Error Saving Matched Accelerograms

03-Unexpected behaviour/errors
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Error Saving Matched Accelerograms

Post by huffte »

I cannot save my matched accelerograms. If I try to save all accelerograms I get the following error:

"Range Check Error"

If I try to save them one at a time I get a different error:

"I/O Error 32"

Has ayone seen these before?
Tim Huff
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Error Saving Matched Accelerograms

Post by seismosoft »

Dear Huffte,

Unfortunately, we cannot reproduce the behaviour you describe. Are you using v1.3.0-Build_101?

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

Re: Error Saving Matched Accelerograms

Post by huffte »

Yes, I am using 1.3.0 Build 101. The error does persist. It is not a serious error becasue it is easu enough to copy and paste the mathced accelerograms in to Excel and save them that way. But perhaps it is a processor issue?

For the record, I am running Windows XP Professional Version 2002 Service Pack 3 on a machine with Intel Core2 Quad CPU.
Tim Huff
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Error Saving Matched Accelerograms

Post by seismosoft »

Thanks.

We will try our best to reproduce the error (we have been unsuccessful so far), so that we can fix it in the future release.

Thanks again,

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

Re: Error Saving Matched Accelerograms

Post by huffte »

Another slight concern has arisen. In copying matched accelerograms from SeismoMatch and pasting into Excel, the final line in the time history pasted into Excel is often a spurious line.

For example, in one instance, the final line in SeismoMatch was at t=26.915. This was the next to final line after pasting into Excel - the final line was at t = 53.830, exactly twice what should be the final line. Note that this spurious line does not appear in the SeismoMathc form data.

In a second instance, the final line was at a time equal to exactly one half of the true final line.

The error occurs on approximately 4 of 7 records on average I would say. It does not occur every time.

Her is one specific example:

Final 2 lines in SeismoMatch form:
26.91000000 -0.00100000 0.03864834 1.34152686
26.91500000 -0.00100000 0.03864834 1.34152686

Final two lines in the Excel file:
26.915 -0.001 0.039 1.342
53.830 0.027 -4.910 1.585
Tim Huff
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Error Saving Matched Accelerograms

Post by seismosoft »

Dear huffte,

Thanks for reporting this apparent problem, which we will certainly look into.

Thanks again,

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

Re: Error Saving Matched Accelerograms

Post by seismosoft »

Dear huffte,

We did manage to reproduce the second bug you signalled (last value of a time-series being corrupted when copied to an external application), and will correct it.

For what concerns the first issue (error in saving matched accelerograms), however, we were not able to reproduce it. Perhaps you could kindly send us one of the time-series that you are using and that is leading to this type of behaviour?

Many thanks,

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

Re: Error Saving Matched Accelerograms

Post by huffte »

I am sending 7 time histories. When I do simultaneous matching for all 7 I get the error. When I do "BhujH1" by itself I also get the error. I have not yet tried any of the other files by themselves. The files have been sent to support@seismosoft.com and are:

BhujH1.txt
TangshanH1.txt
SGMSH1.txt
PS65H1.txt
OlympiaH1.txt
NisquallyH1.txt
JiashiH1.txt

I have also included the file which defines the target spectrum:

Site-02-MCE.txt
Tim Huff
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: Error Saving Matched Accelerograms

Post by seismosoft »

Thanks.

We have now obtained the same error as you did, and will be able to correct things for the next release.

Thanks again,

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

Re: Error Saving Matched Accelerograms

Post by seismosoft »

Dear huffte,

We have now been able to understand that the "Range Check" and "I/O 32" errors that you have pertinently reported are caused by the use of very large original accelerograms, that give rise to the assemblage, within SeismoMatch, of equally large matrices and vectors, which in turn may request very large memory from the system (in some cases 1GB or more) and lead to errors of this type.

We will, in the forthcoming version of the program, try to introduce more user-friendly messages to the users, advising the cropping of such large input accelerograms, which typically contain large portions of negligible motion at their start and end, along the lines of the discussion in http://www.seismosoft.com/forum/topic.asp?TOPIC_ID=611 .

Thank you once again for having called our attention to this issue.

Seismosoft Support
Post Reply

Return to “03-Unexpected behaviour/errors”