AVO analysis & rock physics based AVO inversion Stud. MSc Asger Johansen Supervisors: Thomas Mejer Hansen, IMM DTU Ida Lykke Fabricius, BYG DTU 1/46Thesis.

Slide 1 AVO analysis & rock physics based AVO inversion Stud. MSc Asger Johansen Supervisors: Thomas Mejer Hansen, IMM DTU Ida Lykke Fabricius, BYG DTU 1/46Thesis defence 14/5-2012 Slide 2 Presenting Objectives of the thesis Part I-III –Method –Results –Conclusions Thesis: Discussion Thesis: Conclusion 2/46Thesis defence 14/5-2012 Slide 3 3/46 (picture: Schiøler et. al) Nini-1 Area of seismic data Thesis defence 14/5-2012 Slide 4 Objectives 4/46 Investigate and assess the AVO response of North Sea oil and brine saturated greensand from the Hermod formation Investigate the possibility of identifying oil reservoirs from seismic data Investigate the prospects of predicting the physical properties of greensand from seismic data by a rock physics based non-linear AVO waveform inversion AVO response AVO analysis AVO inversion Thesis defence 14/5-2012 Slide 5 Workflow 5/46 AVO response Prospects of inverting seismic data AVO analysis Where to invert AVO inversion Predicting physical properties Thesis defence 14/5-2012 Slide 6 Part I: AVO response of greensand Forward model the AVO response Estimate elastic properties –Oil and brine sat. greensand –Of an upper and lower reflecting layer Observe the AVO response from seismic data 6/46Thesis defence 14/5-2012 Slide 7 7/46 (picture: Kearey, P. (2002) AVO (Amplitude variation with offset) Thesis defence 14/5-2012 Slide 8 Identified greensand (Oil saturated) 8/46 Thesis defence 14/5-2012 Slide 9 9/46Thesis defence 14/5-2012 Slide 10 10/46 Greensand Balder Ekofisk Thesis defence 14/5-2012 Slide 11 AVO response 11/46 Observed from the seismic data at the Nini-1a well location Thesis defence 14/5-2012 Slide 12 AVO response From seismic data 12/46Thesis defence 14/5-2012 Slide 13 Conclusion - Part I: AVO response of oil sat. greensand AVO response of oil sat. greensand different from brine sat. greensand Oil saturated characterized by: –a top reflector which amplitude increases negatively with offset –A bottom reflector which amplitude increases positively with offset 13/46Thesis defence 14/5-2012 Slide 14 Part II: Identifying possible oil reservoirs Method of identifying the seismic attribute of oil Visualize the possible oil reservoirs 14/46Thesis defence 14/5-2012 Slide 15 15/46 B (slope) ≈ 0 B (slope)= big negative A A Thesis defence 14/5-2012 Slide 16 16/46 Brine sat. Oil sat. Thesis defence 14/5-2012 Slide 17 17/46 Greensand Balder Ekofisk Thesis defence 14/5-2012 Slide 18 18/46Thesis defence 14/5-2012 Slide 19 19/46 Brine sat. Oil sat. Thesis defence 14/5-2012 Slide 20 20/46 Nini-1a location Thesis defence 14/5-2012 Slide 21 21/46 Greatest distance Thesis defence 14/5-2012 Slide 22 Ekofisk 22/46Thesis defence 14/5-2012 Slide 23 23/46Thesis defence 14/5-2012 Slide 24 24/46Thesis defence 14/5-2012 Slide 25 25/46Thesis defence 14/5-2012 Slide 26 Conclusions from part II Method of identifying possible oil reservoirs verified The oil reservoir around Nini-1a and Nini-4 identified Another “known” oil reservoir identified A third possible oil reservoir identified 26/46Thesis defence 14/5-2012 Slide 27 Part III: AVO inversion Predicting physical properties of greensand from seismic data non-linear AVO waveform inversion Metropolis inversion 27/46Thesis defence 14/5-2012 Slide 28 Principle of the algorithm –Iteratively –Sampling a realization of the posterior distribution 28/46 Propose model Forward model Misfit to data Accept / reject model Thesis defence 14/5-2012 Slide 29 Forward model Non-linear rock physics model Non-linear reflectivity model Convolution model 29/46Thesis defence 14/5-2012 Slide 30 The rock physics model Hertz-Mindlin grain contact model –Modelling: K dry and μ –Estimating: Poisson ratio Glauconite bulk and shear moduli Gassmann fluid substitution –Estimating: Mineral modulus 30/46Thesis defence 14/5-2012 Slide 31 Illustrate forward model 31/46Thesis defence 14/5-2012 Slide 32 Sampling a posterior distribution Accept model by: –Likelihood of a proposed model –Accept: If: L(m_propose) >= L(m_current) Else if: L(m_proposed) / L(m_current) > rand([0:1]) –m_current = m_propose Measure of how well the accepted models fits data relative to each other –Number of iterations between accepted models SPE meeting 28/2-201232/46 Slide 33 Illustrate sampling method 33/46Thesis defence 14/5-2012 Slide 34 Synthetic reference model –Different prior scenarios The waveform inversion requires the elastic properties of the 1. and / or 3. layer to be known Inversion using the Hertz-Mindlin rock physics model is found to require the P eff, C and φ glauconite to be known 34/46Thesis defence 14/5-2012 Slide 35 Inverting seismic data 35/46 Φ = 0.25 Φ glau = 0.4 F glau = 0.3 C=15 P eff = 22 MPa Oil sat = 0.9 Thesis defence 14/5-2012 Slide 36 Prior distribution 36/46Thesis defence 14/5-2012 Slide 37 37/46Thesis defence 14/5-2012 Slide 38 38/46Thesis defence 14/5-2012 Slide 39 39/46Thesis defence 14/5-2012 Slide 40 Probabilistic estimates on: Porosity Glauconite fraction 40/46 Oil saturation Thesis defence 14/5-2012 Slide 41 Modeling greensand Estimating the: Mineral modulus Poisson ratio Glauconite bulk and shear moduli Investigating these estimates by inversion of seismic data The estimates proposed by Zakir Hossain are verified from seismic data 41/46Thesis defence 14/5-2012 Slide 42 Inverting seismic data – Nini-1a A discrepancy observed –Seismic data –Forward model –Elastic properties above the greensand –Inversion by a 3-layer model 42/46Thesis defence 14/5-2012 Slide 43 Conclusion part III Feasible to invert the physical properties of greensand from seismic data Requires prior information on both the greensand and the elastic properties above and/ or below Challenge describing the seismic data by a 3- layer model 43/46Thesis defence 14/5-2012 Slide 44 Discussion Thesis: The assessed AVO response depends on the geological setting The method of identifying possible oil reservoirs only indicates a possible seismic anomaly Challenge of inverting seismic data at locations without information from a well log Describing the seismic data by a 3-layer model 44/46Thesis defence 14/5-2012 Slide 45 Conclusion Thesis: The results of part I supported: –the possibility of identifying possible oil reservoirs from seismic data by AVO analysis –An the possibility of inverting physical properties of greensand by AVO inversion Two possible oil reservoirs identified Inversion of physical properties are feasible however a challenge is encountered Thesis defence 14/5-201245/46 Slide 46 Conclusion Thesis: Elastic properties and the physical properties of greensand are required to be known Inverting the physical properties at the two “unknown” possible reservoirs was not possible 46/46Thesis defence 14/5-2012