Step By Step Procedure
for Seismic Modeling
Just as log analysts use core data as ground truth to calibrate their log analysis results, geophysicists use synthetic seismograms created from edited log data as their ground truth.
Structural interpretation of seismic sections involves identification of reflective horizons and picking of seismic travel times for each trace on each horizon. From these data, time, and subsequently depth, maps are made for each horizon. Choosing the exact horizon time is sometimes difficult, since the bed boundary may not be on a peak or a valley of the seismic trace. A synthetic seismogram is often used to calibrate these time picks.
Stratigraphic interpretation is based on the ability to correlate seismic character to subsurface geology, instead of just seismic two-way times. The correlation is established at a well with a synthetic seismogram computed from edited well log data. The shape and amplitude of the reflection on each trace is important in its interpretation.
steps in making a synthetic seismogram are:
Synthetic seismograms derived from unedited and un-modeled data are common commercial products from all vendors. Two typical examples are shown below. They are useful for gross correlation and major reflector identification. Most vendors allow the purchaser to customize the product to include the type of editing and modeling described in the following Sections. However, the responsibility for the model parameters rests with the client as few suppliers have the log analysis skills to prepare adequate log models.
Using the log response equation we can inject new layers, delete existing layers, replace existing layers with new ones, change the fluid content from water to gas or oil (or vice versa). This is called "What-if" modeling. Models of this type will show why some water zones have bright spots, why some coal beds have no reflections, and why some carbonate porosity is visible on seismic (and some is not). What-if models allow the interpreter to test different structural and stratigraphic solutions against the actual data before committing to one interpretation.
Another variant of this modeling method is to replace data, or add data to the bottom of one well, using log data from another well which more closely represents the interpreted seismic section. This data must also be modeled and edited to reflect the unaltered rock/fluid mixture. This is often done to aid in design of a sidetrack or whipstock well to find the fault, reef, or salt dome structure that was missed by the original hole.
far as can be determined, FEW existing seismic modeling workstation
performs the necessary log analysis and log recalculation. Therefore
YOU must do it first, offline from the modeling package, and enter
the edited/modeled logs into the system.
Page Views ---- Since 01 Jan 2015
Copyright 1978 - 2019 E. R. Crain, P.Eng. All Rights Reserved