Computer Aided Engineering tools have come a long way since their inception. The CAE tools have become so sophisticated and comprehensive that sometimes it’s difficult to remember how they started from a simple equation. On the other hand, we have become so reliant on using them that it is difficult to imagine engineering a complex product such as an automobile without using any CAE tools.
There is a lot of choices in the CAE world … Frequency domain, time domain, co-simulation, etc. are all various domains for performing analysis. Model build detail has also progressed tremendously. Various mesh size and types, analysis precision, diverse range of surface contact definitions such as frictional surface, bolted joints, clamped components, etc. all provide superior ways of building the perfect simulation model. And finally, various solvers come to perform the analysis and provide post-analysis tools for critiquing simulation results. With all of that, why do we still see NVH problems on mass production vehicles??
In my opinion a great deal of this comes to HOW we use these tools. A successful simulation requires:
A thorough Design of Experiment
An accurate set of material input data, truly representing the material properties of the actual brake system components
Representative geometry - CAD model
The correct choice of model build attributes such as mesh element type, contact definitions, etc,
The correct choice of analysis domain and solver
An accurate independent set of data for correlation of the model before results can be critiqued
There is A LOT of potential failure modes in this, but what is the most common failure mode? In my experience the biggest challenge is having a representative model. Building a CAE model involves making a lot of assumptions and that’s where a lot of models go wrong and a wrong model results in sign off of a vehicle which will perform differently in the field, which is how the model should have been set up to replicate.
In my early days of learning about brake NVH / CAE, my mentor told me you always should assume this sophisticated piece of software is trying to make wrong assumptions. It’s the responsibility of the user to use the toolset correctly to produce quality results.
Comments