geopsy.org

Hello Marc.

I'm working with L-shape array for MSPAC tests (ambiental vibration). Array has 16+8 geophones (4.5Hz) spaced to 3m. Recording time is 45 minutes at 8ms. Dispersion diagrams from spac2disp shows modal superposition (see image attached). So, to do inversion process... How can I approach the inversion process at Dinver for these types of curves?

We do three HV tests at this site. All of theme shows two peaks at 5.6Hz and 0.75Hz.

What is your recommendation about this challenge?
Larger arrays? Use arrays with greater geophone spacing?

Hi,

As a first approach, you can manually pick the two portions of dispersion curves on spac2disp plot ([~2,~3] Hz and [~7,~12] Hz). Export the curves to a file and import them into dinver specifying the modes appropriately.

Did you try Capon FK? Are there also two modes? Are they three-component geophones? Three-component FK processing is usually better to catch several modes because they can be separated on the wavenumber and on the ellipticity axes.

Multiple peaks on the H/V can be effectively due to multiple Rayleigh modes (peaks of the ellipticity curve of each mode) but not necessarily. Three-component array processing can deliver a straight forward answer if you have three-component sensors.

If the lowest H/V peak is at 0.75 Hz, you can probably extent the dispersion curve towards lower frequencies by increasing the array aperture. Playing with warangps (tab 'Map') can help you to define the require aperture. Load your current geophone coordinates. You can play with the 'factor' in the bottom right to scale your array. You can add the dispersion curve (layer) found with spac2disp as a background of dispersion plots in warangps (bottom right, kmin and kmax or wavelength).