Why a Design Engineer is supreme to an FEA Engineer?

A Design Engineer and an FEA Engineer normally should work side by side.  This doesn't mean they must accompany each other during coffee breaks and lunch breaks (its better if they do it) but knowing each other's capability and job profile.

An FEA Engineer’s task is to model the actual physics into Finite Element Model.  This is his expertise field and hence don't expect him to be able to predict too much on design.  FEA Engineers normally tend to think a lot regarding the correct element types, assumptions, material models applicable for the given problem etc.  However, whenever they make assumptions, they must consult with the Design Engineers in order to make sure that the FEA Analysis approximations are reasonable to represent the physics problem. To validate the FEA Analysis results by simplified hand calculation is the ownership of a good FEA Engineer.  However, in order to simplify the model, he must take the opinion of the Design Engineer since he is the ultimate authority when it comes to understanding the Design.

A Design Engineer on the contrary is much more focused about the performance of his design and knows what to expect whenever a particular loading is applied.

It is very much of a necessity that an FEA person keeps on asking questions regarding the Design.  Normally people shift from Design to FEA.  The general perception is FEA is one step up the ladder than Design.  But the truth is Design is always supreme. 

The process of Engineering Product Development has changed from conventional methods of historical recording to Simulation driven product development with the use of high end FEA Analysis software.   As the technology grows, the focus normally shifts from the conventional approach to emerging ones.  But the fundamentals remains same.  Be it an FEA Engineer or a Design Engineer, both of them need to apply the concept of Structural Mechanics and keep thinking.


Vibration, Music and FEA

Since yesterday night I have been thinking about music.  Then it turned out to be vibrations and later on I related it with FEA.
Each scale has seven "Sur".  Sa, Re, Ga, Ma and Pa.  I tried it on my synthesizer and the thing I noticed is, these "Sur" are relative to each other.  Means, if I play a Sargam from a "Sa", same "Sa" can be heard as "Re" if I play from one down. 

These are basically frequencies.  Because of these only the "Harmonium" is called a "Harmonium" since it uses the Harmonics of the scale.  The part I'm still amazed is that the scales are divided equally into 8 parts (being 7 sur), yet there are other in between frequencies which we know as "Komal" and "Teevra".

Up to this is fine.  But then I thought about guitar FEA Analysis!  It would be really interesting to analyze the natural frequencies for a guitar wire!  I found out that some projects are already done on that.

From here on my perspective on vibration has enlightened.  People think of vibrations as always be noisy and unwanted conditions unknowingly the fact that the same vibrations help us to relax when we listen music.  It's the frequencies which make you laugh when somebody laughs in typical manner. 

And it's the vibrations which annoys the machine operator who complains to his superior about it and such superiors can contact us to get the solutions! 


When Desserve Engineering has done work on fiber glass leaf spring and AUDI is planning to implement it for coiled springs....

I was going through this article at a website (Link: http://machinedesign.com/blog/more-automakers-going-fiberglass-springs#comment-77831). This is all about AUDI planning to implement fiberglass coiled springs. I connected it with one of the projects done by us. We had a research based company working for fiberglass leaf spring. The benefits that it offers are in line (and better in some aspects) than the conventional materials. It reduces weight and Fatigue parameter performance is better. It really pleases you in terms of work quality satisfaction since major car brand is implementing something similar you've supported.


How to Model Reinforced Concrete in FEA

FEA Analysis applications has started from structural mechanics and hence civil engineering or Applied Mechanics finds the closer use of FEA.  However, modelingconcrete materials has not been easy due to the complexity involved in the material properties as well as the use of reinforcements that normally they're used with.  FEA methodology for modeling a reinforced concrete structure has been more or less similar for most of the software.

The most appropriate method to model the reinforced concrete structure is to model the concrete as solid elements and the reinforcements as link elements.  Off course you need to tie the elements with each other at proper nodes.  The good thing about this approach is: It reduces the computational resources and you can still have comparable results with those in actual physics.  Since the reinforcement members are there to take the load (or transfer the load within the structure) and usually we are not interested in simulating their failure, this methodology is beneficial to get the desired outputs in terms of simulating reinforced concrete structures.

However, one of the drawbacks of using this method is: we cannot initiate a crack by describing failure criteria for the link members. There are indirect methods of element death options in certain codes, however it is recommended that whenever we want to check the failure and apply a failure criteria for the reinforcements, we should model them with proper solid elements.


How to Make Valid FEA Assumptions

Finite Element Analysis is being popular tool to evaluate the product performance without a prototype having built.  Though it is being used as a qualitative analytical tool for majority of applications, there are products wherein it is required that a model calibration is done to ensure the quantitative parameters predicted matches with the actual results.

While it is important to include the geometrical details when conducting an FEA Analysis, it is more important to have knowledge that which details should be excluded.  This is required since FEA models are based upon assumptions and the role of FEA engineer is to make valid assumptions. We will be discussing about some of the examples for simplifying the problem.

(1)   Use of Linearity vs Non linearity:

If the problem is more or less a linear, make sure to conduct linear elastic analysis first.  In real physics, almost all problems are nonlinear, however linearity is an approximation method.  There are different types of non-linearity such as geometric(shape based), material non linearity and contact non linearity.  If the geometry is not much slender, you can assume it to be geometrically linear. However, for slender structures having loading at the end of the span, one should chose nonlinear geometry option. Common examples are bending of the fishing rods, deformation of off shore structures (Very long span), deformation of bridges etc. 

For general application materials such as metals (Fe based), one should go with linear material option.  However, if the intention is to carry out an explicit (crash) analysis, one must use nonlinear materials.  For gaskets, rubbers and seals, one has to use nonlinear materials.

Contact non linearity is generally applicable to the frictional, sliding or no separation kind of contacts in an assembly level analysis.  One should ensure that if the parts are assembled with locked degrees of freedom as in case of bolted or glued, MPC based contact algorithm should be used.

(2) Use of beams and shells:

Use of elements such as beams and shells can reduce the model size by significant amount while one can still have the intended result sets out of the FEA analysis.  Beams are defined as a line with a cross section and material properties assigned to them.  Examples of using beams while modeling structures is: FEA models for building pillars and beams, FEA models for machine structures etc.  One variant of beams is pipe element which is used in carrying out strain analysis for piping systems.  Usually these systems are very long (in the ranges of km) and hence usage of line based elements such as pipes and beams helps to solve such problems very quickly.  Shells are popularly used in automotive and aerospace industry to model the panels however nowadays they're being used in the pressure vessel industry as well.  One can easily change the thickness parameters for shells and assess the designs quickly.  The biggest advantage of using shells is one can model composite lay up using shell elements.  This is really helpful with increasing popularity of composites in various industries.


Goto Top