Remote vs. Distributed Mass in SolidWorks Simulation

Engine_FrameOne of the more important and yet challenging aspects to using Finite Element Analysis is constructing a usable model.  This model, whether found natively in a CAD package, imported from a CAD package, or constructed within a dedicated FE package needs to be both efficient and provide accurate results.

When certain models contain a great amount of detail we need to remove the extraneous details while preserving the integrity of the model.  However some "details" represent entire sub-systems.  For example, to the left, is a very simple welded frame. Atop the frame, though, is a model of a small Honda engine. (Thank You 3DContentCentral!!)  I'm only interested in the displacements and stresses of the frame, not the engine.  What is an engineer to do!?!

If we were to solve for the beams and only take the applied weight of the engine we would be neglecting inertial effects as well as any moment around one of the beams' axes.  And so, while a normally applied load to the mounting plate sitting atop the horizontal member may be "close enough" our numerical tools permit us to analyze this system with greater detail.

The question then becomes, using SolidWorks Simulation, what method should be employed?  Our choices are: a remote load/mass that replaces the engine as well as its center of gravity and mass moments of inertia or a "distributed" load that represents the mass of the engine.

In the former, I can use SolidWorks to calculate the mass and the mass moments of inertia (Tools, Mass Properties). In the latter I need only know the engine's mass. (Additionally, since I will use the Gravity boundary condition I could apply the mass * acceleration = weight as a normal force…. but that's too easy!) If the model I intend on removing from the problem is not modeled with detail then guessing the mass and CG is perhaps my only recourse.  My preference would be to include as much information about the model I'm eliminating from the problem.

"What's the difference?  They both include the mass." Not entirely true. A Remote Mass is attached to the face selection via rigid bars and thus adds stiffness to the model. One needs to provide the mass AND the CG. (There is an easier way)

The Distributed Mass is uniformly applied to the selected entity without said rigid bars. Think of it as being "melted" to the face it is applied to.

Both Remote Load/Mass and Distributed Mass can be found by right-mouse clicking the External Loads folder of your study. Additionally, and more easily, a Remote Mass can be created by right-mouse clicking the component from the Parts list of the study. Treat_as_remote_mass

Remote_Mass_in_tree

 The Remote Mass becomes its own folder.  The solid body still requires a material as well as having Gravity turned on as a boundary condition.

How do the results compare between a distributed mass and a remote mass?  Let's compare results using SolidWorks Simulation's easy-to-use Compare Results tool. Firstly, the bending moments. Followed by the shear diagrams. And next displacement results. Lastly, stress comparisons.

As you review the above results you may be saying to yourself, "C'mon Chris, they're so close… does it really matter?" I'll leave that up to you.  But you will note that the moments introduced by the remote mass could not be calculated correctly if you use a distributed mass.  And that's the point: there's more than one way to solve this but there's a more correct way to arrive at the results.

Chris Schaefer is a Simulation Specialist at Graphics Systems, a SolidWorks Value Added Reseller with locations across Wisconsin and Illinois. He is a regular contributor to the Graphics Systems’ blog: SolidNotes.com, your source for SolidWorks, Simulation, Data Management, & Product Communication Tips & Tricks.

  • http://profile.typepad.com/d107022020177139467 D

    Good article Chris!!
    Thanks.
    This works great for static analysis. But is equally important in dynamic analysis as you will get inaccurate frequency values & mode shapes corresponding to the modes that have rotational modes (twisting etc).
    It is more challenging in dynamic analysis to determine whether to leave out a component or not. Any thoughts on that?

    -Sandeep P.