BroadSeis™ extends the bandwidth to both high and low frequencies to deliver over 6 octaves of signal, with the best signal-to-noise ratio at low frequency. BroadSeis has been proven to be AVO compliant on both synthetic and real data, incorporating fully amplitude preserving deghosting algorithms to deliver true-amplitude, high-resolution subsurface images, reliable AVO attributes and stable, quantitative seismic inversion. As a consequence of the proprietary curved streamer configuration, BroadSeis is able to deliver the deepest tow in the industry, up to 50m to record low-noise ultra-low frequencies down to 2Hz. When combined with full source characterization and accurate 3D designature and debubbling, BroadSeis delivers stable low frequencies.
Why does your seismic inversion work so well? Because low frequencies in CGG broadband are so good!"
A major international oil company at Devex 2015
The lower frequencies provided by BroadSeis mean that a low frequency model derived from well logs is not required
The broad bandwidth and ultra-low frequencies achieved using BroadSeis acquisition and processing, translate into more accurate and quantitative seismic inversion. The lack of low frequencies in conventional seismic data means that a low frequency model must be incorporated into the inversion process. With BroadSeis data, high-resolution NMO-derived seismic velocities are used to define the low frequency model in the 0-5 Hz range, while the reflectivity provides information from 2.5 Hz. BroadSeis low frequencies are also ideal for FWI (Full Waveform Inversion) which yields extremely detailed velocity models and so raises the frequency provided by the seismic velocities, and reduces the gap between reflectivity and velocity from the low end. This means that quantitative inversion can be performed, even in areas without existing well information.
In order to quantify the impact of the low frequency content on seismic inversion, the results of 3D acoustic inversion over a test area of around 1330 km2 over a Brazilian offshore field were compared for BroadSeis and conventional flat streamer data. The first target was the Eocene reservoir and the second the pre-salt carbonates. The objective was to see the difference in acoustic inversion results for conventional and BroadSeis seismic datasets, quantified by a blind well test, and to delineate the Atlanta reservoir. The test was executed with effort to synchronize all parameters of both inversions and see the result caused by the differences in the input data. The same initial model was used for both inversions and was created based on the logs from Well X with Vp (P-wave velocity) and density values being interpolated along basic horizons. A low-pass filter of 2-5Hz was applied to the model. All inversion parameters are based on Well X and Well Y was used as a blind test.
The most obvious difference between them is that the broadband result shows much sharper borders of the pay interval of the reservoir (indicated by blue low impedance zones). The same effect can be seen on the vertical sections. The BroadSeis inversion defines the geometry of the reservoir better and decreases the uncertainty in calculating the reservoir volume. There is also better definition of geological bodies separated by faults, even where a fault is not included in the initial model, making interpretation easier and more reliable.
The impedance from the BroadSeis inversion gives a better match to the “blind” well Y (below) – the inversion values are closer to the real log values because of the presence of low frequencies.
In this example from northwest Australia, the inversion of BroadSeis data shows greater dynamic range and sharper boundaries than that from conventional data. There is an excellent tie with the well logs, which were not used in the inversion process. Both the conventional and BroadSeis inversions started from the same initial model.
The elastic parameters derived from the seismic inversion were used to predict the lithology using LithoSI. In this technique, a supervised Bayesian classification is performed to produce probability cubes of the predicted lithology or rock properties.
The gas sand prediction from the BroadSeis data matches the well data much better than the conventional data. This indicates a considerable reduction in the uncertainty of the lithology prediction, which is also shown in the cross-section below, where the BroadSeis data gas prediction matches the well logs with a high degree of probability. Note that the BroadSeis data also indicates a high probability of gas sands in a discrete volume between wells A and B, which is not apparent on the conventional
35,000km2 Cornerstone contiguous broadband PSDM data with added-value Facies Finder and PPP products to help pinpoint the best prospects
Contact us by opening an online inquiry.
EITHER... search our portfolio of innovative solutions:
OR... browse our alphabetical directory of geoscience products and services:
Explore CGG's impressive library of journal and magazine articles.
View CGG's extensive portfolio of products and solutions.