PGS Ultima addresses the long-standing challenge of building full-bandwidth earth models and in a recent First Break article, PGS authors show data examples from the Gulf of Mexico and Brazil to illustrate the effectiveness of this new tool.
PGS Ultima is equivalent to performing full waveform inversion (FWI) and least-squares migration (LSM) simultaneously, using raw seismic data, and modeling the full acoustic wavefield.
The main advantage of PGS Ultima is that an accurate model and reflectivity can be derived faster, as the simultaneous inversion resolves both using the same raw seismic data. PGS can bypass most processing steps in a seismic sequence because we use raw seismic data, and the evolution of the products support each other. Additionally, there is no leakage between parameters, so both products of the simultaneous inversion are not contaminated by each other, unlike many current industry approaches.
Yang explains in the First Break article 'With a single modeling engine, parameterized in terms of velocity and vector reflectivity, the two earth properties are iteratively updated using their appropriate sensitivity kernels based on inverse scattering theory. The vector reflectivity is estimated as a data domain least-squares migration and is the key for updating the velocity model beyond the maximum penetration depth of refracted and diving wave energy.'
He adds 'With field examples, we demonstrate how the new solution for vector-reflectivity modeling combined with proper inversion kernels enables us to address the long-standing challenge of building full-bandwidth earth models.'
Synthetic Tests Validate Simultaneous Inversion Approach
A modified version of the SEG/EAGE overthrust model is used to validate the simultaneous inversion, PGS Ultima. The true velocity, density, and reflectivity are given in A, B, and D below. The initial velocity model is given in C and the output from PGS Ultima is given in E and F. PGS Ultima has correctly resolved the velocity model and correctly positioned the reflectors and estimated the reflectivity and the output compares very well with the true examples in A and D.
SEG/EAGE overthrust model synthetic test example. True velocity (A) and density (B) models. Initial velocity model (C) and true vertical reflectivity (D). Inverted vertical reflectivity (E) and velocity from the simultaneous inversion (F).
The performance of the new simultaneous inversion is illustrated in two field data examples below. In both examples a simple initial velocity model is used and zero reflectivity. There is minimal pre-processing.
Gulf of Mexico Case Study | Delivering Accurate Velocity Models and Reliable Amplitudes
The first example is from the deep-water Gulf of Mexico, De Soto Canyon area. The GeoStreamer data was acquired with a maximum offset of 12 km. All the recorded data, i.e. the full wavefield, were used for the inversion.
In the example below, A shows the initial velocity model and C displays the reflectivity from the first iteration of the inversion, which is equivalent to an RTM image using the starting velocity model. Since the data contained multiples, crosstalk artifacts are observed in the RTM image and are indicated by the yellow oval.
The results after several iterations of simultaneous inversion are shown in B and D. The PGS Ultima models clearly display higher resolution and a reduction of the crosstalk in the final reflectivity model.
De Soto Canyon field data example. Initial (A) and inverted velocity models (B). Vertical reflectivity from the first iteration (C) and the final inverted vertical reflectivity (D). Notice the crosstalk reduction as the yellow oval indicates.
Campos Case Study | PGS Ultima Resolves Steep Dips and Accurately Images Target Area
The second field example is from a deepwater environment in the Campos Basin, offshore Brazil. Although the maximum inline offset in the survey is 10 km, the water column of more than 3 km makes it challenging to update the deep targets using refracted energy. The simultaneous inversion was applied to the total pressure data using full shot records (i.e., no event selection). The slider images below show the vertical reflectivity models corresponding to the first and final iteration of the inversion. The velocity updates extend beyond the maximum penetration depth of the diving waves.
It is important to remember that there is minimal pre-processing applied prior to inversion which significantly improves the time needed to obtain this result compared to traditional workflows.
The amplitudes are much better balanced across the entire section on the PGS Ultima image. This is not only due to an improved velocity model but also the inverted reflectivity compensates for any incomplete data missing from the acquisition scheme and geometric lensing effects causing variable illumination in the subsurface.
There is improvement in the resolution of the shallow fault system and a coherency enhancement in the deep reflectors in the mini-basin and the steep salt flanks. The uppermost arrow indicates a thinly bedded sedimentary unit, truncated by faulting. The middle arrows on the far left and far right highlight steeply dipping salt flanks and the lowermost arrows are located closer to the target interval where carbonates are present above and below the salt.
Campos Basin image comparison. Vertical reflectivity from the first iteration (left) compared with the final inverted vertical reflectivity from PGS Ultima. The top arrow indicates a thinly bedded sedimentary unit, truncated by faulting. The middle arrows on the far left and far right highlight steeply dipping salt flanks. The lower arrows are located closer to the target interval where carbonates are present above and below the salt.
The final example below is a depth slice at 3.4 km through the Campos dataset. On the left side, the initial migration is overlain by the starting velocity model. This model is relatively laterally invariant except for the salt responses, showing as the red color. On the right side, the inverted reflectivity is overlain by the inverted velocity model from PGS Ultima. Looking at the results from PGS Ultima on the right, it is apparent that the inverted velocity model contains significantly more resolution and conforms well to reflectivity without reflectivity being imposed.
The geometry of the salt bodies in the center and on the right has been updated compared to the starting model and the underlying inverted reflectivity is both higher-resolution and spatially better balanced in terms of amplitude.
PGS Ultima Velocity model appears consistent with reflectivity.
Accurately Position Targets with PGS Ultima
PGS believe it is possible to reduce project turnaround by at least 50% using PGS Ultima.
The results demonstrate that while the velocity model is iteratively updated, an accurate estimate of the earth's reflectivity is simultaneously generated. FWI and LSM can be performed jointly as a single inversion workflow using minimally processed data. The simultaneous inversion reduces the turnaround time for model building and imaging projects and eliminates the need for time-consuming manual interpretation, especially in complex geological contexts. PGS Ultima solves these challenges by transforming traditional processing and imaging workflows into a data-driven approach that delivers accurate velocity models and reflectivity using simultaneous inversion.