460 Posts in 227 Topics by 67 members

Imaging

Forum » Imaging » PSTM

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: PSTM 3386 Views

PSTM Link to this post

Please share the Pstm data prep and pstm job deck for land data.

can we run pstm with different sort of data other than CDP,CDPTRACE.

How we improve the Pstm results. please share the suggestion to improve the PSTM results.

Re: PSTM Link to this post

Masood,

For 2D and indeed 3D land data it is worth remembering that Pre-Stack time migration is not always the best option, due to the sparse and irregular offset distribution that is typical of land acquisition you will often get a better image using a more classical DMO and post stack time migration approach than by utilising pre-stack migrations.

This also makes a generic data preparation workflow for Land preSTM difficult to define as it is likely that each dataset would present unique challenges.

In terms of processes that would typically be applied, in Claritas for 2D PreSTM (KPRET2D module) the TDERIV module should always be applied in the data prep flow, in addition to this the likes of FDFILT/BUTTERFILT and the typical noise attenuation processes such as FXDECON maybe useful.

The Claritas KPRET2D module can accept data in CDP/CDPTRACE, CDP/OFFSET, SHOTID/CHANNEL etc order quite happily, there maybe some system overhead associated with ordering data from different primary/secondary key order but it would be quite small in terms of the overall processing flow.

Key parameters in KPRET2D are the ANGLE/RANGE/STRETCH which control the shape of the migration operator and these should be tested to optimise the migration result on any given dataset. Also as KPRET2D does not have a parameter to protect the near offset there is the possibility with Land or very shallow water marine datasets that near surface information maybe lost as a part of the migration process.

I have attached an example KPRET2D processing flow, which is quite simple as the module handles both input and output so the flow consists of just a single module.

Data output from KPRET2D is NMO corrected, so before reviewing the velocities in CVA you would need to unNMO the data.

Re: PSTM Link to this post

what is procedure do define the OFFMIN, OFFINC and NOFFSETS, for eaxmple i have data with offset rage -3100:3100, group interval 50m, and Source interval is 100m.

how i set these above metioned parameters in module. Please explain briefly.

if we apply interpolation between traces, run PSTM, is this possible to get good result in shallow data?

Re: PSTM Link to this post

The OFFMIN, OFFINC and NOFFSETS parameters define the Minimum offset to create in the output gather, increment between offsets in the gather and number of offsets to create. After migration there is no real necessity to produce split-spread CDP's so typical parameterisation would be OFFMIN=100 OFFINC=100 NOFFSETS=31.

Interpolation will have limited effect on PSTM in the shallow, as the muting effect is controlled by the migration operator which is in itself controlled by the DIP/RANGE/STRETCH parameter and the velocity field, it would probably be more effective to tune these parameters and condition the velocity field. Also with the interpolation you would need to manually set and update the SHOT_XY, REC_XY, CDP_XY and OFFSET headers which would be some very intensive header manipulation.

Re: PSTM Link to this post

is there any module present to reduce the random noise from stack?

Re: PSTM Link to this post

For random noise attenuation there are a number of options available in Claritas , these include in no particular order:-

FXDECON - FX Domain 2D random noise attenuation, you can iterate the use of this module to maximise its effect.
RUNMIX - Coherency filter - using a spatial median filter approach.
FXRUNMIX - Coherency filter - FX Domain spatial median filter.
SEMBSMOOTH - Coherency filtering technique.

    3386 Views
Go to Top