457 Posts in 225 Topics by 67 members

Demultiple

Forum » Demultiple » Adaptive subtraction of frequency split multiple model

If this is your first visit, you will need to register before you can post. However, you can browse all messages below.

Moderators: Andy Juniper , Guy Maslen , Keleigh Jones

Page: 1 Go to End
Author Topic: Adaptive subtraction of frequency split multiple model 1750 Views

Adaptive subtraction of frequency split multiple model Link to this post

An approach that splits the multiple model into different frequency bands and adaptively subtracts these separately can prove to be a most effective way of attenuating multiple.
The attached example flow (Freq_Split_Adaptsub_example.job) :
•   Reads in data.
•   Mutes any water column noise by applying a manually picked water bottom mute.
•   Stores the multiple model in a pseudotrace (SRME_MODEL).
•   Saves a copy of the data to a pseudotrace (TRACE_COPY).
•   Replaces the seismic trace with the multiple model (stored in the pseudotrace SRME_MODEL).
•   Stores the position of the first non-zero sample to the DELAY header.
•   Applies a band pass filter.
•   Mutes to the position of the first non-zero sample stored previously in the DELAY header.
•   Stores the filtered multiple model in another pseudotrace (e.g. SRME_MODEL2) .
•   Replaces the seismic trace with the original multiple model (stored in the pseudotrace SRME_MODEL) .
•   Repeats the previous four steps producing several different filtered multiple models all stored in separate pseudotraces.
•   Restores the copy of the data (stored in the TRACE_COPY pseudotrace) to the seismic trace.
•   Mutes to the position of the first non-zero sample stored previously in the DELAY header. As this should be the waterbottom multiple this step ensures that the adaptive subtraction will only occur where there is multiple and not effect the shallow primary seismic.
•   Uses several MONKSUBT modules to adaptively subtract the different filtered multiple models stored in the pseudotraces, from the data read into the jobflow.
•   Merges the original input data stored in the TRACE_COPY psudeotrace with the seismic trace. The merge will follow the values stored in the DELAY header. In this way data above the multiple will not be effected by the adaptive subtraction.
•   Deletes the headers containing the stored pseudotraces, as these are no longer needed and will just make the output file larger than it needs to be.
•   Outputs shot gathers.

Attached Files
    1750 Views
Go to Top