With the integration that is available today between CAD and FEA, it is not uncommon to take model setup for granted, and put out fires at the end if there are conflicting results. For a seasoned FEA user, this approach works effectively enough. However, a little planning ahead can go a long way in ensuring correct results.
Analysis is a lot like tightrope walking. One misstep in the setup can cause results to look disastrous. However, if setup correctly, the results can be superlative and extremely invaluable in terms of time, money and resources. For example, if a cantilever is fixed on one end, and a load is applied on the other, the results will be synchronous with real-life testing. However, if the fixed restraint was not applied, and was substituted with a hinge type of restraint, then the cantilever would not behave as expected. One degree of freedom can make a world of difference if it is not accounted for in the model.
This brings about an interesting discussion on the rules to follow to minimize the chances of errors during analysis setup.
UNDERSTANDING THE PROBLEM
The first step in performing an analysis is to take sufficient time to assess the contribution of analysis to the entire design objective. An optimal design can be a by-product of analysis, or analysis can be treated as just a final spell-check on a functional design. If the scope of the analysis is outlined at inception based upon its role in the design phase, the influence of assumptions/idealization of geometry and loads can be charted clearly.
It is also important to make sure that sufficient details are gathered before setting up the problem. The type of geometry, the nature of the problem, the type of material(s), the loading conditions, the pattern of loading, the required outputs, and the time scale for completion of the analysis are examples of some important data to assimilate before conception of analysis.
While it is not a requirement, setting the platform for analysis by glancing through the theory governing the nature of the problem will help the user make appropriate judgments on validity issues. Performing simple hand calculations by using computer simulations/applets etc. can also be very useful in laying the groundwork for analysis. Successful analysis is often accompanied by spreadsheets that carry detailed research tracking trends in results with variations in pre-processing.
(…to be continued)