What if every engineer, no matter their duties or background, could have easy access to powerful CAE prediction capabilities? What if they could use it directly in developing—or selling—their automotive components? “For example, a GKN Driveline engineer at a customer site in Asia could launch an FEA and pick up the results the next morning,” said Jeff Wohlschlegel, Principal Engineering Specialist for GKN Driveline. GKN Driveline has done just that with its Virtual Product e-Realization (ViPeR) applications, and in particular the Virtual ToolKit (VTK).
The GKN Driveline engineers that use this VTK do not have years of specialist CAE training in complex mathematics. Rather they are design and release, application, or sales engineers with broader responsibilities. The tools they access range from simple spreadsheet calculations to state-of-the-art computer-aided analysis solvers, including strength, fatigue, function, and NVH applications. Some are empirical tools built on GKN collected data. Others are advanced numerical solvers from third-party vendors. All tools are accessed through standard simulation setups using standardized input templates. Results are provided in standard output formats as well.
Constraint is freedom
To make this work, users of the VTK are constrained in selecting standard parts. These could be sideshafts, propshafts, or constant velocity joints among others GKN supplies. It starts with a spreadsheet-like interface, with many of the input parameters prepopulated to provide a starting point, according to Wohlschlegel. These select variables are then editable by the user. Post-processing provides output in standardized formats similar to physical test results familiar to the end users. “It puts the results in their context,” said Wohlschlegel.
He describes a typical, but complex, simulation as requiring a CAD file in STEP format as the starting point. Loads are specified, either directly or from running a kinetic model to predict the loads. Preprocessing of an FEA mesh is performed using, once again, preconfigured templates. These elements feed a static FEA solver, followed by post-processing that is combined with a fatigue prediction. A key element includes results presented in familiar formats. For example, fatigue life is presented in a familiar lin-log format of torque compared to cycles.
Another key factor he stresses as a key enabler is the use of centralized computer servers accessed through thin-client connections. That engineer in Asia is in effect running analysis on powerful computers in the home office, not on his desktop. “Also, the VTK portal is accessible 24/7 over the intranet within the Driveline Internet,” said Wohlschlegel.
This provides freedom to the CAE specialists as well, he claims. “This frees up our simulation specialists for more complex R&D and to develop and build the next level of models for the future,” he said. In the past, these specialist engineers were tasked with performing routine CAE analysis to support product development and design engineering.
Foundations in enterprise web architecture
The journey for GKN Driveline’s progression to CAE democracy started in 2002 with an adoption of enterprise-wide PDM. Since then, capabilities have grown to include Virtual Program Management, Enterprise Change Management, and finally to company-wide Enterprise PLM workflows and democratized CAE. The VTK is one of three application portals into this company-wide PLM system, the other two are Product Selection and Product Testing. All three mutually support each other, stresses Glenn Valine, Director, Engineering ISIT for GKN. Physical testing is used to correlate simulations while stored product CAD is the starting point for simulations. Most data, test or simulated, is stored and accessible, reducing the cost of having expensive engineering time spent simply locating data.
“What made ViPeR necessary was the fact that there were engineers in 11 locations around the world using few standardized tools,” said Valine. Simulation was slow, sometimes not well correlated, and produced inconsistent results. Also, since most simulation was being done by specialists, there were capacity bottlenecks. “Routine work was using valuable resources needed for R&D.”
Application knowledge still needed
There are some cautions. While they stress this is for non-specialists, it is not for non-engineers. Basic understanding of the engineering problem and result interpretation is still required. This is a basic sentiment that Keith Meintjes, Practice Head, Simulation & Analysis for CIMdata, agrees with in commenting on the general trend toward making CAE easy to use.
“The goal of democratization is to make the power of simulation available to more people, to make it pervasive,” he told SAE Magazines in a separate discussion. “The issue is to reduce or hide the complexity of the simulation tools. It is not to reduce the need for understanding of the engineering problem.”