Content¶
User Guide
GUWlib - Python package for Guided Ultrasonic Wave simulation with ABAQUS¶
About¶
GUWlib is a Python package that helps setting up models in the Finite Element software ABAQUS for the simulation of Guided Ultrasonic Wave (GUW) propagation and defect interaction in thin plates. Simulations can be defined by simple Python scripts, which set the geometry and loading parameters. These scripts can then be run in ABAQUS/CAE to generate the FE model.
The ABAQUS scripting interface is utilized to automate the modelling work in ABAQUS/CAE. Mesh size can be controlled through easy-interpretable parameters. The required calculations are automatically performed using the stored material data. Additionally, functions for the batch processing of models are provided, simplifying the generation, solving, and post-processing of finite element models (either local or on a computer cluster).
Features¶
Geometry:
rectangular plates
circular through-thickness holes
through-thickness cracks
Meshing:
mesh size determination based on user-defined parameters (number of elements per wavelength / in thickness direction) and dispersion data
appropriate partitioning of the geometry to ensure well-structured hexahedral meshes, even around defects and transducers
Excitation:
transducers represented by concentrated forces, exciting either symmetric, asymmetric, or both Lamb wave modes
(transducers represented by circular piezo-electric patches with input/output voltage)
transducer signals: burst, unit impulse can be applied to each transducer individually as load cases
Materials:
import of isotropic and piezoelectric material definitions through a JSON file interface
import of dispersion data, generated by DLR dispersion calculator through a TXT file interface
Batch processing:
local pipeline: building and solving the models with the specified number of cores, extract history and field output and save to NumPy binary files
cluster pipeline: wrappers for automatic upload of model scripts to the cluster via SSH, parallel solving of multiple models on multiple cores, post-processing and results download