Reaction Force

[Anonymous]

Hi everyone,

At the moment I am using shape optimization to minimize material usage of a complicated part.

In the Static Stress everything seems logical, except the Reaction Force.

Since I'm not allowed to share the design of the part, I made a simple drawing to explain the situation. (Simple drawing)

I've also made a simple part to show my question.

As seen on Example.png, I've placed a downwards 10 kN Force on the bottomside of the part.

The hole has been constrained with a pin. (With fixed constraint, the reaction force shows up at the some location.)

The simulation shows that the Reaction force is on the bottomside of the hole.

This is not right, since the reaction force should be on the topside of the hole.

Is there another way to constrain the hole? or is there another reason why its showing up at the bottomside?

Cheers,

A

[Anonymous]

Hi everyone,

At the moment I am using shape optimization to minimize material usage of a complicated part.

In the Static Stress everything seems logical, except the Reaction Force.

Since I'm not allowed to share the design of the part, I made a simple drawing to explain the situation. (Simple drawing)

I've also made a simple part to show my question.

As seen on Example.png, I've placed a downwards 10 kN Force on the bottomside of the part.

The hole has been constrained with a pin. (With fixed constraint, the reaction force shows up at the some location.)

The simulation shows that the Reaction force is on the bottomside of the hole.

This is not right, since the reaction force should be on the topside of the hole.

Is there another way to constrain the hole? or is there another reason why its showing up at the bottomside?

Cheers,

A

[henderh]

I believe what you're seeing is the correct behavior according to how the study was set up (but I agree with you that it may not be intuitive at first impression).

For a Pin constraint, the default fixed directions are axial and radial.  Since radial is fixed, all points on the cylindrical face cannot move in the radial direction.  This means that each point on the face (each node in FEA terms) is fixed such that they'll take a portion of the load without regard to whether the load is 'pushing' (compression) or 'pulling' (traction) at that point on the face.  It's a bit tricky to conceptualize, since any constraint is taking place of a physical thing that would be holding it there, so it's really a kind of idealization / abstraction of physical reality.

In order to get the effect you're after, consider using a bearing load where it cannot pull, it can only push (in a parabolic shape load profile).  Then, reverse the targets that are selected for loads and constraint (see image below).

Another option is to continue using the Pin constraint, but only apply it to the 'upper half' of the hole.  Use a workplane to split the full cylindrical face into two half cylindrical faces beforehand.

By the way, the Reaction Force result type isn't as useful as the Reactions command (in the Inspect panel) in my opinion.  With the Reactions command, you can select the face(s) with the applied constraints, and we'll sum up the reactions (forces and moments) so you can verify that those are equal and opposite to the applied loads (forces and / or moments) to check that the structure remains, and is calculated in static equilibrium.

Hope this helps!

Edit: One caveat to the Pin constraint on half the cylinder approach to be aware of when inspecting the reaction moments:  In calculating the reactions, we sum the forces / moments about the centroid of the constrained entity.  Therefore, for the half cylinder face the centroid is not located at the axis of the cylindrical face, so there's a chance the reaction moment will not be zero for all directions (even though it's in static equilibrium, and can 'rotate' tangentially).

[henderh]

I believe what you're seeing is the correct behavior according to how the study was set up (but I agree with you that it may not be intuitive at first impression).

For a Pin constraint, the default fixed directions are axial and radial.  Since radial is fixed, all points on the cylindrical face cannot move in the radial direction.  This means that each point on the face (each node in FEA terms) is fixed such that they'll take a portion of the load without regard to whether the load is 'pushing' (compression) or 'pulling' (traction) at that point on the face.  It's a bit tricky to conceptualize, since any constraint is taking place of a physical thing that would be holding it there, so it's really a kind of idealization / abstraction of physical reality.

In order to get the effect you're after, consider using a bearing load where it cannot pull, it can only push (in a parabolic shape load profile).  Then, reverse the targets that are selected for loads and constraint (see image below).

Another option is to continue using the Pin constraint, but only apply it to the 'upper half' of the hole.  Use a workplane to split the full cylindrical face into two half cylindrical faces beforehand.

By the way, the Reaction Force result type isn't as useful as the Reactions command (in the Inspect panel) in my opinion.  With the Reactions command, you can select the face(s) with the applied constraints, and we'll sum up the reactions (forces and moments) so you can verify that those are equal and opposite to the applied loads (forces and / or moments) to check that the structure remains, and is calculated in static equilibrium.

Hope this helps!

Edit: One caveat to the Pin constraint on half the cylinder approach to be aware of when inspecting the reaction moments:  In calculating the reactions, we sum the forces / moments about the centroid of the constrained entity.  Therefore, for the half cylinder face the centroid is not located at the axis of the cylindrical face, so there's a chance the reaction moment will not be zero for all directions (even though it's in static equilibrium, and can 'rotate' tangentially).

[henderh]

You could also model the pin, then use a Separation contact between the faces of the two bodies to arrive at the most realistic result.

[henderh]

You could also model the pin, then use a Separation contact between the faces of the two bodies to arrive at the most realistic result.