PUBLICATIONS

Finding "Rosetta Stone"
for Marine Seismic Epoch

Nick Amelin, Aleksandr Nikitin and Sergei Pokrovskii

An Urge to Improve

Seismic velocity is a key to clear image and reliable time to depth conversion. It is our “Rosetta Stone” of understanding and interpreting seismic data for hydrocarbon exploration.

Properly interpreted seismic data is vital for making drilling decisions and an accurate estimation of reserves. The standard complaint from O&G companies’ managers is about interpreters coming up with unreliable predictions of trap locations and thickness of hydrocarbon bearing rocks, particularly close to salt bodies. As a result, we have unsuccessful drilling in the majority of cases.

Global marine seismic acquisition market is estimated at 2,8 bln USD in 2017 and is expected to recover to 5,1 bln USD by 2019 as companies have to maintain hydrocarbon total reserves which have been affected by collapsed oil market. Now companies are looking for inexpensive data acquisition techniques and in a low oil price environment they seek to limit exploration expenditure and retain profitability.

GWL FloatSeis™ is an acquisition technique that allows to build high-resolution and robust velocity models for depth up to 10 km and is a cost-effective solution. FloatSeis™ is applicable both to the existing marine seismic data sets to improve it by FWI based re-processing and new seismic projects aimed to considerably improve the output of the surveys.

GWL FloatSeisTM is the most cost-effective way to record data perfectly suitable for FWI application.

GWL FloatSeis data contains both ultra-long offsets and low frequency component that are required for a robust FWI algorithm output

Convenient Industry Solution

In 90% of cases companies use 2D/3D CDP marine seismic acquisition to reveal new prospective areas as well as to improve understanding of existing reserves.

CDP approach gives structural image, but in the time domain. A robust velocity model is required to scale it to depth that CDP approach, unfortunately, does not provide.

Based on overall experience, CDP velocities all within ±15-20% of the true velocities. In complex geological environments (salts, basalts, irregular bedding, high-velocity formations and so called “gas chimneys”) CDP velocity field tends to be even less reliable. CDP velocity analysis suffers from multiples, diffractions and scattering which is treated as clutter. Elaborate processing algorithms are being developed to reduce influence of these factors on CDP velocities but we still have just an approximate velocity field.

In 90% of cases companies use 2D/3D CDP marine seismic acquisition to reveal new prospective areas as well as to improve understanding of existing reserves.

CDP approach gives structural image, but in the time domain. A robust velocity model is 

CDP and VSP Velocity Comparison​

A good illustration of the rule of a thumb - CDP velocity tends to be higher and doesn't match with the well data (The Barents Sea, Offshore Norway)

required to scale it to depth that CDP approach, unfortunately, does not provide.

CDP approach gives structural image, but in the time domain. A robust velocity model is required to scale it to depth that CDP approach, unfortunately, does not provide.

Based on overall experience, CDP velocities all within ±15-20% of the true velocities. In complex geological environments (salts, basalts, irregular bedding, high-velocity formations and so called “gas chimneys”) CDP velocity field tends to be even less reliable. CDP velocity analysis suffers from multiples, diffractions and scattering which is treated as clutter. Elaborate processing algorithms are being developed to reduce influence of these factors on CDP velocities but we still have just an approximate velocity field.

To a certain extent well ties or VSP can be used to retrieve velocity in some sparse points but in other cases we only interpolate, extrapolate or do nothing. It is not a robust and a very controversial method of getting depth conversions. Especially, when it comes to offshore areas where only a few wells are available.

CDP method drives standards for seismic equipment. Streamer length is limited by 12 km and seismic source usually has a 10-70Hz bandwidth with a dominant frequency of about 30-40 Hz. Almost every seismic equipment and processing software are produced for the needs of CDP seismic exploration. That fact leaves very little room for application of other methods different from CDP acquisition if they require acquisition parameters other than CDP. As a result, it has an influence on the development of the new acquisition techniques and equipment.

To break out of the vicious circle GWL has developed and produced seismic equipment able of recording much longer offsets and operating within the frequencies lower than the standard marine seismic source can emit.

A New Promise!

Full Waveform Inversion (FWI) is a game changer approach for velocity model building. It uses two-way wave equation and performs forward modeling to compute the difference between the acquired seismic data and the current model. As a result, high-fidelity velocity model that misses all the drawbacks typical for CDP velocity analysis and traveltime tomography is derived.

Full Waveform Inversion (FWI) has potential to become a key tool to interpret seismic data acquired in complex geological settings and grow into a new standard for velocity model building.

In addition to seismic information, joint FWI allows to involve other types of data, such as gravity and well logs for better stability of the inversion. FWI output can be a reliable ground to recover a range of strata’s parameters like Vs, density, anisotropy, absorption coefficient. Algorithm abilities only determined by density and diversity of input geophysical data for an unambiguous solution.​

FloatSeis FWI Velocity Model - 16 Hz​

FWI  velocity model  obtained with  help of FloatSeis  long-offset data  is well  resolved and  can be used for an initial geological  interpretation itself. The  higth velocity carbonate layer  is  prominent  on  the  velocity  secretion with the available well data. itself  (in  blue)  and  decently  corresponds 

Companies actively applying FWI approach:

  • Schlumberger;

  • CGG;

  • BP;

  • DUG;

  • PGS;

  • etc.

According to the last publications and industry major events, FWI is booming, but a few factors still deter its broad application.

Limits on Offset Length

FWI reveals its potential by using both wave fields: reflection and refraction. Refracted waves carry a considerable amount of information about the physical properties of the medium that other types of waves do not have. Thus, using only reflected waves for the FWI imposes the same limitations on the result as the CDP velocity analysis does.

Previously, refracted waves were considered mainly as a tool for the upper part of the section and statics corrections and usually were muted before CDP velocity analysis. FWI brings reflection and refraction data together into a single model.

Example of FloatSeis Survey.

GWL Seismobuoy Super-Long Offset Common Receiver Point Gather

1 - Offsets recorded during long-offset seismic streamer acquisitions;
2 - Amount of usefull data that is missing on the seismic streamer records and can be easily recorded by GWL Seismobuoy

The existing stipulation is that the lack of long offset data can negatively influence FWI in terms of illuminating more subsurface angles and deeper sections.

Refraction waves for application in FWI require recording arrays with lengths of approximately 5 to 6 times the depth of imaging (in complex geological environments it can bounce up to 10-15 times). Thus, to be able to utilize FWI for up to 7 km depths, efficient recording of 40-60 km offsets is needed.

Present-day marine seismic equipment offers a 12 km streamer length, and longer offset recording is difficult and expensive. This in turn, limits the FWI depths up to 2 km, that is insufficient for commercial applications.

Ocean-bottom seismic recorders or multi-vessel shooting is quite rarely used because of its high costs and could not be considered as a meaningful substitution. Evidently, FWI requires a fundamentally new technique for seismic data acquisition.

GWL Seismobuoy™ is designed to record high quality ultra-long offset seismic data at up to 120 km offsets. Real-time GPS positioning and online QC gives it substantial advantage in comparison to OBS/OBN surveys toward such applications. Operational costs of FloatSeis are drastically lower than traditional techniques, especially in deep water. Actually, this is the only instrument that allows the deployment of dense arrays to record seismic data with required offsets and high spatial resolution and is able to maintain the same speed of data acquisition as the towed streamer seismic.

The GWL equipment has already been commercially produced and is ready for use on commercial surveys. GWL FloatSeis™ equipment consists of GWL Seismobuoy™ units, control desk, navigation and QC systems. All equipment has been tested and certified.

Low Frequency Component

The need to use a low frequency component of the seismic signal is dictated by the following factors:

  1. Increase in the signal-to-noise ratio at ultra-long distances. The lower dominant frequency significantly reduces the signal attenuation and increases the penetrating abilities keeping the same total volume of the source. It is problematic to increase the volume of standard air-gun sources (over 3,500 cu.in) due to toughening environmental constraints.

  2. Full Waveform Inversion (FWI) application requires low frequency components to ensure a robust solution. Conventional seismic records generally lack low frequencies below 10 Hz that need to be present to form a low frequency background model. Missing low frequencies will affect the:

  • Information in the data to lead it to a minimizer that geologically makes sense; 

  • Parts of the model (loss in resolution);

  • Sustained depth structures (loss in the total model depth)

The lack of low frequencies may be resolved with either well log information (dense grid of wells with a proper log data is required, that is not available for the majority of offshore areas) or by low frequency spectra emission during seismic data acquisitions (special low frequency source is required).

Alternative Low-Frequency source is required to emit Low-Frequency spectra beneficial for Full Wave Inversion.

List of companies developing prototypes of LF source:

  • BP;

  • PGS;

  • Seismic Source Company;

  • Sinopec;

  • etc.

Major challenges in Low-Frequency Source are its durability and signal stability. Source maintenance interval should be at least 50,000 shots for efficient use in seismic operations.

GWL LF Source™ is a combination of the best technical solutions and modern materials & components giving a stable low frequency signal with no compromise to production.

GWL offers a complete source vessel equipped with navigation (synchronizing with seismic vessel if dual shooting is required), source controller, deployment system and source array.

Conclusion

We see the opportunity to take up a new Blue Ocean in a matured Seismic Sea opening right now.

For superior implementation of FWI you are only limited by:

  • Quality of seismic impulse;

  • Presence of low frequencies;

  • Presence of long offsets being enough to image target horizons.

The combination of GWL Seismobouy™, GWL LF Source™ and FWI Approach provides an efficient operational and budget solution ready to commercial use right now. GWL has all equipment required for FloatSeis data acquisition, crew and processing facilities to respond to a call for any seismic project worldwide.

There is a lot of seismic 2D/3D data already acquired that cannot be adequately processed and interpreted because the standard CDP velocity analysis fails. Performing a FloatSeis™ survey over the existing prospects is a key to significantly improve seismic image by reprocessing the data with a FWI built velocity model.

Sub-salt Imaging. PSDM Based on CDP Velocity Model

Sub-salt Imaging. PSDM Based on FWI Velocity Model

Seismic streamer data used for CDP velocity model building is not able to provide us with a proper image of the sub-salt formations (on the left). Ultra-long offset low frequency data used for FWI velocity model building routine helped to reconstruct the bottom of the salt body and gave us the clear image of sub-salt layers (on the right)​

For past proprietary datasets now considered junk, FloatSeis allows capitalizing on exploration investments. For multiclient data, especially over geologically complicated areas, such improvement could jump-start a new spiral of MC data licensing.

FloatSeis™ offers great operational economics compared to alternatives. FloatSeis™ adds up just 15-25% extra costs to a 2D survey and less than 10-15% on 3D projects. We believe that FloatSeis™ has an attractive value-for-money ratio and has a great potential for commercial application.

 

For more details, please contact amelin@gwl-geo.com