Bending spring analysis

[Parxx]

Hello everybody,

I am trying to obtain the linear velocity at the end of the bar after a release from the horizontal position at the end of this bar linked to the ground by a spring connector:    

I am only interested by the linear velocity resulting of the bending spring unloading.

It is exactly the same movement than the children spring swings like this:

I did direct transient response analysis, but I never succeed to have good results.

I have troubles of unrealistic displacements and linear velocities (the velocity is in m/s) :

Here you have an example of setting that I did for my sprinng connector (N/m & Nm/rad):

Here you have the details of my problems:

• The linear velocity increase if the enforced motion or moment charges is reset to 0 instantaneously like this:
• The linear velocity also depends of my dynamic settings, more I decrease the last of the time steps, more the linear speed will increase, which is completely illogical again
• The displacements when I do the Multiset Animation Setting option are completely unrealistic (you have the video in attached)

To summarize, it seems that the system behaves as if the bar don’t have any moment of inertia, the acceleration will tend to the infinite if the spring is unloaded instantaneously, but for the link with the times steps I have no idea.

It is very strange because I putted a density which isn’t equal to 0 in my materials, I also putted an area on my bar and gravity loads.

I tried all the way, I replaced enforced motion by moments, I restarted the analysis with new parts, I tried to do it with solid elements instead of the beam, I replaced the spring connector by a coil, etc.

I tried also with the damping and not, with diffent kind and size of meshs, etc.

Maybe it is not a direct transient response that I have to do? I thought to the explicit dynamic analysis, but it is only for the solid and the time to solve them is to high, the nonlinear transient responses don’t allow the enforced motions, …

In attached you have my model.

Thank you very much in advance for your help!

[Parxx]

Hello everybody,

I am trying to obtain the linear velocity at the end of the bar after a release from the horizontal position at the end of this bar linked to the ground by a spring connector:    

I am only interested by the linear velocity resulting of the bending spring unloading.

It is exactly the same movement than the children spring swings like this:

I did direct transient response analysis, but I never succeed to have good results.

I have troubles of unrealistic displacements and linear velocities (the velocity is in m/s) :

Here you have an example of setting that I did for my sprinng connector (N/m & Nm/rad):

Here you have the details of my problems:

• The linear velocity increase if the enforced motion or moment charges is reset to 0 instantaneously like this:
• The linear velocity also depends of my dynamic settings, more I decrease the last of the time steps, more the linear speed will increase, which is completely illogical again
• The displacements when I do the Multiset Animation Setting option are completely unrealistic (you have the video in attached)

To summarize, it seems that the system behaves as if the bar don’t have any moment of inertia, the acceleration will tend to the infinite if the spring is unloaded instantaneously, but for the link with the times steps I have no idea.

It is very strange because I putted a density which isn’t equal to 0 in my materials, I also putted an area on my bar and gravity loads.

I tried all the way, I replaced enforced motion by moments, I restarted the analysis with new parts, I tried to do it with solid elements instead of the beam, I replaced the spring connector by a coil, etc.

I tried also with the damping and not, with diffent kind and size of meshs, etc.

Maybe it is not a direct transient response that I have to do? I thought to the explicit dynamic analysis, but it is only for the solid and the time to solve them is to high, the nonlinear transient responses don’t allow the enforced motions, …

In attached you have my model.

Thank you very much in advance for your help!

[John_Holtz]

Hi @Parxx 

What version of Inventor Nastran are you using? And what frequency of vibration do you expect?

For unknown reasons, my computer has problems showing the model. I am not sure if the following observations are correct or not.

1. The time step size is 0.2 seconds. This seems to be very large unless the model vibrates very slowly. You want 10 to 20 calculations per cycle which implies the period would need to be longer than 2 or 4 seconds. This is not what is being shown by your graph. (The graph implies several cycles per second, but it is hard to interpret the total velocity. You should be displaying the velocity in the X, Y, or Z direction so that you can see negative and positive values.)
2. Not sure if you can use an enforced motion load. First off, you want to pull down the end and release it. The load curve (Transient Table Data) of (0,0) to (2,1) to (2,0) says to ramp the enforced motion to full value over 2 seconds and then push it back to 0 displacement instantaneously. A value of 0 does not release it! I am not sure if you can remove an enforced motion from the analysis. (Perhaps you can by using two subcases?)
3. An enforced rotation of 90 radians is wrong. (That is 5156.6 degrees!)

Inventor has crashed multiple times on me. Maybe it is telling me to stop doing work on a holiday weekend. 🙂

[John_Holtz]

Hi @Parxx 

What version of Inventor Nastran are you using? And what frequency of vibration do you expect?

For unknown reasons, my computer has problems showing the model. I am not sure if the following observations are correct or not.

1. The time step size is 0.2 seconds. This seems to be very large unless the model vibrates very slowly. You want 10 to 20 calculations per cycle which implies the period would need to be longer than 2 or 4 seconds. This is not what is being shown by your graph. (The graph implies several cycles per second, but it is hard to interpret the total velocity. You should be displaying the velocity in the X, Y, or Z direction so that you can see negative and positive values.)
2. Not sure if you can use an enforced motion load. First off, you want to pull down the end and release it. The load curve (Transient Table Data) of (0,0) to (2,1) to (2,0) says to ramp the enforced motion to full value over 2 seconds and then push it back to 0 displacement instantaneously. A value of 0 does not release it! I am not sure if you can remove an enforced motion from the analysis. (Perhaps you can by using two subcases?)
3. An enforced rotation of 90 radians is wrong. (That is 5156.6 degrees!)

Inventor has crashed multiple times on me. Maybe it is telling me to stop doing work on a holiday weekend. 🙂

[Parxx]

Hi John,

Thanks for your answer.

Here you have mines:

What version of Inventor Nastran are you using? Inventor 2022 Professionnal 

And what frequency of vibration do you expect? It is a traffic sign with a human size so I would say less than 10 Hz.

The time step size is 0.2 seconds. This seems to be very large unless the model vibrates very slowly. You want 10 to 20 calculations per cycle which implies the period would need to be longer than 2 or 4 seconds. This is not what is being shown by your graph. (The graph implies several cycles per second, but it is hard to interpret the total velocity. You should be displaying the velocity in the X, Y, or Z direction so that you can see negative and positive values.) OK, I just decrease the time steps and did a view on the Z axis, I have nothing on the others:

Here the time step size is between 0,01 and 0,05 s:

 

Not sure if you can use an enforced motion load. First off, you want to pull down the end and release it. The load curve (Transient Table Data) of (0,0) to (2,1) to (2,0) says to ramp the enforced motion to full value over 2 seconds and then push it back to 0 displacement instantaneously. A value of 0 does not release it! I am not sure if you can remove an enforced motion from the analysis. (Perhaps you can by using two subcases?). I already tried with two subcase, my screenshot just above is with two subcases (the first for the loading and the second for the unloading). Unfortunately, the trouble is always here, I have speed which are more than 2.10^5 m/s! It is just a bar of a human size with a spring, it is not possible to have this. As I said before, the only way to decrease the speed is increase the last of the time steps or decrease the enforced motion gradually, but it is not that I want. The bar will have to come back alone to the vertical position from the horizontal one, so it is important to release brively the enforced motion once the 90° angle is reached.

An enforced rotation of 90 radians is wrong. (That is 5156.6 degrees!) Yes it was a mistake, but I already did other analysis with the correct rotation (90°), as it is the case with the graph above.

In attached you have my document modified, the trouble came from the language version, I worked on it in French, I it will be difficult for you to open it with the french version of Inventor; so I will start again the document in english and send it to you. I know this because I tried also with the english version and Inventor crash as like you.

For the moment I send you the french document again with the 2nd subcase update.

If it is not possible to release an enforced motion, is there another way to have the values that I want?

[Parxx]

Hi John,

Thanks for your answer.

Here you have mines:

What version of Inventor Nastran are you using? Inventor 2022 Professionnal 

And what frequency of vibration do you expect? It is a traffic sign with a human size so I would say less than 10 Hz.

The time step size is 0.2 seconds. This seems to be very large unless the model vibrates very slowly. You want 10 to 20 calculations per cycle which implies the period would need to be longer than 2 or 4 seconds. This is not what is being shown by your graph. (The graph implies several cycles per second, but it is hard to interpret the total velocity. You should be displaying the velocity in the X, Y, or Z direction so that you can see negative and positive values.) OK, I just decrease the time steps and did a view on the Z axis, I have nothing on the others:

Here the time step size is between 0,01 and 0,05 s:

 

Not sure if you can use an enforced motion load. First off, you want to pull down the end and release it. The load curve (Transient Table Data) of (0,0) to (2,1) to (2,0) says to ramp the enforced motion to full value over 2 seconds and then push it back to 0 displacement instantaneously. A value of 0 does not release it! I am not sure if you can remove an enforced motion from the analysis. (Perhaps you can by using two subcases?). I already tried with two subcase, my screenshot just above is with two subcases (the first for the loading and the second for the unloading). Unfortunately, the trouble is always here, I have speed which are more than 2.10^5 m/s! It is just a bar of a human size with a spring, it is not possible to have this. As I said before, the only way to decrease the speed is increase the last of the time steps or decrease the enforced motion gradually, but it is not that I want. The bar will have to come back alone to the vertical position from the horizontal one, so it is important to release brively the enforced motion once the 90° angle is reached.

An enforced rotation of 90 radians is wrong. (That is 5156.6 degrees!) Yes it was a mistake, but I already did other analysis with the correct rotation (90°), as it is the case with the graph above.

In attached you have my document modified, the trouble came from the language version, I worked on it in French, I it will be difficult for you to open it with the french version of Inventor; so I will start again the document in english and send it to you. I know this because I tried also with the english version and Inventor crash as like you.

For the moment I send you the french document again with the 2nd subcase update.

If it is not possible to release an enforced motion, is there another way to have the values that I want?

[Parxx]

Hi @John_Holtz ,

I did the document in english version, now I can start analysis without crash, so I think it will be the same for you, you have it in attached.

I also tried a new kind of analysis, instead of an enforced rotation, I tried with an initial condition of rotation:

Unfortunately it didn't work, I tried with only one, the two subcase but the bar didn't move at all.

May I do something wrong, initial conditions may be the solution? I suppose initial conditions loads need to be constraint in the same direction like the other loads? I tired with contraint and without but the result was the same...

Thanks in advance for your help, have a nice day.

[Parxx]

Hi @John_Holtz ,

I did the document in english version, now I can start analysis without crash, so I think it will be the same for you, you have it in attached.

I also tried a new kind of analysis, instead of an enforced rotation, I tried with an initial condition of rotation:

Unfortunately it didn't work, I tried with only one, the two subcase but the bar didn't move at all.

May I do something wrong, initial conditions may be the solution? I suppose initial conditions loads need to be constraint in the same direction like the other loads? I tired with contraint and without but the result was the same...

Thanks in advance for your help, have a nice day.

[John_Holtz]

Hi @Parxx 

I cannot think of a situation where an initial condition of rotation would make any sense, but maybe I do not understand the load type. As far as I know, the initial condition does not create an initial stress. How can there be an initial displacement (or rotation) without creating an initial stress?

I tried using an enforced motion with two subcases, but I could not get it to work. I will ask a co-worker if there is some way to do it.

Until then, this is what I would do:

1. Run a Normal Modes analysis to get the natural frequency. (The period = 1/natural frequency, and you will use that information to set the time step size.)
2. Run a linear static stress analysis with an enforced motion to find out how much force is required to pull the sign the desired amount. (Or if "close" is good enough, you can do a hand calculation to get the force.)
3. Run a direct transient analysis with two subcases. Subcase 1 has the force applied, a relatively long duration, and a few steps. (You want this portion to be close to static so there is no vibration occurring during the loading.) Subcase 2 is a short duration (5 to 10 full cycles) with small time steps (1/20 the period). Subcase 2 has the same constraints applied (meaning the constraint applied to the model is assigned to subcases 1 and 2) but no load applied. (Or maybe a small dummy load so that Inventor does not complain about a missing load in subcase 2.)

[John_Holtz]

Hi @Parxx 

I cannot think of a situation where an initial condition of rotation would make any sense, but maybe I do not understand the load type. As far as I know, the initial condition does not create an initial stress. How can there be an initial displacement (or rotation) without creating an initial stress?

I tried using an enforced motion with two subcases, but I could not get it to work. I will ask a co-worker if there is some way to do it.

Until then, this is what I would do:

1. Run a Normal Modes analysis to get the natural frequency. (The period = 1/natural frequency, and you will use that information to set the time step size.)
2. Run a linear static stress analysis with an enforced motion to find out how much force is required to pull the sign the desired amount. (Or if "close" is good enough, you can do a hand calculation to get the force.)
3. Run a direct transient analysis with two subcases. Subcase 1 has the force applied, a relatively long duration, and a few steps. (You want this portion to be close to static so there is no vibration occurring during the loading.) Subcase 2 is a short duration (5 to 10 full cycles) with small time steps (1/20 the period). Subcase 2 has the same constraints applied (meaning the constraint applied to the model is assigned to subcases 1 and 2) but no load applied. (Or maybe a small dummy load so that Inventor does not complain about a missing load in subcase 2.)

[Parxx]

Hi @John_Holtz,

I tried what you advised, I puted a force in the X axis of the bar reference.

But it did not work, I set the analysis as you advise me, did the modal one and follow all your steps...

I had no velocity and displacement horizontally, but a little vertically, it is not that we want.

You really wanted a force, not a moment?

Thanks in advance have a nice day.

[Parxx]

Hi @John_Holtz,

I tried what you advised, I puted a force in the X axis of the bar reference.

But it did not work, I set the analysis as you advise me, did the modal one and follow all your steps...

I had no velocity and displacement horizontally, but a little vertically, it is not that we want.

You really wanted a force, not a moment?

Thanks in advance have a nice day.