Convert a TexGen Weave to SwiftComp Solid and Plate SGs#

Problem Description#

One microstructure can serve different macro models. The TexGen plain weave of Convert a TexGen Weave to SwiftComp is homogenized here both as a 3D solid (SD1) and as a Kirchhoff-Love plate (PL1), so it needs two SwiftComp 2.1 .sg files.

sgio.convert() could write each file, but it reads the input deck once per call. This example reads the deck once and writes both files.

Solution#

"""Read a TexGen plain weave once and write SwiftComp SGs for two macro models."""

from __future__ import annotations

import os
from pathlib import Path

import sgio


EXAMPLE_DIR = Path(__file__).resolve().parent
INPUT_FILE = EXAMPLE_DIR / "plain_weave_3d.inp"
OUTPUT_FILE_SD = EXAMPLE_DIR / "plain_weave_3d_sc21_sd.sg"
OUTPUT_FILE_PL = EXAMPLE_DIR / "plain_weave_3d_sc21_pl.sg"


def main() -> None:

    # Read the TexGen model once; the macro model is given per write below
    sg = sgio.read(
        filename=os.fspath(INPUT_FILE),
        file_format="abaqus",
        sgdim=3,
    )
    sg.analysis_config.physics = 1  # Thermoelastic analysis

    # Write one SG per macro model. sg.omega is None, so each write computes
    # omega from the mesh bounding box for its own model: the SG volume for
    # SD1, the in-plane (y1, y2) area for PL1.
    for model_type, output_file in (("SD1", OUTPUT_FILE_SD), ("PL1", OUTPUT_FILE_PL)):
        sgio.write(
            sg=sg,
            filename=os.fspath(output_file),
            file_format="sc",
            format_version="2.1",  # SwiftComp 2.1
            model_type=model_type,
        )
        print(f"Wrote SwiftComp SG ({model_type}): {output_file}")


if __name__ == "__main__":
    main()

  1. sgio.read() reads the deck with sgdim=3 and no model_type: for an Abaqus deck the macro model can be given at write time, so the SG read stays model-neutral.

  2. sg.analysis_config.physics = 1 selects a thermoelastic analysis. The deck defines the thermal expansion of both materials.

  3. sgio.write() runs once per macro model, with model_type="SD1" and model_type="PL1". Each write uses its own model type without changing sg.

omega is not set by hand. While sg.omega is None, each write computes it from the SG bounding box over the dimensions the SG shares with its macro model (see the omega table in Formats and Model Types):

Model

Shared dimensions

omega

SD1

y1, y2, y3

volume, 2 x 2 x 0.22 = 0.88

PL1

y1, y2

in-plane area, 2 x 2 = 4.0

To set omega by hand, pass omega= to sgio.write().

Run the example:

uv run python examples/convert_texgen_weave_to_sc_sd_pl/run.py

Result#

The script writes plain_weave_3d_sc21_sd.sg and plain_weave_3d_sc21_pl.sg. Running SwiftComp 2.1 on them gives the .k files:

  • SD1: the 6 x 6 effective stiffness matrix, e.g. C11 = 4.876e10 and C33 = 6.196e9.

  • PL1: the A, B, D plate stiffness, e.g. A11 = 8.923e9, B11 = 8.923e8 and D11 = 1.007e8.

Note

SwiftComp computes plate stiffness about the plane y3 = 0. The TexGen mesh spans z = -0.01 to 0.21, so that plane lies 0.1 below the mid-plane, which is why B11 = 0.1 x A11. For the stiffness about the mid-plane, shift the node coordinates in the script before the PL1 write, e.g. sg.mesh.points[:, 2] -= 0.1.

File List#