Force direction / decomposition

[TomHia]

Hi,

General Nastran question (not version specific). Talking about a derived welded object to avoid contacts (single part file for analysis).

A quick question having not found an answer in google and Nastran guides.

If I have an inclined surface (not orthogonal to the global or part coordinate system) such as this one:

 

and want to load a split line on it (thinking of the force equal to the mass of a cylindrical object resting on the faces):

a) does Nastran decompose the force and load it adequatly,

b) shall I calculate the forces normal to the faces on which the split lines reside by hand, create a sketch with lines normal to the split lines and use them as "geometric entity" under "Direction",

c) shall I calculate the forces by hand and create a model with correct coordinate system,

d) a better solution <insert here>.

[TomHia]

Hi,

General Nastran question (not version specific). Talking about a derived welded object to avoid contacts (single part file for analysis).

A quick question having not found an answer in google and Nastran guides.

If I have an inclined surface (not orthogonal to the global or part coordinate system) such as this one:

 

and want to load a split line on it (thinking of the force equal to the mass of a cylindrical object resting on the faces):

a) does Nastran decompose the force and load it adequatly,

b) shall I calculate the forces normal to the faces on which the split lines reside by hand, create a sketch with lines normal to the split lines and use them as "geometric entity" under "Direction",

c) shall I calculate the forces by hand and create a model with correct coordinate system,

d) a better solution <insert here>.

[John_Holtz]

Hi @TomHia 

You need to apply the force in the direction that it occurs in reality.

In your image, the force is in the Y direction. Is that the direction you want the force to be in?

• Yes. You do not need to do anything different. (Nastran is going to apply the load in the Y direction!)
• No. You need to apply the force in the correct direction. If the correct direction has components in X, Y, Z, then either b, c, or d will work. Use whichever method you prefer.

You mentioned a cylinder sitting on the inclined surface. If the 1000 lb force in Y is due to the weight in the Y direction, then your load may be okay. Most cylinders will roll 😊 so depending on how that is prevented may change what load you want to apply. (Is the rolling prevented by your model or is it prevented by something outside of your model?) If you do a free body diagram of the cylinder and put the forces on it, the forces coming from your model are the forces you want to apply to your model.

Hope this helps.

John

[John_Holtz]

Hi @TomHia 

You need to apply the force in the direction that it occurs in reality.

In your image, the force is in the Y direction. Is that the direction you want the force to be in?

• Yes. You do not need to do anything different. (Nastran is going to apply the load in the Y direction!)
• No. You need to apply the force in the correct direction. If the correct direction has components in X, Y, Z, then either b, c, or d will work. Use whichever method you prefer.

You mentioned a cylinder sitting on the inclined surface. If the 1000 lb force in Y is due to the weight in the Y direction, then your load may be okay. Most cylinders will roll 😊 so depending on how that is prevented may change what load you want to apply. (Is the rolling prevented by your model or is it prevented by something outside of your model?) If you do a free body diagram of the cylinder and put the forces on it, the forces coming from your model are the forces you want to apply to your model.

Hope this helps.

John

[TomHia]

Hi John,

Keeping it simple - if I had a case with a body on a slope resting on an obstacle, such as:

Do I need to decompose gravity force acting on the body to Fn and Fs + use either of b/c/d steps or is it an overkill and exactly the same solution will be obtained if I only set the Fg force acting towards the ground (Y axis)?

(sure, I'll test later in different circumstances but I'd like to get a gist of the 'most correct' approach).

 

[TomHia]

Hi John,

Keeping it simple - if I had a case with a body on a slope resting on an obstacle, such as:

Do I need to decompose gravity force acting on the body to Fn and Fs + use either of b/c/d steps or is it an overkill and exactly the same solution will be obtained if I only set the Fg force acting towards the ground (Y axis)?

(sure, I'll test later in different circumstances but I'd like to get a gist of the 'most correct' approach).

 

[John_Holtz · Accepted solution]

Hi @TomHia 

You need to be very careful here because we do not know what is included and not included in your model. (As far as I know, you have a dark blob of something. 🙂)

Your free body diagram shows bodies A, B, C. (B and C may be the same body or separate bodies.)

• If A is in the model, you apply the force Fy to A. The constraints and/or contact with B and C, however body A is retained, will result in the reaction forces FRn and FRs.
• If B is the only thing in the model, the only load acting on it is Fn and you should only apply Fn to part B. Whatever causes Fs only happens on body A which is not in the model in this scenario.
• If B and C are connected together, applying Fn and Fs (normal to the faces of course) is equivalent to applying Fy in the vertical direction. At least at the level of the free body diagram. At the level of the analysis, distributing Fy over the faces on B and C means than component Fn is also applied over B and C and Fs is applied over B and C. That does not happen in reality, but if the area of interest is elsewhere in the model, how the load is applied on B & C is less important.

John

[John_Holtz · Accepted solution]

Hi @TomHia 

You need to be very careful here because we do not know what is included and not included in your model. (As far as I know, you have a dark blob of something. 🙂)

Your free body diagram shows bodies A, B, C. (B and C may be the same body or separate bodies.)

• If A is in the model, you apply the force Fy to A. The constraints and/or contact with B and C, however body A is retained, will result in the reaction forces FRn and FRs.
• If B is the only thing in the model, the only load acting on it is Fn and you should only apply Fn to part B. Whatever causes Fs only happens on body A which is not in the model in this scenario.
• If B and C are connected together, applying Fn and Fs (normal to the faces of course) is equivalent to applying Fy in the vertical direction. At least at the level of the free body diagram. At the level of the analysis, distributing Fy over the faces on B and C means than component Fn is also applied over B and C and Fs is applied over B and C. That does not happen in reality, but if the area of interest is elsewhere in the model, how the load is applied on B & C is less important.

John