Need FEA assistance

[chipwitch]

I'm having difficulty understanding Inventor's FEA analysis.  Just when I'm beginning to think I have a useable understanding of it, I run into something like the attached assembly.  The image below illustrates the analysis modeled in the assembly file.  The load is at the hole on the far right in the part called the "Top Link."  The Top Link is connected to the 3" square tubing body via a .625" steel pin, invisible for clarity.  Also invisible is the only grounded, fixed part to the left.  The illustration below shows two sets of flanges protruding.  Through these holes, .75" steel pins attach the body to the invisible grounded, fixed part.  All of the pin joints are Sliding/no Separation. 

As you can see, the greater stresses are at the upper left pin connection.  I expected the greatest stress to be at the lower two pin connections.

The only other thing I should mention, but don't think is relevant is that I included a frictionless joint on the top surface of the top link.  Without it, the top link has too many DOF and was causing divergence.  Even the parts would grow dramatically in size, even when set to "Actual".

I'm very confused and would greatly appreciate it if someone could tell me where I'm going wrong.

[chipwitch]

I'm having difficulty understanding Inventor's FEA analysis.  Just when I'm beginning to think I have a useable understanding of it, I run into something like the attached assembly.  The image below illustrates the analysis modeled in the assembly file.  The load is at the hole on the far right in the part called the "Top Link."  The Top Link is connected to the 3" square tubing body via a .625" steel pin, invisible for clarity.  Also invisible is the only grounded, fixed part to the left.  The illustration below shows two sets of flanges protruding.  Through these holes, .75" steel pins attach the body to the invisible grounded, fixed part.  All of the pin joints are Sliding/no Separation. 

As you can see, the greater stresses are at the upper left pin connection.  I expected the greatest stress to be at the lower two pin connections.

The only other thing I should mention, but don't think is relevant is that I included a frictionless joint on the top surface of the top link.  Without it, the top link has too many DOF and was causing divergence.  Even the parts would grow dramatically in size, even when set to "Actual".

I'm very confused and would greatly appreciate it if someone could tell me where I'm going wrong.

[admaiora]

Hi Chip,

There are part with missing relationships and contacts, not enough remove visibility, I suggest you to exlude the parts out of simulation, and use the pin constraint.

[admaiora]

Hi Chip,

There are part with missing relationships and contacts, not enough remove visibility, I suggest you to exlude the parts out of simulation, and use the pin constraint.

[chipwitch]

Hi, I only made those parts invisble in order to view the analysis.  I still wanted them in the analysis.  That's why I didn't exclude them.

I guess pin constraints would work, but doing so doesn't allow me to analyse the other parts in the assembly.  I didn't mean to say that the two locations you indicated attached to a grounded part.  There's actually another part they're pinned to and THAT part is attached to the grounded/fixed part.  I prefer the entire assembly be done in a single analysis rather than break it up into multiples, analysing one or two parts at a time.  Am I expecting more out of the FEA plugin than it's capable of?

[chipwitch]

Hi, I only made those parts invisble in order to view the analysis.  I still wanted them in the analysis.  That's why I didn't exclude them.

I guess pin constraints would work, but doing so doesn't allow me to analyse the other parts in the assembly.  I didn't mean to say that the two locations you indicated attached to a grounded part.  There's actually another part they're pinned to and THAT part is attached to the grounded/fixed part.  I prefer the entire assembly be done in a single analysis rather than break it up into multiples, analysing one or two parts at a time.  Am I expecting more out of the FEA plugin than it's capable of?

[innovatenate · Accepted solution]

Are you okay with the resulting reaction forces in the frictionless constraint?

I suspect having to correct for the rigid body motion of "top link" component may be interferring with results, unexpectedly. Seems like it would be a lot easier to isolate the Stinger component and apply the pin constraints and bearing loads directly. 

I hope that helps.

Thanks,

[innovatenate · Accepted solution]

Are you okay with the resulting reaction forces in the frictionless constraint?

I suspect having to correct for the rigid body motion of "top link" component may be interferring with results, unexpectedly. Seems like it would be a lot easier to isolate the Stinger component and apply the pin constraints and bearing loads directly. 

I hope that helps.

Thanks,

[chipwitch]

Nathan, thank you.  You correctly identified the problem. 

The model I've shown you is a pared down version I used to try and isolate the problem.  The actual assembly consists of additional links in the system.  The hitch is a movable scaffold of sorts that supports a connection for towing.  It will be subjected to loads of varying directions as well as magnitude.  Couple that with the geometry of the design also changing shape means those loads will act on individual parts that are less predictable depending on the current configuration of the "scaffolding."

In any case, I've previously modified the simulation to use the pin constraints as you've suggested, and the results were as expected.  The key was the frictionless constraint.  I didn't expect that kind of reaction moment.  Nor was I aware of the Reaction Forces dialog!  That was huge for me.  Thanks.  The pin constraints are not an option for reasons I stated above.  However, in response to your suggestion regarding the frictionless constraint, I removed it and bonded the contacts between the stinger and top link.  That simulation was successful.

I had originally placed the frictionless constraint to overcome a divergent result.  I had incorrectly assumed that the force, being applied in a direction parallel to an edge on the fixed component in the system would resolve to a state of equilibrium.  Instead, the uncontstrained top link would swing to a wildly unrealistic postion.  Clearly, the frictionless constraint was a poor choice.  Ironically, had I ran the simulation on the full-blown assembly, there would have been no unconstrained movement in the system and I probably wouldn't encountered a problem.

Thanks again.

Sherri

[chipwitch]

Nathan, thank you.  You correctly identified the problem. 

The model I've shown you is a pared down version I used to try and isolate the problem.  The actual assembly consists of additional links in the system.  The hitch is a movable scaffold of sorts that supports a connection for towing.  It will be subjected to loads of varying directions as well as magnitude.  Couple that with the geometry of the design also changing shape means those loads will act on individual parts that are less predictable depending on the current configuration of the "scaffolding."

In any case, I've previously modified the simulation to use the pin constraints as you've suggested, and the results were as expected.  The key was the frictionless constraint.  I didn't expect that kind of reaction moment.  Nor was I aware of the Reaction Forces dialog!  That was huge for me.  Thanks.  The pin constraints are not an option for reasons I stated above.  However, in response to your suggestion regarding the frictionless constraint, I removed it and bonded the contacts between the stinger and top link.  That simulation was successful.

I had originally placed the frictionless constraint to overcome a divergent result.  I had incorrectly assumed that the force, being applied in a direction parallel to an edge on the fixed component in the system would resolve to a state of equilibrium.  Instead, the uncontstrained top link would swing to a wildly unrealistic postion.  Clearly, the frictionless constraint was a poor choice.  Ironically, had I ran the simulation on the full-blown assembly, there would have been no unconstrained movement in the system and I probably wouldn't encountered a problem.

Thanks again.

Sherri