Principle of Seismic Value Chain

In network organizations, work processes do not follow simple linear paths but they occur in a cyclic manner with feed-forward and feed-back paths. The result is integration across traditional organizational boundaries.

In network organizations, interaction processes in workflows do not follow simple linear paths but they occur in a cyclic manner with feed-forward and feedback paths (Senge, 1990; Berkhout, 2000). In this cyclic concept, the process model consists of coupled loops and the organization that facilitates the information flow in those loops is flat with short interconnections. Cyclic interaction applies to the E&P workflow in the widest sense, but considering the research topics in the Delphi consortium, we will focus on the inner loops of geophysics only. Figure 3.1 shows the Seismic Value Chain (SVC). It visualizes the cyclic interaction processes between the three disciplinary nodes: seismic acquisition, structural imaging and reservoir characterization.

Current research in Delft aims at extending the seismic value chain to the entire upstream workflow (E&P innovative circle), including geological modeling and reservoir management.


Stacks Image 198
Figure 3.1: Cyclic interaction in the Seismic Value Chain (SVC), showing the importance of cross-disciplinary communication between the disciplinary seismic nodes. Bear in mind that the forward and backward information flow between the two inner loops, connecting acquisition and characterization, is describe

In the double-loop seismic value chain the needs in reservoir characterization should drive structural imaging and the needs in structural imaging should drive seismic acquisition (double feedback). Moreover, new capabilities in acquisition should inspire new developments in the imaging node and new capabilities in imaging should inspire new developments in the characterization node (double feed-forward). Hence, both feed-forward and feed-back processes are required, defining the ‘context of discovery’ and the ‘context of justification’ respectively.

Note the important property that interaction is not only first order but also second order. This is the organizational architecture that we have established in the research program of the DELPHI consortium, crossing also organizational boundaries within the Delft University (Berkhout, 2005).

If we look at seismic research today, the following strategic issues can be observed in the different nodes of the Seismic Value Chain:

1. Seismic acquisition node

The largest investments along the seismic value chain occur in data acquisition. Therefore, more knowledge is required how to design economical acquisition geometries (node 1) that potentially yield the best images in terms of spatial resolution and amplitude accuracy (node 2), as well as the best reservoir models that allow accurate estimates of recoverable reserves at any stage of the life cycle (node 3). A fundamental research question is to what extent sparse spatial sampling in acquisition can be compensated for by interpolation beyond aliasing (preprocessing in the acquisition node). Another important issue in node 1 is how preprocessing can properly solve the near surface problem in land and sea-bottom acquisition. Until today, ‘statics’ is still an unsolved problem. Note that if data-driven processing techniques such as wavelet deconvolution, primary-multiple separation, velocity estimation, and full wavefield imaging are going to play a more important role, data acquisition design need to gain significantly more attention.

The philosophy of the DELPHI consortium is: ‘let the data speak’. We believe that too many of today’s solutions are biased by preconceived models, particularly in the second part of the value chain. Economizing on data acquisition may be penny-wise and pound-foolish. Recently, we emphasized the large advantages of blended acquisition, involving the utilization of incoherent wavefields. In addition, we are exploring the potential of robotization in seismic acquisition, being considered the next big step.


2. Structural imaging node

Until today, imaging requires primary reflection data. A lot of progress has been made on the removal of multiple scattering events. This is particularly true for multiples that cannot be easily distinguished from primaries with the aid of move-out. Here, the data driven approach of DELPHI in terms of multi-record convolution (Berkhout, 1982) has shown excellent results in practice (Verschuur, 1991) and is now being used worldwide. 3D versions of the data-driven algorithm (SRME) are commercially available now. The current focus is solving the challenges with respect to utilizing primaries and multiples, both surface and internal.

Similarly, a lot of progress has also been made on the development of seismic migration algorithms. However, the weakest point of today’s migration technology is the estimation of accurate migration velocity models. For instance, imaging below salt is still a largely unsolved problem.
In the DELPHI consortium, full wavefield migration is being developed, utilizing multiple scattering (surface and internal) as well. In addition, we believe that velocities should be derived by the migration process itself (‘joint migration-inversion’). This topic is actively pursued in our research.

3. Reservoir characterization node

In reservoir characterization high-resolution seismic im- ages, angle-dependent reflection information and accurate velocity-density models are needed to determine the volumetric properties of a (potential) field. In addition, using rock phys- ics the internal micro properties of the reservoir need be es- timated in terms of rock composition and pore properties. In reservoir characterization we aim at extracting these me- dium properties in an accurate manner via broadband, high-resolution full waveform inversion: FWI-res. DELPHI’s joint migration-inversion technology (JMI) provides the full wavefields that are needed to apply FWI-res in an efficient way.

The ambition in the characterization node is to make also use of EM measurements and production data. The target is to obtain accurate estimates of saturation and stress by using an integrated estimation approach that combines 4D seismic inversion and production history matching in a geologically meaningful way.
In time-lapse seismic inversion, the double-loop interaction between seismic acquisition and reservoir characterization becomes very clear: repeated acquisition aims at monitoring changes in the reservoir – in terms of changes in saturation and stress – during the production phase. These changes drive the reservoir management decisions.