Open source implementation of a user material subroutine for composite materials

Abstract

The modeling and simulation of composite materials requires the incorporation of anisotropic material models, as well as appropriate failure criteria or damage models. This was normally done by developing in-house codes or writing additional programs to accompany existing finite element (FE) software. However, in recent years the ability to model and simulate the complex composite behaviour in commercial FE software packages has matured. That said, even commercial FE software has a limited range of failure criteria, damage models and degradation laws to choose from. Using open source FE packages, such as Calculix, to analyse complex behaviour of composite materials requires considerably more user interventions. One of the only options is to define multiple materials to build the composite or use a composite shell element, but the software provides no method to initiate a progressive failure and damage analysis which is required to accurately simulate composite material behaviour. From a research perspective, new contributions in this field are made by either developing new progressive damage models (with various modes of failure and damage) for composite materials or by implementing existing models through a user subroutine in a commercial or open source environment. A user subroutine allows for the customisation of the FE software by easily adding new material models, behaviour laws, boundary conditions, etc. without having to fully understand the complex structure and flow of the underlying FE code. The purpose of this paper is to focus on the open source implementation of a composite material model by making use of the user subroutine to add material models (UMAT) not already available. The process for implementing a UMAT in the open source software package Calculix is first established as an example for other users. The first step is therefore to develop a UMAT for a simple material model which is already available in the FE package. The implementation process is then verified and validated by comparing the UMAT to the built-in material definition. Next, a composite material model is chosen from literature and implemented via a UMAT. The composite material model is verified through a simple single element test and validated with experimental test data. In short, this paper provides a method for implementing a composite material model, through the use of a UMAT, to model the behaviour of composite structures in an open source FE software package.