Re: Differential Gasket Thickness


[Follow Ups] [Post Followup] [Dodge Power Wagon Forum]


Posted by Matt Wilson [108.162.221.179] on Sunday, January 21, 2018 at 12:38:37 :

In Reply to: Re: Differential Gasket Thickness posted by Clint Dixon [162.158.75.127] on Sunday, January 21, 2018 at 08:25:38 :

For many years, my job was (and still is to some extent, though my role has changed somewhat) to assess fatigue strength and fatigue life of aircraft components. I would like to have gotten into the FEA modeling, but never really have had the opportunity. Like your company, ours has individuals dedicated to performing FE analysis, and there are others who perform stress analysis mostly by hand calculations. Then there are those like me, who perform a limited amount of hand stress calculations and very little or no FEA, but who do more specialized fatigue strength and life calculations, fatigue tests, review of new designs for adequate fatigue resistance, evaluation of field problems, etc.

The deflection and stress calculations for a bar of uniform cross section (like the axle shaft basically is) are pretty simple and straightforward. When it comes to FEA, I look at it with a very suspicious eye, as seemingly minor changes to the modeling technique can produce vast differences in the output of the model. Validation of the model, either by known reliable hand calculations, or by strain gaging a representative part, is an important part of showing that the model provides adequately accurate results. Once the model is shown to be sufficiently accurate, it can be used for many other load cases. How accurate it needs to be will, of course, depend on how critical the component is, i.e., the consequences of its failure. It's generally considered acceptable if the model mimics reality to within 10%, so unvalidated models are probably off by a lot more than that in many instances. So think about that next time someone presents the pretty pictures from a model to you. I'm not trying to dog anyone or make anyone uncomfortable, but it's good to be aware of these limitations. And this can be true even with an experienced modeler, although usually not as often, but more often than we would all like to think.

Anyway, I've taken this tangent pretty far.



Follow Ups:



Post a Followup

Name:
E-Mail:
Subject:
Message:
Optional Link
URL:
Title:
Optional Image Link
URL:


Enter the characters in the image to the left or Load New Code