In summer 2019, I was given such a great opportunity at Quartus Engineering in Herndon as an engineering analysis intern. I learnt a lot in finite element analysis (FEA) and its well-development in the industry. For a long time, I thought that FEA only serves as a comparative method. However, FEA along with hand calculation validation has been accepted and widely used in the engineering world especially by FDA and FAA. Throughout my internship, I got to familiarize myself with several programs such as Abaqus, Hypermesh, FEMAP API and NX Nastran.
Due to the company confidential information, the images are limited. My main job at Quartus is to perform finite element analysis of aerospace and mechanical parts using varieties of software.
NOTABLE PROJECTS:
I started my summer by performing a non-linear static analysis comparison between the NX Nastran PCOMP card and actual built-in inner honeycomb geometry. I then investigated the accuracy of Nastran PCOMP card and the function its used in the code.
Later on the summer, I got to work on an actual aircraft project. I used Femap and Hypermesh to prepare detailed finite element model (FEM) of external aircraft-mounted satcom antenna system for static and dynamic analysis. It is a challenging and repetitive task since I need to prepare CAE model for a fuselage mount hardware with thousands of rivets holes.
Part of modeling preparation is connecting the adjacent holes with rigid elements (RBEs 2 or 3) and spring elements (CBUSH, CBEAM, CBAR). It is a super repetitive and time-consuming task.
Therefore, I wrote custom code using Femap API to automatically detect fastener holes in a FEM and create connections using rigid and spring elements. The code also allows users to choose which types of connections based on their preference.
It is one of my most proud projects over the summer as the code significantly decreases modeling time, particularly for fuselage-mounted hardware which may have thousands of rivets. A day job with finding and connecting holes could be shortened to only an hour.
----------------------------------------------------------------------------------------------------------------------------------------------------
Moving from huge aircraft parts to smaller parts. I am also working on a project for a customer in a biomedical device world focusing on joint healing devices. The device is still in development.
I used Abaqus to simulate multi-step forming process of a NiTi biomedical device. The analysis utilized superelastic material models, contact, and annealing steps, to mimic the customer’s actual forming process and predict forming strains. Additional analyses were performed to demonstrate proposed physical testing that can be used in DOE to evaluate potential design changes.
Throughout the projects, I learnt a lot about Abaqus especially the complexity in forming and contact analysis and constantly dealing with big models with a lot of nodes and elements. At the end of the summer, I am comfortable using Abaqus on both pre and post process. I am also able to work with and change input or data files, perform import analysis and familiarize with superelasticity material behaviour.
When I am performing the analysis for the project, I realized Abaqus instability in the plastic region definition of Superelastic material. The problem does not fix even with changing element types, stabilization, damping, friction, etc.... The experiment is a displacement control experiment. Abaqus should not have problem of following the plastic curve. A presentation was prepared and sent to Abaqus to inform them about their problem. After a few tries and email exchange, Abaqus gladly accepted that there is a bug in their stiffness matrix and promised to release the new version with the bugs fixed.