PGA

02-Analytical capabilities
Post Reply
zbadar
Posts: 9
Joined: 18 Mar 2014, 04:57

PGA

Post by zbadar »

Dear Seismosoft tool experts,

I am perhaps confused with a very insignificant question. For several recorded accelograms the PGA is reported as 0.3g or 0.4g etc in case of major earthquakes. But in case of response spectrums mentioned in various country specific seismic codes the PGA is as 1g or 1.5g etc quite big compared to major measured real accelograms. What is the explanation for the same? When the spectrum compatible artificial accelograms are generated they are also scaled to these very high PGAs which is confusing me a bit as that look a bit unrealistic and over-conservative.

Thank you for your support in advance! zbadar
User avatar
ruipinho
Posts: 165
Joined: 07 Dec 2013, 14:37

Re: PGA

Post by ruipinho »

Dear zbadar,

The way response spectra are defined in different seismic design codes tends to vary, hence it is not necessarily straightforward for me to provide an answer that is general and applicable to all cases.

However, I can certainly note that, in the case of the European regulations (EC8), for instance, the response spectrum needs indeed to be scaled to the level of PGA appropriate for the site in question. In other words, EC8 provides the spectrum shape, and then users need to anchor it to the correct level of PGA.

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

Re: PGA

Post by huffte »

I'll also add that code spectra are generally envelopes of spectra from many different sources. Typically, distant and large magnitude earthquakes control the spectrum shape at longer periods (say, 1 second and higher, roughly) where large displacements are felt, while closer and smaller magnitude earthquakes control the low period range - including the PGA - where large accelerations are felt.

So I believe that your observation is correct. So, it is likely that you would need to select smaller magnitude records to scale to a PGA and larger magnitude records to scale to a longer period (1 second for example).

The keys are twofold:
(1) code based spectral shapes do not represent a single earthquake, but the potential shaking from many different sources of varying magnitude and distance
(2) large magnitude earthquakes generally produce large displacements at longer periods; smaller magnitude earthquakes generally produce larger accelerations in the lower period range. Note that these are generalizations and cannot be used as hard and fast rules.

This is actually very important to realize. I have seen various manufacturers of structural elements advertise that their product has been tested for the highest acceleration ever recorded in an actual earthquake. This is not necessarily as impressive as it sounds because a structure with a long period will be affected much more severely by shaking with low frequency/large displacement characteristics than it will be by shaking possessing high frequency/large acceleration characteristics.

As a final note, the science of scaling records to code spectra is a hot topic of late and scaling to the PGA only is not necessarily the best criteria. There is great coverage of the topic in:

1. NIST GCR 11-917-15, "Selecting and Scaling Earthquake Ground Motions for Performing Response-History Analysis", November 2011.

2. Haselton, Whittaker, Hortascu, Baker, Bray, and Grant, "Selecting and Scaling Earthquake Ground Motions for Performing Response-History Analysis", 15th World Conference on Earthquake Engineering, Lisbon, Portugal 2012.
Tim Huff
zbadar
Posts: 9
Joined: 18 Mar 2014, 04:57

Re: PGA

Post by zbadar »

Dear ruipinho and huffte, thanks so far for your very supportive feedbacks. In order that I have understood ruipinho's answer properly, I would like to take one practical example. For example, my location in Germany falls in Type 1 and ground type E category as per EC8 with PGA equivalent to 1.4g. Where should now I look for the appropriate PGA of my location? Is their any comprehensive resource where it could be found for Germany or Europe? Another problem if the PGA of my location is far below like 0.3g than by correcting for the same in SeismoArtif the convergence might be challenging or impossible. How could we address this issue? Thank you very much in advance!
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: PGA

Post by huffte »

zbadar, I am not sure I understand your situation entirely. You say your location has a PGA equal to 1.4g (that's a really high PGA)? But then you say where can you find the appropriate PGA for your location? So, I wonder if you could elaborate a bit, because I do not really comprehend clearly what your question is. Thanks very much.
Tim Huff
zbadar
Posts: 9
Joined: 18 Mar 2014, 04:57

Re: PGA

Post by zbadar »

Hello huffte, indeed it might look like an ambiguous question. Actually, I was asking it keeping in mind the answer of ruipinho in the same thread from yesterday. As I understood it yesterday for example EC8 specifies response spectrum for given site conditions. The acceleration at zero period is the PGA according to response spectrum. Now my site PGA known from any other reliable source is different and far lower compared to PGA from the response spectrum. So the response spectrum needs to be scaled to the appropriate PGA for the site. If I have understood the message wrong please correct me otherwise I would repeat my question again:
1) Is there any reliable source where I could find the PGA for different sites e.g. for Germany/Europe(in response spectrum the envelope of many earthquakes is used and therefore, PGA might be too high and overconservative)
2) How to correct the response spectrum or artificially generated accelogram considering site PGA is lower than PGA from response spectrum mentioned in country code e.g. EC8? In Seismoartif I have seen if specified PGA is far lower than PGA from response spectrum than convergence is the issue.

Hope I have been able to convey my quetion clearly this time? Thank you!
User avatar
ruipinho
Posts: 165
Joined: 07 Dec 2013, 14:37

Re: PGA

Post by ruipinho »

Hi zbadar,

You may consider visiting this web portal (http://www.efehr.org:8080/jetspeed/portal/hazard.psml) for the most up-to-date values of seismic hazard for European countries.

Rui
Post Reply

Return to “02-Analytical capabilities”