Discrepancy Between Input Motion and Output Response at the Absorbing Boundary Condition

[Zouatine]

Dear Stko Team,

I would like to kindly ask for your insight regarding the following problem:

I have a simple 2D model of a single soil layer, modeled as elastic. Both the soil and the absorbing boundaries have the same material characteristics. The model runs fine. I used the same input motion presented in your webinar (a Ricker wavelet) and compared my input velocity to the velocity recorded at the bottom as part of the model verification (they should be the same). This turned out to be true (Figure Below) ( there a bit of shifting i guess do the stage earlier) .

Input_Recorded_Velocity.png (32.93 KiB) Viewed 254 times

Then I changed my input velocity to something more representative of an earthquake acceleration record, as shown in the figure. I compared the input to the output as before, and at this point I started observing discrepancies, as you can see in the figure below.

Input_Recorded_Velocity_earthquake.png (53.97 KiB) Viewed 254 times

What I tried :

I experimented with the mesh size (increasing and decreasing it), but this did not have a major effect on the results. However, I wondered whether wave reflections might be occurring—that is, the wave reaches the surface, reflects back to the bottom, and gets added to the signal. To check this, I reduced the wave velocity of the layer so that the waves travel more slowly. I obtained the following results:

Input_Recorded_Velocity_earthquake_high_velocity.png (83.98 KiB) Viewed 254 times

As we can see, there is a fairly good match in the first part (better than in the earlier figure). However, it seems the problem is not solved—the input and the recorded signal should be the same. The recorded signal appears to be delayed, and higher velocities are observed later. Could you please share your thoughts on this issue?

Please find the model in the figure below :

Simple_2D_One_Layer_Soil.zip

(174.98 KiB) Downloaded 47 times

Thank you and I am looking forward to your insights.

Zouatine,

[STKO Team]

You got the point. When you input a motion at the bottom, it will reflect at the top and you will see it back at the bottom just before it gets absorbed