CODES

The codes presented here are used in the paper titled “A Multi-Phase-Field Model for Fiber-Reinforced Composite Laminates Based on Puck Failure Theory”. The implementation extends previous multi-phase-field formulations for fiber-reinforced composites toward laminate-level failure modelling. The model combines Puck failure theory with separate phase-field variables for fiber-dominated and inter-fiber-dominated failure mechanisms, enabling the simulation of damage initiation, interaction, and propagation in composite laminates with different ply orientations.

The repository includes Abaqus input files and user-subroutine implementations for several benchmark examples, including coupon tests, open-hole tension, compact tension, and double-notch configurations. The examples are intended to support reproducibility, further development, and open scientific exchange in computational fracture mechanics of composite laminates.

Authors: Pavan Kumar Asur Vijaya Kumar and Aamir Dean

The codes presented here are used in the paper titled “A Multi-Phase-Field Model for Fiber-Reinforced Composite Laminates Based on Puck Failure Theory”. The implementation extends previous multi-phase-field formulations for fiber-reinforced composites toward laminate-level failure modelling. The model combines Puck failure theory with separate phase-field variables for fiber-dominated and inter-fiber-dominated failure mechanisms, enabling the simulation of damage initiation, interaction, and propagation in composite laminates with different ply orientations.

The repository includes Abaqus input files and user-subroutine implementations for several benchmark examples, including coupon tests, open-hole tension, compact tension, and double-notch configurations. The examples are intended to support reproducibility, further development, and open scientific exchange in computational fracture mechanics of composite laminates.

Authors: Pavan Kumar Asur Vijaya Kumar and Aamir Dean

The codes presented here are used for the paper titled “A Phase-Field Fracture Model for Fatigue using Locking-Free Solid Shell Finite Elements: Analysis for Homogeneous and Layered Composites”. These codes are applicable with or without fatigue considerations and are written as UEL and UMAT subroutines for use in Abaqus. We have tested the files for compatibility with Abaqus 2020 and Abaqus 2022.

When using the code, remember to update the variable allelem with the number of UEL elements in each layer. This variable is located in the module Kvisual. Please exercise caution when defining the fatigue model from the input file.

Authors: Pavan Kumar Asur Vijaya Kumar, Aamir Dean, Jose Reinoso, Heinz Pettermann, and Marco Paggi 

This repository provides a Python/TensorFlow implementation of a multiphysics Deep Energy Method for fourth-order phase-field fracture with piezoresistive self-sensing. The code is developed for simulating fracture in smart and conductive composite materials, where damage evolution can be monitored through changes in the electrical resistance.

The framework combines phase-field fracture mechanics, deep energy minimization, and an electrical sensing subproblem. The mechanical fracture problem is solved using a neural-network-based Deep Energy Method, while the electrical response is evaluated through a conductivity model that depends on strain and crack-induced damage. This enables the simulation of crack evolution together with its resistance signature.

The implementation includes adaptive refinement near damage zones, exact imposition of essential boundary conditions through distance functions, and hybrid ADAM/L-BFGS optimization. The code also provides tools for visualizing the damage field, deformed configuration, electric potential, current-density distribution, resistance history, and energy evolution.

Authors: Aamir Dean and Betim Bahtiri

A deep learning workflow for detecting and segmenting crack propagation in electroactive materials under coupled mechanical–electrical loading. The approach combines high-fidelity electromechanical phase-field FEM simulations with a computer-vision segmentation model based on a U-Net decoder and ResNet backbones. Grayscale field outputs from the simulations are used as learning inputs, primarily the phase-field damage variable and the electrical potential field.

The segmentation task labels material regions at the pixel level, for example background, crack, and hole for phase-field maps, and left, right, and hole for potential maps. A key finding is that electrical potential patterns provide a strong proxy for fracture, often yielding more accurate and more robust crack segmentation than using phase-field damage maps alone.

Authors: Aamir Dean, Jaykumar Mavani, Betim Bahtiri, Behrouz Arash, Raimund Rolfes

This document provides a brief introduction to the numerical computation of the material tangent in implicit finite element analysis, with a focus on implementation within the Abaqus UMAT subroutine. It explains the importance of the material tangent in finite element simulations and discusses the challenges associated with deriving it analytically.

The document introduces finite difference methods, particularly the forward-difference first-order scheme, for approximating the Jacobian material tangent. Practical aspects such as numerical accuracy, perturbation size, and floating-point effects are discussed. A step-by-step procedure and a Fortran UMAT code snippet are also provided, making the document suitable as a beginner-friendly guide for implementing numerical material tangents in Abaqus.

Authors: Aamir Dean