CornerPointGrid

class petres.grids.CornerPointGrid

Bases: object

3D corner-point grid with explicit corner coordinates.

Parameters:

• pillars (PillarGrid) – Lateral pillar geometry defining i–j topology.

• zcorn (ndarray) – Corner depths shaped (2*nk, 2*nj, 2*ni) in Eclipse order.

• active (ndarray or None, optional) – Boolean mask (nk, nj, ni); None marks all cells active.

• zone_index (ndarray or None, optional) – Zone id per cell, same shape as the grid. 0 denotes gap.

• zone_names (dict[int, str], optional) – Mapping of zone id to zone name.

Notes

Properties are managed via GridProperties and are not stored in zcorn. Use from_zones() to build grids from stratigraphic zones.

pillars: PillarGrid

zcorn: ndarray

active: ndarray | None = None

zone_index: ndarray | None = None

zone_names: dict[int, str]

__post_init__()

Validate grid arrays and initialize zone lookup metadata.

Raises:

• ValueError – If zcorn, active, or attached property shapes do not match the grid dimensions, or when a property belongs to a different grid.

• TypeError – If an attached entry in _properties is not a GridProperty instance.

Notes

When active is None, an all-active broadcast view is created for memory efficiency.

summary()

Create a human-readable summary of grid dimensions and metadata.

Returns:

Multi-line summary including shape, active/inactive counts, registered property names, and zone names.

Return type:

str

Examples

>>> text = grid.summary()
>>> "Grid Summary" in text
True

set_zones(zone_index, zone_names)

Assign zone membership arrays and zone-name metadata.

Parameters:

• zone_index (np.ndarray) –

  Integer array of shape (nk, nj, ni) assigning each cell to a zone. Values use this convention:

  • 0 means gap, undefined, or unassigned.

  • > 0 means a valid zone id.

• zone_names (dict[int, str], optional) – Mapping from zone id to zone name. If not provided, names will be auto-generated as “Zone {id}”.

Raises:

• ValueError – If shapes mismatch, invalid ids are present, or names are inconsistent.

• TypeError – If zone_index cannot be converted to an integer ndarray.

property properties: GridProperties

Return the property manager bound to this grid.

Returns:

Dictionary-like facade used to access and validate cell properties.

Return type:

GridProperties

property ni: int

Get the number of cells in the i direction.

Returns:

Cell count along the i axis.

Return type:

int

property nj: int

Get the number of cells in the j direction.

Returns:

Cell count along the j axis.

Return type:

int

property nk: int

Get the number of layers in the k direction.

Returns:

Layer count inferred from zcorn.

Return type:

int

property shape: tuple[int, int, int]

Get grid dimensions ordered as (nk, nj, ni).

Returns:

Layer, row, and column cell counts.

Return type:

tuple[int, int, int]

property n_cells: int

Get the total number of cells in the grid.

Returns:

Product of nk * nj * ni.

Return type:

int

property n_active: int

Get the number of active cells.

Returns:

Count of True values in active.

Return type:

int

property n_inactive: int

Get the number of inactive cells.

Returns:

Difference between total and active cell counts.

Return type:

int

property cell_centers: ndarray

Cell centers computed as the mean of 8 corners.

Returns:

Coordinates shaped (nk, nj, ni, 3).

Return type:

ndarray

property cell_volumes: ndarray

Cell volumes computed via hexahedral decomposition.

Returns:

Volumes shaped (nk, nj, ni).

Return type:

ndarray

classmethod from_grdecl(path, *, use_actnum=True, use_metadata=True, properties=None)

Load a grid from a GRDECL file.

Parameters:

• path (str or Path) – Path to GRDECL file containing COORD/ZCORN (and optional ACTNUM).

• use_actnum (bool, default True) – When True, read ACTNUM to set active cells; otherwise all active.

• use_metadata (bool, default True) – When True, read metadata keywords from the file.

• properties (Sequence[str] | None, default None) – Property names to import. If None, all available properties are imported.

Returns:

Grid populated from GRDECL arrays.

Return type:

CornerPointGrid

to_grdecl(path, *, properties=None, include_actnum=True)

Write grid geometry and selected properties to a GRDECL file.

Parameters:

• path (str or Path) – Output file path.

• properties (Sequence[str] or None, optional) – Property names to export. When None, all attached properties are written.

• include_actnum (bool, default True) – Whether to export ACTNUM from the grid active mask.

Raises:

• TypeError – If properties is not a valid sequence of names.

• MissingEclipseKeywordError – If an included property has no Eclipse keyword configured.

show(show_inactive=False, scalars=None, color='#BFFF00', cmap='petres', title=None, z_scale=1.0, wells=None, **kwargs)

Render the grid in 3D PyVista viewer.

Parameters:

• show_inactive (bool, default False) – Whether to display inactive cells.

• scalars (str or None, optional) – Property name to color by; if None uses solid color.

• color (Any, default DEFAULT_COLOR) – Solid color when scalars is not provided. Defaults to the project default color in petres.config.colors (DEFAULT_COLOR).

• cmap (str or None, default DEFAULT_CMAP) – Colormap applied when scalars is provided.

• title (str or None, optional) – Figure title.

• z_scale (float, default 1.0) – Scale factor for the z-axis.

• wells (VerticalWell or Sequence[VerticalWell] or None, optional) – Well(s) to plot on top of the grid. Can be a single VerticalWell or a sequence of them. If None, no wells are plotted.

• **kwargs (Any) – Forwarded to viewer add_grid.

classmethod from_rectilinear(x, y, z)

Create a corner-point grid from rectilinear coordinate vertices.

Parameters:

• x (np.ndarray) – Monotonic x-vertex coordinates with shape (ni + 1,).

• y (np.ndarray) – Monotonic y-vertex coordinates with shape (nj + 1,).

• z (np.ndarray) – Layer-interface depths with shape (nk + 1,).

Returns:

Grid with vertical pillars and broadcasted layer surfaces.

Return type:

CornerPointGrid

classmethod from_regular(*, xlim=None, ylim=None, zlim=None, ni=None, nj=None, nk=None, dx=None, dy=None, dz=None)

Build a rectilinear corner-point grid from limits or spacings.

Parameters:

• xlim (tuple[float, float] or None, optional) – Minimum and maximum x bounds.

• ylim (tuple[float, float] or None, optional) – Minimum and maximum y bounds.

• zlim (tuple[float, float] or None, optional) – Minimum and maximum z bounds.

• ni (int or None, optional) – Number of cells in i direction.

• nj (int or None, optional) – Number of cells in j direction.

• nk (int or None, optional) – Number of cells in k direction.

• dx (float or None, optional) – Cell size in i direction.

• dy (float or None, optional) – Cell size in j direction.

• dz (float or None, optional) – Cell size in k direction.

Returns:

Corner-point grid resolved from the provided geometric constraints.

Return type:

CornerPointGrid

classmethod from_zones(*, pillars, zones)

Create CornerPointGrid from zones with gap detection.

Parameters:

• pillars (PillarGrid) – Lateral pillar geometry

• zones (Sequence[Zone]) – Zones in stratigraphic order (top to bottom)

Returns:

Grid with gap-filling cells marked as inactive (ACTNUM=0)

Return type:

CornerPointGrid

apply_boundary(boundary)

Mask grid activity using a boundary polygon.

Cells whose centers fall outside the polygon become inactive (ACTNUM=0).

Parameters:

boundary (BoundaryPolygon) – XY boundary polygon used to clip the grid.

Returns:

Updated grid with activity masked by the boundary.

Return type:

CornerPointGrid

Notes

Updates active in place; cached property masks may become stale.

well_indices(well)

Locate the surface grid column indices intersected by a vertical well.

Parameters:

well (VerticalWell or tuple[float, float]) – Well object providing x and y map coordinates, or a tuple of coordinates.

Returns:

Zero-based (i, j) indices of the top-surface column containing the well head location. Returns None when the well lies outside the grid footprint.

Return type:

tuple[int, int] or None

Notes

This method performs a geometric XY containment lookup on the grid top surface only. It does not evaluate completion intervals or active-cell status.

__repr__()

Return a compact debug representation of the grid instance.

Returns:

Representation including shape, active-cell count, and properties.

Return type:

str

__init__(pillars, zcorn, active=None, zone_index=None, zone_names=<factory>, _properties=<factory>, _eclipse_metadata=None)