Adams Advanced Package > FE Part > Appendix > FE Part Compared to Other Methods for Geometric Nonlinearity

FE Part Compared to Other Methods for Geometric Nonlinearity

The FE Part modeling object is accurate for beam-like structures undergoing very large deformation; that is, geometric nonlinearity. Therefore one will want to compare this approach with other common approaches for modeling geometrically nonlinear flexibility of beam-like structures in Adams. The other principal approaches are:

■Discrete Flexible Link: a series of rigid bodies connected by Adams Solver BEAM, FIELD or BUSHING force objects

■Multi MNF: a series of modal neutral file (MNF) based Adams Flex linear flexible bodies connected by fixed joint type constraints or bushings

There are advantages (pros) and disadvantages (cons) of each approach:

■FE Part

■Pros

♦Geometrically nonlinear and distributed mass formulation does not require subdivision in lumped masses or linear flexible components

♦Modification and parameterization is often easier than the other two approaches

♦Modeling a distributed load via the FE Load is far less time consuming than using discrete force vectors or MFORCES

♦Support for stress and strain recovery in APostProcessor (X-Y plots)

♦Reduced noise in nonlinear contact where a geometry "wraps" around another since the geometry is not discretized

♦No "seams" in the stress/strain results due to discretization

♦Refining or coarsening the model (number of nodes in this case) is often easier than in the other two approaches

♦2D formulation option for faster analysis on planar problems

■Cons

♦No support for AControls

♦No support for Adams2Nastran export

♦No support for stress and strain recovery in APostProcessor (animated color contour plots)

■Discrete Flexible Link

■Pros

♦Can be defined in a preloaded/pre-stressed condition (and support model save/reload)

♦Support for ALinear

♦Full support for AControls

♦Supports use of 2D parts for faster analysis on planar problems

■Cons

♦Time consuming to modify and parameterize (Adams command language skill required in many instances)

♦Time consuming to apply distributed loads to it

♦No direct support for stress/strain recovery (need to write your own requests)

♦Refining or coarsening the discretization is somewhat time consuming without Adams command language skill

♦Mass properties approximated by discretization

■Multi MNF

■Pros

♦Can be defined in a preloaded/pre-stressed condition (and support model save/reload)

♦Support for ALinear

♦Full support for AControls

♦Modal stress/strain recovery (X-Y plots and animated color contour plots)

♦Distributed mass properties

■Cons

♦Model setup is time consuming (MNF generation for discretization of what is usually a single piece of geometry)

♦Higher skill required for model setup (definition of interaction between discrete flexible bodies is not straightforward to do correctly)

♦Time consuming to modify and parameterize aspects of it requiring MNF re-generation (shape, material and so on.)

♦Sometimes time consuming to work with applied distributed loads (MFORCE workflow)

♦Refining or coarsening the discretization is time consuming

♦No 2D option

Another consideration is model solve time. On this consideration it is difficult to compare the three approaches generally speaking. The comparisons are expected to be model dependent. One would want to compare the coarsest model which achieves sufficiently accurate results from the various approaches. Model coarseness/fineness is defined by number of nodes (FE Part), parts (discrete flexible link) or flexible bodies (multi-MNF).