PGA matching

02-Analytical capabilities
Post Reply
Dave Hallman
Posts: 1
Joined: 15 Aug 2012, 14:53

PGA matching

Post by Dave Hallman »

I am getting a good match with the input spectra, but not with the PGA. I am concerned about scaling the matched records by PGA in subsequent analyses thus throwing off the match to the rest of the spectra. Is there some way to specifically force a closer match or scale to PGA? Perhaps scaling the input time-history to the target PGA prior to importing and performing the match?

thanks
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: PGA matching

Post by seismosoft »

Dear Dave Hallman,

That is correct; you may/should set the 'Scale factor' paramater to a value that is approximately equal to the ratio between the target PGA and the peak acceleration value of the original record.

Seismosoft Support
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: PGA matching

Post by seismosoft »

We would like to correct our previous response by noting that defining the Scale Factor as described above may not necessarily lead to a matched record with a PGA value that is coincident with the PGA of the target response spectrum.

Indeed, the PGA value of the matched record will depend also on the matching period range:
- if the latter includes period values close to zero (e.g. 0.05 < T < 0.5) then it is more likely that one will manage to get a matched PGA that is equal to the target PGA
- if, instead, the period range is away from T=0 (e.g. 0.5 < T < 2.0), then it will be difficult to match the target PGA, even if one uses a Scale Factor as defined above

Seismosoft Support
jhoroll
Posts: 1
Joined: 20 Oct 2015, 18:01

Re: PGA matching

Post by jhoroll »

I scale a seismic signal to a spectrum in addition to the maximum acceleration to be registered in the escalating signal is of 0.4g, and I made the escalameinto but the maximum acceleration is 0.51g. This can be done? command that I can use? that is what I would recommend.
Thank you
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: PGA matching

Post by huffte »

Hi jhoroll. Let me see if I understand your problem.

1. You scaled an accelerogram to a target.
2. The target PGA is 0.40.
3. Your scaled accelerogram has a PGA of 0.51.

This is not unusual, I would say. You really have to make a determination as to what range of periods is important in your analysis.

If the accelerogram is going to be used to load a structure with a natural period of 1 second, then you might consider a matching criteria in the range of, say, 0.2 seconds to 2 seconds, or something similar. This is just an example.

If PGA really is your only concern, then simply multiply the matched accelerogram by the factor (0.40/0.51). Note, however, that PGA is not necessarily a good measure of damage potential. Nonetheless, some researchers find the need to have a loading of some specific PGA.

If you are really interested in short periods in addition to PGA, then you could set the matching period range to something like 0.02 - 0.2 seconds or something similar. You may get the best result for your particular problem by a sensitivity study on spectral match as you change the period range of matching. Best of luck.
Tim Huff
mhfansuri
Posts: 1
Joined: 17 Sep 2019, 16:21

Re: PGA matching

Post by mhfansuri »

Hi everyone

I am doing the like same problem. I am interesting what somebody says in above. Because I am very curious how does to scale acceleration to target, e.g., from 0.19 to 0.21 g. Do you have any provisions that say to scale using the equation of scale is PGA'/PGAo, where PGA' is PGA target and PGAo is PGA existing?
Could we say so? Please give me suggestions.
Thank you.
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: PGA matching

Post by huffte »

If indeed, you do wish to scale a record to a specific PGA, then you are correct and it is quite simple. Very seldom, if ever however, do design codes and specifications recommend or require records to be modified in this manner. Typically, a more realistic approach is to minimize mean-square-error between the target and the record spectra over a range of periods rather than exactly match the target at a single period (PGA). For the latest on ground motion selection and modification, I recommend the freely available:

NIST GCR 11-917-15: Selecting and Scaling Earthquake Ground Motions for Performing Response-History Analyses
Tim Huff
Post Reply

Return to “02-Analytical capabilities”