Discover Better Designs, Faster – with HEEDS MDO

Don’t Just Simulate, Innovate!

HEEDS MDO

Streamlining the Design Exploration Process

Featured Customer Case Study: Improving Engine Mount Performance

• This Case study aims to Illustrate through a step-by-step process how an industry-leading company applied this technology to solve a real problem

So let’s talk a bit more about the required steps to streamline the design exploration process. To do that, we’ll use a real-world example from an automotive supplier. A major automotive company tasked this supplier with providing a new, low-cost engine mount that would better match new performance targets than the existing mount.
 

 
As shown here, the OEM provided target curves for the “force vs. deflection” curve for each side of the mount.  The solid line represents the desired target stiffness, while the dotted line represents the simulated stiffness of the current mount.  The supplier decided to employ HEEDS to quickly identify a design that would fit the requirements of the OEM and win the business.
 

A challenging and realistic Automotive design problem 

To simplify the typical process automation set-up used for the building of virtual prototypes of the engine mount, they used HEEDS to automate the modeling and simulation workflow.  They first employed NX CAD to generate the 11-variable, parametric, 3D geometry for the engine mount.   Abaqus was used to automatically mesh the engine mount, perform the 3D structural (FEA) simulation, and analyze the results. HEEDS was used to monitor the simulation responses and intelligently modify the mount geometry variables in an effort to help engineers discover mount designs which would exhibit force vs. displacement characteristics closer to the target.

Distributed Execution Process

To accelerate the testing of virtual prototypes of their engine mounts, they assigned each task of the workflow to the most appropriate and available compute resources.  So, the NX CAD geometry modifications were performed on a single-core, Windows desktop computer while the ABAQUS FEA simulations were assigned to be executed four-at-a-time on a 24-node Linux cluster with 6 cores per node.  This allowed the supplier to quickly evaluate hundreds of design variations.

Efficient Search for High Performing Designs

Once the process automation and distributed execution set-up had been completed, the engineers initiated a study for HEEDS to explore 250 design variations in an effort to identify designs that would meet the intended target.
 
The engine mount shape is being varied automatically and a structural simulation is being performed.  HEEDS automatically calculate the root-mean-square (RMS) error between the simulated results and the target force vs. deflection curves.  Each dot in the chart represents the RMS error of one specific design.  The blue line connects the best designs found throughout the study.  In just days, the supplier identified multiple designs that satisfied the OEM targets and they won the business.

Illustrate Insight & discovery Phase of Design Exploration

The OEM also asked the supplier to provide information on the sensitivity of the design to each design variable. In this case, HEEDS calculated the sensitivity of both the starting baseline design and the improved design to each variable. It is interesting to note that some of the variables, such as thickness RD1, have a large effect on the performance of the improved design, but were not very influential around the baseline. In traditional design exploration, these types of variables may have been screened out which would have prevented the supplier from finding such a good design.
 
HEEDS also illustrates the level of interaction of variables on performance.

The supplier had previously tried to solve issues like this be time-consuming, tedious, manual iteration. HEEDS simplified and accelerated the process, improving the success rate of bidding on new programs. It also established the supplier as an innovation partner to the OEM, solidifying downstream business.