Seismomatch- Matching of Time History

03-Unexpected behaviour/errors
Post Reply
rbsce.ph
Posts: 30
Joined: 29 Jul 2020, 09:29

Seismomatch- Matching of Time History

Post by rbsce.ph »

Good day!

I tried matching a particular time history curve for dynamic time history analysis using Seismomatch. The result of the matched curve was not what I desired. Do you have any guidelines in matching time history curves? Please see attached picture. Any comments and recommendations are highly appreciated. Thank you so much.

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

Re: Seismomatch- Matching of Time History

Post by huffte »

First, this is not unusual. A pioneer of spectral matching has been quoted as saying that 80% of the records generated through such processes should be discarded. This does not mean it is not a valuable tool. It just means you have to put a good deal of work into doing it correctly.

Your period range is large, from 0.05-8.0 seconds. It can be difficult to get a match over such a wide period range. If you can justify narrowing the range to only the periods of interest for your project, that would help.

The record is likely not a good candidate for spectral matching to your defined target, at least in the form shown. As you correctly observe, the modified accelerogram is fairly unrealistic. You may have to try several records to find appropriate candidates.

You can measure match to spectral shape using various methods from the literature. This may be used to identify potential candidates. That is beyond the scope of the forum, however.

You might wish to tighten the tolerances and expand the maximum number of iterations in the settings. You could investigate using the different algorithms (there are two available in SeismoMatch in the settings).

This record has an extremely short duration, 8 seconds. This is unusual. A longer total duration, even with zero-padding at the beginning and end, could help. This appears to be PEER RSN243, which I know to have a total duration of 40 seconds. You may have loaded the record incorrectly.

Best of luck in your work.
Tim Huff
rbsce.ph
Posts: 30
Joined: 29 Jul 2020, 09:29

Re: Seismomatch- Matching of Time History

Post by rbsce.ph »

Hi Sir,

We tried to download the RSN243 from NGAWest, the Time History is complete as displayed, however, when we try to download it shortens to 8 secs (see attached images). Are there any ways to completely download the file? Any comment is highly appreciated. Thank you.

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

Re: Seismomatch- Matching of Time History

Post by seismosoft »

Hi rbsce.ph,

It seems that you are reading only one column of the record. Make sure that you use the correct settings when importing it in SeismoMatch.
Seismosoft Support
rbsce.ph
Posts: 30
Joined: 29 Jul 2020, 09:29

Re: Seismomatch- Matching of Time History

Post by rbsce.ph »

Your answer definitely helped. Thank you so much!
rezalatifi
Posts: 13
Joined: 13 Nov 2019, 13:37

Export to Excel

Post by rezalatifi »

Dear,

I would like to Export the spectral matching results from SeismoMatch/SeismoSpec to Excel. Because I want to increase the quality of the output graphs. Please let me know how it is possible to export the output graph's data to Excel?

Sincerely,
Reza Latifi
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Seismomatch- Matching of Time History

Post by seismosoft »

Exporting the time-histories is only available with a valid (commercial or academic) license of the program.
Seismosoft Support
Kirkaguel
Posts: 66
Joined: 31 Dec 2018, 10:04

Re: Seismomatch- Matching of Time History

Post by Kirkaguel »

seismosoft wrote: 16 Sep 2020, 10:17 Hi rbsce.ph,

It seems that you are reading only one column of the record. Make sure that you use the correct settings when importing it in SeismoMatch.
Seismosoft Support
The technical support is right clearly the input was not properly set up to read the exact records. Looking at the graphs It is obvious that the peak ordinate of the acceleration was shifted, the original has peak ordinate at around 3 seconds whereas the imported was shifted to less than 1 sec.

My two cents
Post Reply

Return to “03-Unexpected behaviour/errors”