baseline correction

02-Analytical capabilities
Post Reply
liu shuai
Posts: 6
Joined: 16 Mar 2014, 23:26

baseline correction

Post by liu shuai »

1
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: baseline correction

Post by huffte »

I'm not sure what you mean liu shuai. Can you elaborate a bit? Are you doing a real accelerogram adjustment in SeismoArtif? If so, then perhaps you have the "Perform Baseline Correction" box checked in the Accelerogram Generation section of the Settings dialogue? In any event, the displacement history you observe in SeismoArtif has been adjusted from the specified real accelerogram input and would not be identical to the original accelerogram displacement history, baseline corrected or not.

If I have misunderstood and you wish to both elaborate and send the suspect time history to me, I'll be glad to take a look at it liu shuai.
Tim Huff
liu shuai
Posts: 6
Joined: 16 Mar 2014, 23:26

Re: baseline correction

Post by liu shuai »

Thank you for your reply, huffte, maybe my words make you confused, I will try to explain the question again.
I had used the method of ¡°Artificial Accelerogram Generation¡± in the software 'SeismoArtif' to generate a artificial earthquake accelerogram (be named as AWE1), and had used the 'Quadratic' to perform baseline correction, the displacement time history curve of the acceleration AWE1 (I get from the 'SeismoArtif') is not drift ,the question is that, when I input the datum of the acceleration AWE1(I get from the 'SeismoArtif') into the software 'Seismosignal', I found out that the displacement time history curve of the acceleration AWE1 is drift in the software 'Seismosignal', I want to know why the same acceleration AWE1 has two different displacement time history curve, one is not drift,but the other is drift?
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: baseline correction

Post by huffte »

Hi liu shuai. I have tried to reproduce the error by following your steps but have been unable to do so.

I wonder if it could simply be an issue of round-off. Do you have the number of decimal places set to something larger than the default in the programs? I use 10 decimal places and it can make a difference when doing the double integration to get displacement from acceleration. I would recommend this higher value for both SeismoArtif and SeismoSignal.

Also, double check to see that your time step is correct in Seismosignal when you input the SeismoArtif accelerogram. It is easy to mistakenly accept the default always when the true time step may be different. And make sure you read in all the lines. If there are more lines in your input accelerogram, then, once again, accepting the default last line in SeismoSignal would produce inaccuracies.

The only other thing I could think of was maybe you mistakenly read a displacement or velocity history into SeismoSignal? Note that it only accepts an acceleration history as input.

I'll be curious to see the outcome of this, liu shuai.
Tim Huff
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: baseline correction

Post by seismosoft »

Hi Lui Shuai,

We guess that with drift you mean permanent deformation. Apparently, in the case of permanent deformation baseline correction has not been performed, whereas in the case of no permanent deformation baseline correction has been carried out. Please, refer to the help of the programs on how to perform baseline correction

SeismoSoft Support
liu shuai
Posts: 6
Joined: 16 Mar 2014, 23:26

Re: baseline correction

Post by liu shuai »

I had tried again considering your suggestions (¡°issue of round-off¡±,¡± time step¡±,¡± acceleration history as input¡±), but the result still be the same, the displacement history is still drift toward one side in the software ¡°SeismoArtif¡±.
Did you use the method of ¡°Artificial Accelerogram Generation¡± in the software ¡°SeismoArtif¡± to generate a artificial earthquake accelerogram?,
Can you give me your email, so I can send you the details of my steps.
Thank you very much for your reply again.
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: baseline correction

Post by huffte »

You should be able to send me an e-mail by clicking on my username liu shuai. I'll be happy to look over your steps to see if I can reproduce your error. For your information, here is what I did:

Step 1. Start SeismoArtif and generate an artificial accelerogram using the "Artificial Accelerogram Generation" option.

Step 2. Define a target spectrum.

Step 3. Define an exponential envelope function and a duration of 30 seconds. These are somewhat arbitrary for pinpointing your issue and could be changed. I just picked them randomly.

Step 4. Make certain that the "Perform Baseline Correction" box is checked in the "Accelerogram Generation" tab of the "Settings".

Step 5. Generate the artificial accelerogram.

Step 6. Observe the resulting time histories of acceleration, velocity, and displacement. The displacement and velocity histories do, in fact, return to zero at the end of the 30 second record.

Step 7. Save the acceleration history using the "File" -> "Save Time Histories" dialogue and selecting only the acceleration and time boxes.

Step 8. Start SeismoSignal and open the file which had been saved in the previous step. Don't forget to adjust the time step and the last line as needed for the file you are trying to open.

The velocity and displacement histories are identical to those from SeismoArtif and both do, in fact, return to zero at the end of the 30 seconds.

So, liu shuai, I still have not been able to reproduce the error of having SeismoSignal read a SeismoArtif-generated accelerogram improperly.

I'll await your e-mail.
Tim Huff
huffte
Posts: 978
Joined: 22 Jul 2011, 10:19
Location: Cookeville, Tennessee, USA
Contact:

Re: baseline correction

Post by huffte »

I tried that too liu shuai. But I still get the velocity and displacement to return to zero when I have baseline correction applied in the settings, regardless of whether I use a time step of 0.01 or 0.03 seconds. At least for the parameters I selected.

I can see how too large a time step could create problems. So it's a good idea to find a time step which both makes processing the data manageable and produces accurate results (don't use a time step of 0.005 seconds when 0.010 seconds is adequate).

So, since I cannot reproduce any of the errors, liu shuai, I think you have the best solution to your problem - tweak the time step as needed. This will likely depend on the duration for the envelope function and the time history generation you select.

Best of luck, liu shuai.
Tim Huff
liu shuai
Posts: 6
Joined: 16 Mar 2014, 23:26

Re: baseline correction

Post by liu shuai »

Thank you for your careful response,huffte
Post Reply

Return to “02-Analytical capabilities”