Hi, 

This past summer, I attempted to write two VUMATs for hyperelastic material simulation in Abaqus. One was for the Neo-Hookean model, and a second for one-term Ogden. We require VUMAT for dynamic head impact simulations in Abaqus Explicit. We seek to implement material models not available in Abaqus and wanted to begin by recreating available models to get the feel for VUMAT.

 After writing the VUMAT, I conducted single element testing to ensure that my script was correctly predicting stress in various deformation modes (uniaxial tension and compress, biaxial tension and compression, shear, volumetric, etc.). Both the Neo-Hookean and Ogden VUMATs predicted stresses that agreed with Abaqus' built-in models in every deformation mode. However, when these VUMATs were implemented into a model with many elements, severe issues began to appear. I've attached an image showing this, with the in-built Abaqus Neo-Hookean material model on the left, and my VUMAT Neo-Hookean model on the right. I believe this may be an issue of hourglassing, since we are using C3D8R elements. 

However, issues with the Ogden VUMAT are more severe. Even after switching to C3D8I elements, with more integration points, issues were still appearing. I've attached an image of a simulation with the Ogden VUMAT below, showing strange element deformation. Any ideas would be greatly appreciated.