Scaling factor significant figures

02-Analytical capabilities
Post Reply
jcarter40
Posts: 2
Joined: 26 May 2020, 02:42

Scaling factor significant figures

Post by jcarter40 »

I am trying to process some Japanese earthquake records from K-Net that are saved in an unusual scale for acceleration. I need to convert the records to acceleration in gravity. I am running Seismomatch 2016. The program only allows 4 decimal places for the scaling factor resulting in a rounding error that throws off the velocity and displacement plots. Is there a fix to allow a scaling factor with additional digits beyond 4 decimal places?
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: Scaling factor significant figures

Post by huffte »

One idea: load the accelerogram with a specified scale factor of 1.0. Set to 9 decimal places in the settings for subsequent loading of a revised accelerogram. Copy and paste the ensuing table of acceleration versus time into Excel and scale the record in Excel before saving as a new text file which can be opened in SeismoMatch. A tedious solution, I know, but it should work. Best of luck with your work, jcarter40.
Tim Huff
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Scaling factor significant figures

Post by seismosoft »

In more recent versions of the program the program allows for more than 4 decimal points for the scaling factor.
Seismosoft Support
jcarter40
Posts: 2
Joined: 26 May 2020, 02:42

Re: Scaling factor significant figures

Post by jcarter40 »

OK, I have installed SeismoMatch 2020. The program allows me to put in a longer scaling factor, but now I'm getting an "Out of Memory" message every time I try to import a earthquake record. I'm running the program on a machine with 8GB of RAM. Any thoughts?
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: Scaling factor significant figures

Post by huffte »

Many Japanese records are very large, with 60,000 or more data pairs. One option is to specify a frequency of 2 and double the specified time step to read only every other value instead of every value, essentially cutting the record size in half. I have used this before several times. It will be important to examine the record with alternate points read in order to make certain it is still valid with regard to velocity and displacement drift at the end of the record.

I have actually even used a frequency of 4 and quadrupled the time step and obtained good results for large records with a very low time step. You will also need to make certain that the new time step is ok for your analysis.
Tim Huff
Post Reply

Return to “02-Analytical capabilities”