geo inversion

Discussion views - Monday: 

Discussion views - Monday Non-uniqueness is everywhere Seismic data is highly oversampled - less Non-seismic methods are generally undersampled - more Equivalence for some methods What constraints/apriori is valid?? Deliver on what is needed Explain probility/risk of interpretaion

Discussion views - Tuesday: 

Discussion views - Tuesday Technology transfer is a separate process Parallel applications should be allowed Client mentorship is essential Communication processes are essential

Discussion views - Wednesday: 

Discussion views - Wednesday Errors in data need to be errors in model Data benchmarks need to use realistic noise Understand physics better (anisotropy, rock physics) Match data and synthetics Improve acquisition Data combination/integration Bandwidth must include space domain Shahred Earth model

Summary discussion... Page 1: 

Summary discussion... Page 1 More data, more direct data (prestack, less influenced by processing) D.H. Direct rock properties D.H. Tayloring problem to desire (fit-for-purpose) D.H. Output must be simple andamp; final answers to make operating decisions D.H. Analysing the uncertainities D.H. Integrated surface and borehole seismic techniques (M.K.S.)

Summary ...continued: 

Summary ...continued State-of-the-art: Prestack genetic algorithm, seismic guide geostatistical inversion (M.K.S) Seismic (well log constrained(andamp;VSP)) (2 yrs) (M.K.S.) Geostatistical (pre-stack constrained) (M.K.S.) Need more research/out off box thinking (better handling of non-uniqueness, 'undershooting') C.W. III Joint inversion (M.B) focussing on a joint model..can be complimentary.

Summary ...continued: 

Summary ...continued Describe probabilities (=risk) parameters M.B. Use tomographic techniques with prestack C.C. Develop standardized procedures C.C. Inversion will be more used for design of new experiments (C.C) Careful marketing/sale/communication (F.G) Integrated team andamp; education (F.G.)

Summary ...continued: 

Summary ...continued Ask usefulness question from the beginning Invert for useful parameters (porosity) Joint inversion, multi-attributed, shared Earth model...goal is 'Understand interwell space' Interwell measurements (3 km) Integration to make a ‘Model’ of the subsurface optimized andamp; constrained with all possible data Acknowledge uncertainity andamp; translated to risk

Summary ...continued: 

Summary ...continued Joint inversion (surface and borehole) is coming (seismics, EM, gravity) Interaction of human with data/model Look for consistent model integrating expert knowledge New workshop w/ mathematical modelers Shared Model must be ‘alive’ (=easy to update) Different techniques will be going together

Summary ...continued: 

Summary ...continued Use true depth for inversion Use 4 C to address lithology in depth Better visualization of different data