of Bempp’s available built-in grids can be found here.

Importing a grid

To import an existing grid file Bempp provides an import_grid command, e.g.

grid = bempp.api.import_grid('my_grid.msh')

to import a grid in Gmsh format. Bempp uses the file ending to recognize a number of grid formats. This works through the external meshio library. The meshio website contains details on supported formats. Frequently used formats with Bempp are msh (Gmsh), vtk (Legacy VTK), and vtu (VTK Xml Format).

Grids can also be generated from arrays containing vertex coordinates and connectivity information. For example, to create a grid consisting of two triangles with vertices \((0, 0, 0), (1, 0, 0), (0, 1, 0)\) and \((1, 0, 0), (1, 1, 0), (0, 1, 0)\) we use the following commands.

vertices = np.array([[0, 0, 0],
                     [1, 0, 0],
                     [0, 1, 0],
                     [1, 1, 0]],
elements = np.array([[0, 1, 2],
                     [1, 3, 2]],
grid = bempp.api.Grid(vertices, elements)

Note that we have specified the arrays to be transposed. The grid objects expects the vertices array to be of the form (3, M), where M is the number of vertices, and the elements array to be of the form (3, N), where N is the number of elements.

The array vertices contains the 3d coordinates of all vertices. The array elements contains the connectivity information. In this case the first triangle consists of the vertices 0, 1, 2, and the second triangle consists of the vertices 1, 3, 2. Optionally, we can specify a list domain_indices that gives different groups of elements the same id. This can be used for example to generate different types of boundary data on different parts of the grid, or to specify function spaces only on parts of the grid. In this example, both triangles automatically have the identifier 0 since nothing else was specified. This corresponds to the call

grid = bempp.api.Grid(vertices, elements, domain_indices=[0, 0])

Querying grid information

One can directly query a number of quantities about grids. In the following we demonstrate some of them.

The number of elements, edges and vertices are given by


To query the maximum and minimum element diameter use the following attributes.


The vertex indices of the 5th element can be obtained through

grid.elements[:, 5]

Hence, to display the vertex coordinates of the 5th element use

grid.vertices[:, grid.elements[:, 5]]

The area of the 5th element is shown as


The edge indices associated with the 5th element are given as


and we can see the definition of the edges in terms of vertex indices as

grid.edges[:, grid.element_edges[:, 5]]

This returns a (2, 3) array of the vertex coordinates associated with the three edges of the element. Edges in Bempp are counted such that the zeroth edge is the one connecting vertex 0 and 1. The first edge connects vertex 2 and 0, and the third edge connects vertex 1 and 2.

A number of other properties are defined, which are explained in more detail in the object description for the Grid class.

Plotting and exporting grids

To export a grid use the export command.

bempp.api.export('grid.msh', grid=grid)

This commands export the object grid as Gmsh file with the name grid.msh.

In order to plot a grid we can simply use the command


However, by default this works only within Jupyter notebooks. In order to enable plotting from command line sessions change the plot backend as

bempp.api.PLOT_BACKEND = 'gmsh'

This requires Gmsh to be available in the system path.