What are point zones?

Point zones are subsections of the computational domain which are selected based on a given geometrical criterion. A point zone is defined by the point_zone_t abstract type. Each point_zone_t object has a unique name attribute, and a mask containing a list of linear indices referring to the GLL points whose coordinates verify the above-mentioned geometrical criterion. Zones can then be used for different purposes, an example being applying a localized source term or probing a particular zone of interest.

Predefined geometrical shapes

There are three predefined shapes from which to initialize a point zone in the case file: boxes, spheres and cylinders. Each shape is described by its own subtype box_point_zone_t, sphere_point_zone_t and cylinder_point_zone_t, extending the abstract class point_zone_t.

Box

A box is defined from its x,y and z boundaries.

[
    {
        "name": mybox,
        "geometry": "box",
        "x_bounds": [-1.0, 1.0],
        "y_bounds": [-1.0, 1.0],
        "z_bounds": [-1.0, 1.0]
    }
]

Sphere

A sphere is defined by its center and its radius.

[
    {
        "name": "mysphere",
        "geometry": "sphere",
        "center": [0.0, 0.0, 0.0],
        "radius": 0.01
    },
]

Cylinder

A cylinder is defined by its end points and its radius.

[
    {
        "name": "mycylinder",
        "geometry": "cylinder",
        "start": [0.0, 0.0, 0.0],
        "end": [0.0, 0.0, 1.0],
        "radius": 0.01
    },
]

Operations on point zones

Inversion

By setting the keyword invert to true, you can indicate that you want to select all the points but those inside the point zone.

[
    {
        "name": "mycylinder",
        "geometry": "cylinder",
        "start": [0.0, 0.0, 0.0],
        "end": [0.0, 0.0, 1.0],
        "radius": 0.01
        "invert": true,
    },
]

Combination

It is possible to combine predefined shapes into a single point zone. This is achieved by the combine_point_zone_t type which can construct a point zone based on the geometrical parameters of each sub point zone or refer to already existing point zones. The mask of a combined point zone will not contain any duplicate points in case the subset point zones overlap.

Specifying which point zones to combine is done through the subsets array. In that array, you can specify:

a reference to an already existing point zone, or
an internal point zone.

Note: The first method is useful for creating a new point zone out of a combination of already existing point zones. If you are referencing zones p1 and p2, you will get a new point zone p3. All three of those will be accessible in the registry. The second method is useful if you want to create only one point zone using the geometric criteria of the internal ones. This means that if you create zone p3 out of internal zones p1 and p2, only p3 will be created. This is useful if you do not explicitly need p1 and p2, which will not be accessible in the registry. Use this method if you are using very large point zones and want to save memory.

In the code snippet below, the combine point zone named "combined_box_and_sphere" uses a subset of two zones:

"mysphere" is a reference to the external point zone above, and
"mybox" is built internally. This means that "mysphere" will be accessible in the registry, whereas "mybox" will not.

The combination will be performed according to a specified operator, which can take the values "OR", "AND" or "XOR". For example, if you try to combine two point zones p1 and p2:

The operator OR will select all the points that are both in p1 or p2,
The operator AND will select all the points that are both in p1 and p2,
The operator XOR (exclusive OR) will select all the points that are either in p1 or in p2, but not both at the same time.

Note: You can also invert a combine point zone.

Attention: You can combine other combine point zones as long as they are in the correct order in the case file, as demonstrated in the code snippet below.

[
    {
        "name": "mysphere",
        "geometry": "sphere",
        "center": [0.0, 0.0, 0.0],
        "radius": 0.01
    },
 
    {
        "name": "combined_box_and_sphere",
        "geometry": "combine",
        "operator": "OR",
        "subsets":
        [
            {
                "name": "mysphere"
            },
            {
                "name": "mybox",
                "geometry": "box",
                "x_bounds": [-1.0, 1.0],
                "y_bounds": [-1.0, 1.0],
                "z_bounds": [-1.0, 1.0]
            },
 
        ],
    },
    {
        "name": "inverted_combination",
        "geometry": "combine",
        "operator": "OR",
        "invert": true,
        "subsets":
        [
            {"name": "combined_box_and_sphere"},
        ],
    },
 
]

User-defined geometrical shapes

The current version of Neko does not support user-defined shapes from the case file. That said, shapes can be defined manually into new types by extending point_zone_t and implementing the abstract criterion interface.

Using point zones

Point zones defined in the case file are stored in a point zone registry, neko_point_zone_registry. The point zone registry allows for the retrieval of any point_zone_t object when needed. Once a point_zone_t object is retrieved, it can be used for e.g. applying a source term to a localized zone, as demonstrated below:

subroutine forcing(f,t)
  class(fluid_user_source_term_t), intent(inout) :: f
  real(kind=rp), intent(in) :: t
 
  integer :: i
  class(point_zone_t), pointer :: my_point_zone
  
  my_point_zone => neko_point_zone_registry%get_point_zone("myzone")
 
  ! Assign a constant forcing to my_point_zone
  do i = 1, my_point_zone%size
     f%u(my_point_zone%mask(i), 1, 1, 1) = 2.0
  end do
 
end subroutine forcing

Table of Contents