What lies beneath

Materials World magazine
1 Jul 2018

A new method for processing data from seismic surveys could help give a more detailed view of underground oil reserves, as Ellis Davies reports. 

Oil exploration isn’t easy. Reserves are often miles underground, and give little indication of their presence on the surface. This makes seismic surveys a valuable tool to those seeking oil – one that is now improved. Researchers at King Abdullah University of Science and Technology (KAUST), Saudi Arabia, have developed an advanced computational method for processing seismic data that can observe the detailed structure of deep oil reserves in high resolution, and also provide additional data such as fracture density and orientation.

Seismic surveys involve the use of shock waves directed towards an area of suspected oil reserves, after which the weak sonic vibrations that return to the surface are recorded to give an idea of what lies beneath. The vibrations allow different rock types to be identified as the waves travel at different speeds depending on the material they bounce off or pass through. This reveals basic geological structures.

Based on this, drilling commences, which is very expensive – tens of millions of dollars per drill, according to KAUST – so an unsuccessful drill is a drain on finances. 

A more detailed picture

To improve these surveys, researchers used full waveform inversion (FWI), which goes beyond analyses of the returning signals by adding additional data gained from the amlpitude and phase of the waves. The new method adds further data from prior geological knowledge of the area and drillhole experience, and is able to provide a higher level of structural resolution and provide useful information, such as the presence of potential geohazards, for drilling decisions and horizontal well placement. 

The use of FWI has so far being limited because of challenges relating to a lack of prior knowledge of local rock physics. 

However, parameters such as the anisotropic – different material properties when moved in different directions, for example wood is stronger following the grain than across it – and isotropic properties – same value when measured in different directions, like glass – of local rocks and other Earth materials can be informed by comparing the rock formations with similar examples elsewhere. Data in other areas may have already been collected, making up for the lack of sensitivity sometimes seen when using these parameters. 

As similar rock formations in the Earth have consistent properties and relative elastic and anisotropic parameters, they can be defined as mapable seismic units. This means that they can be used in FWI to achieve a higher resolution structure of another area.

Researchers took this approach, changing the reflection data into a description of the reservoir using the Bayesian theory – the probability of an event based on prior knowledge of conditions. Results showed that the method could improve the resolution. Four examples were used to verify the effectiveness of the method, each using a different set of assumptions of rock physics and model circumstances.

‘The key idea is a more complex physics description of the reservoir region with parameterisation,’ stated Tariq Alkhalifah, head of the team at KAUST, in a university release. 

Doctoral student Zhen-dong Zhang, added, ‘Although our method is computationally heavy and puts higher quality requirements on the seismic data, there is a lot of interest in this hot topic within the oil exploration and production communities,’ he said. ‘With the pace of advancement in computing power, we are well placed to benefit from the expected wave of interest in methods that can provide more accurate descriptions, particularly for fractured reservoirs.’