Response spectrum soil types

02-Analytical capabilities
Post Reply
akanyilmaz
Posts: 3
Joined: 08 Dec 2009, 14:37

Response spectrum soil types

Post by akanyilmaz »

Hello,

Something is not clear to me with seismoartif. How can I produce artifical accelerograms compatible with the response spectrums with different soil type? When I define 2 different Spectrum with type A and type E ground (Eurocode), and apply a synthetic accelerogram with "no site effects" what comes out are the two different accelreograms only with different peak acceleration values. But I expect also a difference in terms of frequency content of the accelerograms. What am I doing wrong? I doubt my choice about "no site effects" during synthetic accelerogram creation, but how else can I impose to the software to create accelerograms compatible with Eurocode soil types ?

Best regards
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: Response spectrum soil types

Post by huffte »

Hi akanyilmaz.

There are 3 primary steps. I assume you are using the "Synthetic Accelerogram and Adjustment" option. The three major steps are:

1. Target Spectra Definition
2. Synthetic Accelerogram
3. Accelerogram Generation

The major change in frequency content comes with the third step in the process, "Accelerogram Generation". This is where the frequency content is modified to match the target response spectrum. In this step you will select which previously defined target response spectrum for which you wish to attempt a match. Selecting the A spectrum will now give very different results compared to selection of the E spectrum.

The initial "Synthetic Accelerogram" definition in step 2 should, in my view, be generated using the closest match to site conditions for the spectrum you wish to match in order to maximize the convergence in step 3.

So if you want to match to the A spectrum in step 3, select a hard rock option in step 2. If you wish to match to the E spectrum in step 3, select a soft soil in step 2.

It's a great program akanyilmaz. I hope this helps.

Best of luck with your work.
Tim Huff
akanyilmaz
Posts: 3
Joined: 08 Dec 2009, 14:37

Re: Response spectrum soil types

Post by akanyilmaz »

Hi huffte,

Thanks for the help. I followed your suggestions and defined two cases:

EC8 Ground type A with vs: 940 m/s
EC8 Ground type E with vs: 310 m/s

Still What I get is only a difference in PGA, same frequency contents in both cases. I am not sure what I am doing wrong.
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: Response spectrum soil types

Post by huffte »

Ok akanyilmaz.

Let's just make certain you are performing all the steps before we see if there is a problem with the software.

When you start Artif, you see the "Calculation Method" box with 4 options. You select "Synthetic Accelerogram Generation and Adjustment".

You see 5 tabs representing the 5 data/results fields:
1. Target Spectra
2, Synthetic Accelerogram
3, Accelerogram Generation
4, Time Series
5. Response Spectra

In the "Target Spectra" tab, you successfully define the desired response spectrum, or spectra.

You then select the "Synthetic Accelerogram" tab and specify:
1. Inter-plate, Intra-plate, or Extensional tectonics
2. Desired Moment Magnitude
3. Joyner Boore distance
4. Site effects option
5. Generate the specified number of synthetic accelerograms

You then select the "Accelerogram Generation" tab and specify:
1. The spectrum to which you wish to adjust the accelerograms from the previous tab
2. The period range over which matching is to to be attempted
3. Spectrum factor if other than 1.0
4. Select the "Generate Artificial Accelerograms" button and wait for convergence

You now proceed to the "Time Series" tab to inspect the generated and adjusted time series and note which ones converged and which ones did not. Even if it does not converge, it may still be usable as matching slightly out of spec may still be applicable.

Finally, you proceed to the "Response Spectra" tab and inspect the target and matched ground motion spectra.

So, are you saying that the matching never converges and the spectra in the last step have no resemblance to the targets?

If so, I have never had this occur and I have used the program a lot.

Make sure the units in your defined spectra are correct and match the units specified in "Tools" --> "Settings". Make sure you perform each step in succession. If none of that works, then send me the Artif file and I'll see if I can reproduce the non-convergence.
Tim Huff
akanyilmaz
Posts: 3
Joined: 08 Dec 2009, 14:37

Re: Response spectrum soil types

Post by akanyilmaz »

Ok, just to update everybody with our conversation with Huffte (As I sent my model to Huffte, he replied my email with his view on the subject): In summary, my problem seems that I was expecting a more visible difference in terms of frequency content of accelerograms in two soil types. However, the code based spectra gives a more general view covering several earthquakes and distances. And the absence of a large discrepancy in the frequency content of my artifical ground motions was related to the lack of difference in frequency content of the input spectra (for type A and type E soil types).


huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: Response spectrum soil types

Post by huffte »

After reading your e-mail, I have the following additional comments, akanyilmaz.

Regarding frequency content differences between an 'A' site and a 'D' site, you are probably correct in the case of a single event. For a given earthquake the difference could be quite pronounced. However, code-based spectra are not for a given single earthquake, but attempt to envelope several earthquakes and distances – that is the reason there is not as much difference as you might have suspected. So, you see, the absence of a large discrepancy in the frequency content of the resulting ground motions is related to the lack of difference in frequency content of your input spectra. Nonetheless, code spectra may be used as targets for such analyses. It is generally conservative to do so.

And yes, you are correct - there is no 'E' classification included in SeismoArtif spectrum definition. Generally such conditions may require a site-specific analysis and the code spectrum is not available. Your issues are not SeismoArtif issues. They are code spectral shape issues. Certainly, if the code based spectral shape does not suit you, SeismoArtif offers you alternative means of defining the target spectra.

In each of your cases, SeismoArtif has done precisely what you have asked it to do - matched the synthetic accelerogram to a spectrum which you had previously specified.

Best of luck akanyilmaz.
Tim Huff
Post Reply

Return to “02-Analytical capabilities”