Inversion of the seismic data means extracting layer properties from reflectivity, which involves removing the effect of the seismic wavelet from the data, as well as calculating seismic layer properties (e.g. impedances, Vp/Vs ratio, density) from the interface reflections which depend on contrast.
The best output is often obtained using algorithms that take geology and rock physics into account, such as Ikon’s Ji-Fi (Joint Impedance and Facies Inversion).
Besides the reduction of interference effects due to the limited seismic bandwidth, the major advantage of seismic inversion is that it provides a common platform for communication because reservoir engineers and geologists work in a layered configurations as well.
Image shows a very good correlation between inversion predicted and well litho-facies. Strong gas-sand indication at recent Pyxis-1 location below K horizon. Highlighted by ellipse is another strong gas indication between Pluto-4 and Pyxis-1 at about the same depth as Urania-1 gas interval
RokDoc Ji-Fi is an innovative and unique Bayesian simultaneous inversion that takes partial angle stacks and solves jointly for Facies and Impedances. RokDoc Ji-Fi combines the significant benefits of simultaneous seismic inversion with depth dependent Bayesian classification to leverage additional rock physics constraints that help to produce higher quality, more consistent and geologically reasonable results.
Some of the key advantages that RokDoc Ji-Fi brings over other conventional model based inversion techniques include:
RokDoc Ji-Fi has been applied to a wide variety of hydrocarbon plays including deep water turbidites, onshore unconventionals and carbonates amongst others. RokDoc Ji-Fi has proven effective and efficient and succeeded in increasing geological understanding in many circumstances, delineating additional reserves and improving reservoir quantification.
Multiple reflectivity models are available for use in the RokDoc simultaneous inversion, including Bortfeld (1961), Wiggins (1993) and Fatti (1994). The inversion uses a fast direct algorithm to minimise an objective function computed between the synthetics and input stack traces and between the inverted properties and the input base model. Above the usual parameters available for a simultaneous inversion, the RokDoc implementation includes the ability to set weightings on each angle stack, and model weight factors to determine how the model weight is applied to the individual inverted impedances. Both lateral and time variant wavelets can be used in the inversion to account for changes in seismic bandwidth throughout the survey.
The RokDoc simultaneous inversion implementation has multiple quality control tools including:
The RokDoc simultaneous inversion module is fast and allows thorough QC of the inverted impedances. The latest version of RokDoc includes a fast direct inversion (rather than iterative) and also allows model weighting factors to be specified for the second and third impedance terms. In fact, the new algorithm is 25 times faster than the original method!
To fully describe an elastic, isotropic earth, three parameters are required. These are normally P-wave velocity, S-wave velocity and bulk density (RhoB). Density estimates in traditional P to P wave reflection seismic (or PP) inversions are made based on the high incidence angle P-wave reflection behaviour close to critical angle. With limited good quality high angle PP data normally available, the three parameters (two velocities and density) cannot easily be determined. To overcome this, generally two attributes are inverted for from PP seismic (e.g. Fatti et al., 1994 - AI and SI). The amplitude of reflected S-waves (P to S or PS) depends on only two parameters for isotropic elastic media, allowing better estimates of S-wave velocity and density from PS data.
The RokDoc PPPS simultaneous inversion tool simultaneously inverts PP and PS seismic data to acoustic impedance (AI), shear impedance (SI) and bulk density (RhoB). The use of both PP and PS data allows more robust estimates of AI, SI and RhoB. Each inversion (PP and PS) can be parameterised separately (e.g. using appropriate wavelets for the PP and PS angle stack data) and a weighting function allows user control over the influence of the PS data in the final impedance estimates.
Where PPPS seismic data has been recorded, the RokDoc PPPS simultaneous inversion allows the geoscientist to get the most out of it via the generation of more robust AI, SI and density estimates.
Model-based inversion is the industry standard approach to single stack seismic inversion for a single absolute impedance. The method uses a starting background impedance model and equivalent seismic trace (e.g. AI and zero-offset reflectivity) to generate an impedance estimate via the minimisation of an objective function. The background model is perturbed until the difference between the resulting synthetic and seismic trace is minimised.
RokDoc allows the use of laterally and time variant wavelets to capture the variation in seismic frequency content across the survey. As with the RokDoc simultaneous inversion module the RokDoc model-based inversion has a variatey of quality control tools available to the user:
Stochastic inversion allows uncertainty in reservoir properties to be taken into account, producing multiple possible geological realisations. These can then be used to create probabilistic forecasts of oil-in-place, NTG and other model properties.
We compute the reflectivity (or contrast) terms that are defined in terms of petrophysical quantities when solving equations used by Aki and Richards (1980), Smith and Gidlow (1987), etc. and which have the common form of:
Rpp(θ) = a(θ) R1 + b(θ) R2 + c(θ) R3
Introduction: In May 2017 Ophir Energy announced that the ultra-deep water Ayame-1X exploration well had reached its total depth of 5394m TVDss. The w Read More
Most current reservoir-characterization workflows are based on classic amplitude-variation-with-offset (AVO) inversion techniques. Although these meth Read More
Full waveform inversion (FWI) aims fully benefit from all the data characteristics to estimate the parameters describing the assumed physics of the su Read More