Seismosignal can not read long-term measurement data.

04-Future developments & requests
Post Reply
llem0
Posts: 4
Joined: 06 Sep 2018, 04:49

Seismosignal can not read long-term measurement data.

Post by llem0 » 09 Jan 2019, 05:59

Dear Seismosignal team.
I would like to ask you for a recommendation for some software that can analyze long-term measurement data. For example at a sampling rate of 512Hz for 8 hours.

Kind regards

User avatar
Fanis_Moschas
Posts: 3
Joined: 21 Aug 2018, 14:27

Re: Seismosignal can not read long-term measurement data.

Post by Fanis_Moschas » 09 Jan 2019, 15:46

Dear llem0,
It would help if you could provide us with more information.

Concerning the length of your time series, what is the length (in time series points) of the maximum time series you are interested in analyzing? Are you interested in analyzing the full 8 hours recording (some millions of points) or is it possible to break it into smaller ones? For Example SeismoSignal is capable of handling time series of a maximum length close to 260000 points.

Also in what type of analysis are you interested in? Time Series analysis (Fourier analysis, filtering etc.) or something different than that?

Thank you in advance for your feedback
Fanis Moschas
Seismosoft Srl.

cibyjac
Posts: 1
Joined: 11 Feb 2016, 09:33

Re: Seismosignal can not read long-term measurement data.

Post by cibyjac » 21 Jan 2019, 05:35

I am using Seismosignal version 2018, Release 2, Build 10, academic license. "For records larger than 32768 lines, the Fourier and Power spectra as well as the mean period Tm, cannot be calculated" error is being shown. Since this limitation is there, for long duration earthquake data, the Fourier analysis is not possible.

huffte
Posts: 832
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: Seismosignal can not read long-term measurement data.

Post by huffte » 21 Jan 2019, 14:01

Accelerograms, from real earthquakes, having fewer than 32,000 time acceleration pairs are abundant. There are, however, records having more than 32,000 points. In such a case it may be possible to load the data and use the 'frequency' value when the record is loaded into SeismoSignal.

For example, if I have a PEER-formatted record (5 acceleration values per line starting with line 5) with 60,000 acceleration values and a time step of 0.005 seconds, I could specify a frequency value equal to 2 (read only every other data point) and a time step equal to 0.010 (2 times 0.005). This would result in a 30,000 data point record. Similarly, I could specify a frequency value equal to 4, a time step equal to 0.020, and end up with a record having 15,000 data points.

Note that this does not alter the data read. It only skips a certain number of points. So, if the integrated velocity and displacement records are still accurate, the effect of skipping data points may be minimal, depending on the desired intensity measures to be studied. When using this approach, care should be taken in using extremely long period intensity measures.
Tim Huff

Post Reply