Multi-User Methods for FEA Pre-Processing

• Main Author: Weerakoon, Prasad
• Format: Others
• Published: BYU ScholarsArchive 2012
• Subjects: multi-user collaboration; collaborative design; multi-user decomposition; multi-user architectures; collaborative architectures; CAx; multi-user FEA pre-processing; CUBIT; CUBIT Connect; workspace assignment; workspace decomposition; Mechanical Engineering
• Online Access: https://scholarsarchive.byu.edu/etd/3255; https://scholarsarchive.byu.edu/cgi/viewcontent.cgi?article=4254&context=etd

Collaboration in engineering product development leads to shorter product development times and better products. In product development, considerable time is spent preparing the CAD model or assembly for Finite Element Analysis (FEA). In general Computer-Aided Applications (CAx) such as FEA deter collaboration because they allow only a single user to check out and make changes to the model at a given time. Though most of these software applications come with some collaborative tools, they are limited to simple tasks such as screen sharing and instant messaging. This thesis discusses methods to convert a current commercial FEA pre-processing program into a multi-user program, where multiple people are allowed to work on a single FEA model simultaneously. This thesis discusses a method for creating a multi-user FEA pre-processor and a robust, stable multi-user FEA program with full functionality has been developed using CUBIT. A generalized method for creating a networking architecture for a multi-user FEA pre-processor is discussed and the chosen client-server architecture is demonstrated. Furthermore, a method for decomposing a model/assembly using geometry identification tags is discussed. A working prototype which consists of workspace management Graphical User Interfaces (GUI) is demonstrated. A method for handling time-consuming tasks in an asynchronous multi-user environment is presented using Central Processing Unit (CPU) time as a time indicator. Due to architectural limitations of CUBIT, this is not demonstrated. Moreover, a method for handling undo sequences in a multi-user environment is discussed. Since commercial FEA pre-processors do not allow mesh related actions to be undone using an undo option, this undo handling method is not demonstrated.