Temporal analysis for seismic simulation in RSAP2010

[limd]

Hi,

I am studying a steel pier subject to an axial force and I have to carry out some dynamic analysis. So, I have done a modal analysis and then, a seismic analysis under the eurocode 8.

But I was asked to analyze the behavior of the pier under an existing earthquake. I have the data of the earthquake, i.e., the acceleration depends on the time.

So, I wanted to know if there is a way to put those data into RSAP2010. I tried a temporal analysis by considering an imposed acceleration. But, I don't really understand what is the purpose of the function "F(T)" in the :/temporal analysis/definition of the function... Is it a coefficient for the seismic load ? Like if I have my seismic load equal to 1m/s2 and I put F(T)=5; the seismic load will be 5m/s2 ?

If you need any precision, feel free to ask.

Thanks.

PS: Sorry for my english if it's not very clear.

[waldemar.okapa]

Time function F(t) descibes simply change in time of the loads applied to the temporal analysis case. You will get results at each of the time step.
For the seismic analysis better use spectral analysis case. There is a tool to convert acceleration in time function to spectrum, it is function of aceleration depending on vibration periods. As a result of spectral analysis you will get maximum response for each dynamic modes or their combinations.

[limd]

The seismic analysis I've done in robot under eurocode 8 is using the acceleration spectrum response. I guess it's the same thing as the spectral analysis.
But according to my teacher, it's not enough accurate to simulate the real behavior since he wants me to carry out non-linear analysis. Because, the spectral analysis which is in elastic field will be obsolete. That's why I wanted to try the temporal analysis. My professor want me to simulate a real earthquake which is possible in ABAQUS but I totally don't know this software.

So, is there any way to simulate a real earthquake in robot ? I have all the datas needed, i.e. acceleration depends on time of a former earthquake.

[medo]

It is possible to perform Time history analysis. Go to Analysis>Analysis Type>New>Time History. Then define the parameters for the analysis as well as the function. You can use 'Open from file" to import your earthquake data. But, be careful to use the appropriate data format in order to be able to read your data correctly. Your file should be saved with .thf extension. You have example files in your defcfg folder.

[limd]

So, I succeed to put my data into robot. Thanks a lot ! But I still have issues ...

Well, I study a pier which I already have the results from ABAQUS. Thanks to that I can check the results from Robot.

Please find the study characteristics below:

Height=12m // Diameter=2m // Thickness=5.2cm // Material=Steel E=206 000MPa - Yield point=315MPa // Boundaries= Free at the top, encastred at the bottom // I considered a mass at the top, M=945t. And my data from ABAQUS is the accelaration depends on the time, so I guess it's kind of F(t)=a(t).

So, I tried some cases but none are correct.

Case 1: I considered the pier and the mass in X/Y/Z axis => after time history analysis, no deflection, no stress, nothing ...

Case 2: I still considered the mass at the top and add a forced acceleration at the top equal to 1m/s^2 => nothing again ... (I also tried to put the acceleration at the bottom ... no result ...)

Case 3: I only considered the pier and add an horizontal force in X axis equal to 945kN => I have results but wrong one ... according to ABAQUS, I should have a deflection of 3,3cm but I get 11,9cm with Robot.

So, what is the correct way to consider this earthquake ?

[medo]

Please, post your model here to check the data.

[limd]

Please find enclosed my rtd file and my earthquake data. It's the case with acceleration+mass.

Sorry for the delay.

[limd]

Well, I almost succeed to carry out the analysis.

In the time history analysis parameter, I have to choose the case "X direction" and also consider a mass at the top of the pier. I have to put a value for alpha (damping) in the analysis parameter. And then I obtained some results ...

But there is still some issues about the results. Indeed, when I compared with Abaqus results, the difference is too big. I mean that there is 18% of "error" in the displacement result whereas there is only 0.8% of "error" in the modal analysis.

I tried to check all the parameter but the difference is still there ... May be you can find out the error ! Please find enclosed the robot and abaqus files.

Thanks !