Eccentric load problem

[JKLee]

Hello, I'm asking you a question because I have a difficulty in analysis it.
An error occurred during the analysis of applying eccentric load.
Since the eccentric load cannot be applied to stko using the beam element, the analysis was conducted using moment as shown in the picture below.
I set both ends of the column as hinges and analysis it, but the following error occurred.
I think the boundary condition and algorithm are wrong, so how can I solve it?

[JKLee]

I corrected a few things and got the result values without errors.

However, unlike the desired result value as shown in the picture below, there was too much displacement.

Why are there so many wrong results?

[LuisB]

Dear JKLee,

I had quick look at your model and there are some things that got me confused. You have a prescribed nodal rotation, I do not know if what you really wanted to apply because the value is quite high. If you want to apply a moment you can use NodeCouple Condition.

In addition, be careful with the duration of the test, you input a value of 6000 which means that the applied load at the end of the analysis would be 6000 times more than what you input.

[JKLee]

Luis, thank you for answering my question even though you are busy.

I wanted to apply a 20mm eccentric load that increases by 20kN/min for 6,000 seconds.

So I entered a value of 6000, but I think I misunderstood the meaning of 'duration'.

However, i revised and analyzed the input value of the new load, but there is still an error.

This seems to cause an error when I applied a moment by multiplying the load by 20 mm to apply the eccentric effect.

I tried to apply 6666.6N·mm of moment to both ends of the column using 'sp', but 'sp' seems to be a different function.

In Opensees,
load 16 0.0 0.0 -333.33 6666.6 0.0. 0.0
load 1 0.0 0.0 333.33 -6666.6 0.0 0.0
I applied moment in this way, but how does STKO apply it?

[LuisB]

Dear JKLee,

The built-in function called NodeCouple allows to enter the moment in STKO. The output of this command will print a .tcl file with the following:

# Loads.Force NodeForce
load 16 0.0 0.0 -2000000.0 0.0 0.0 0.0
load 1 0.0 0.0 2000000.0 0.0 0.0 0.0

# Loads.Force NodeCouple
load 16 0.0 0.0 0.0 0.0 40000000.0 0.0
}
Which is equivalent of what you wrote. I viewed the model and modified some things, in this case I change to a duration of 1 and increment of 6000. The other way is what you have done in the previous model that you applied the vertical load which was 333 N with a duration of 6000.

I hope this help!

[JKLee]

Hello, Luis.

Thanks to this, I entered Moment in STKO and obtained the results close to the experimental results I wanted.

However, I entered the concrete strength low to get a closer value to the experimental results, and the unexpected results came out as shown in the picture below.

In addition, even if the concrete strength is increased by 1 MPa, strange results.

If you enter a value other than 33 MPa of concrete strength, you will get a strange result.

Which part should I correct to solve this problem?

When defining concrete, should we change not only the strength of concrete but also the other values input to concrete 02?

For Model 27, input 33 MPa of concrete strength.
Enter 25 MPa for Model 27_1 and 34 MPa for Model 27_2, and there is no difference.

[JKLee]

The above problem seems to have occurred because I did not check the p-delta in the element properties part.

The results come out well after the p-delta check. I'm sorry.

[marafini.f]

Dear JKLee,

I am happy you solved your problem.
Let us know if you need any more help, and if you could post the results of your analysis we would be interested to see them.

Enjoy your modelling.
Francesca

[JKLee]

Hello, Francesca.

In the figure below, the cylinder strength of concrete was used for Test and Model, but in some papers, the concrete strength 8MPa was modeled lower for Model_ because it was 8MPa lower than the cylinder strength. However, the professor said that it is right to use the cylinder strength shown in the experimental paper when analyzing.

So, the difference between the experimental results of the model and the test seems to require studying about materials.

[marafini.f]

Good job!
I don't remember what you are using, but if you run your analysis in displacement control you should be able to obtain the post-peak response. And yes, material calibration is really important!
Enjoy your modelling
Francesca