Arcadi
Civil/Environmental
- Nov 16, 2011
- 2
Hi to everybody, I know there are some pro SAP2000 users here, so I would like to submit my doubt.
I usually get floor spectra applying the synthetic accelerogram matching the spectrum of the applied code at the base of the structure in one direction, running a modal time history and recording the time accelerations at a node on the floor. Then in order to obtain the floor spectrum for that direction I use the Newmark Average Acceleration method with a large number of periods and damping values to get a good resolution and use them for any equipment on the structure. Then repeat for the other directions, and so on.
My question comes when I've been asked to take into account the soil damping, not only the structural. I have an offshore platform on piles into the seabed and as usual I just defined linear lateral springs according to soil properties and got the floor spectra, but now I have to consider the impedance of the soil structure interaction.
As far as I know the soil damping is frequency dependant, so higher excitation frequencies yield lower damping values, so the damping definition in SAP2000 is an hysteretic damping, but I can use this kind of link property only with the Power Spectral Density analysis (or Steady state, but not the case here).
Ok, but now, how do I introduce the accelerogram as a PSD function? I tried with the Fast Fourier Transform but I get complex positive and negative values and the SAP2000 PSD function allows only real and positive.
Does anyone have an idea? The only approximation that comes to my mind (not tried yet) is to set the damping value for the first mode of the platform and run a nonlinear time history, and since the natural period for that mode will be lengthened, iterate modifying the damping with the new structure frequency and so on until convergence. This will give me the response around that frequency but not out of that range (problem 1) and I would have to do that for the most relevant modes and each direction (problem 2), of course if someone knows how to do it using a PSD function will spare me that pain!
Any guidance will be highly appreciated.
By the way, I suppose this is the correct forum to post, not the vibrations one, if not sorry!
I usually get floor spectra applying the synthetic accelerogram matching the spectrum of the applied code at the base of the structure in one direction, running a modal time history and recording the time accelerations at a node on the floor. Then in order to obtain the floor spectrum for that direction I use the Newmark Average Acceleration method with a large number of periods and damping values to get a good resolution and use them for any equipment on the structure. Then repeat for the other directions, and so on.
My question comes when I've been asked to take into account the soil damping, not only the structural. I have an offshore platform on piles into the seabed and as usual I just defined linear lateral springs according to soil properties and got the floor spectra, but now I have to consider the impedance of the soil structure interaction.
As far as I know the soil damping is frequency dependant, so higher excitation frequencies yield lower damping values, so the damping definition in SAP2000 is an hysteretic damping, but I can use this kind of link property only with the Power Spectral Density analysis (or Steady state, but not the case here).
Ok, but now, how do I introduce the accelerogram as a PSD function? I tried with the Fast Fourier Transform but I get complex positive and negative values and the SAP2000 PSD function allows only real and positive.
Does anyone have an idea? The only approximation that comes to my mind (not tried yet) is to set the damping value for the first mode of the platform and run a nonlinear time history, and since the natural period for that mode will be lengthened, iterate modifying the damping with the new structure frequency and so on until convergence. This will give me the response around that frequency but not out of that range (problem 1) and I would have to do that for the most relevant modes and each direction (problem 2), of course if someone knows how to do it using a PSD function will spare me that pain!
Any guidance will be highly appreciated.
By the way, I suppose this is the correct forum to post, not the vibrations one, if not sorry!
