PERCEPT - the “ProcEss stRuCture propErty Prediction Tool” - is a rapid process-structure-property software tool for metallic materials, developed by I-Form. Microstructural images such as Electron Beam Back Scatter Diffraction (EBSD) containing grain morphology and crystallographic orientation information are the main input data applied to the tool, in addition to the material type. The tool instantly (within seconds) predicts the mechanical performance of the material in terms of a tensile stress-strain curve, giving stiffness and yield strength. The results are rapidly calculated with the support of an AI methodology called Deep Neural Networks (DNN), trained with a large battery of input-output samples from Crystal Plasticity Finite Element (CPFE) modelling. PERCEPT successfully recognizes phase regions and the associated unique crystallographic orientation variations. It also captures differences in macroscopic stress response due to the varying microstructure. To date, it has been demonstrated on multiphase stainless steels manufactured by powder bed fusion but the framework applies to any material or process. PERCEPT brings the accuracy of computationally expensive CPFE modelling together with the speed of image based AI, thus reducing the time and cost for industry associated with in-house mechanical quality control testing.
A run through of the user interface and tool in operation is at the video link below:
The PERCEPT tool continues to be developed and we are interested in hearing from potential evaluators and collaborators. Please contact the academic lead on this project Dr Noel Harrison (noel.harrison@universityofgalway.ie) or info@i-form.ie to trial the tool and get more details.