Inverting SP Logs Using Artificial Neural Networks and the Application in Reservoir Characterisation
The detailed character of SP logs can hold invaluable information, in particular in old fields without modern logs. Unfortunately, the resolution of the SP is rather low, so that in laminated reservoirs, its direct use tends to overlook thin sands, in particular when they are saturated with hydrocarbons. On the other hand, the sand count for thick sands with inter-bedded shale laminations would be overestimated.
In this article, a new fast and robust technique for more accurate sand thickness calculations is described, by inverting the SP log using neural networks and thus enhancing the resolution considerably. The first step of the method involves forward filter modelling in order to mimic SP responses in sand-shale sequences. The filters are based on theoretical considerations and adjusted to correspond to the standard SP correction charts. According to theory, the filters depend on the ratio of mud resistivity to formation resistivity. They are designed so that this ratio scales the characteristic length as well as the relative weight in each depth interval. These filters are applied to artificial wells with predefined sequences of shales and sands of varying thickness, shaliness and resistivity, typically seen in clastic reservoirs.
Then an inverse filter process is created as an artificial neural network. One artificial well is used as a training set for the neural network brain which is able to window processing. The input is the actual normalised SP, the desired output the original PSP (pseudo-static SP). A brain was finally chosen which both revealed all characteristic features of the original PSP, but also was robust enough not creating spurious deflections in the test sets, including a real SP log. A comparison of the method with the result from a full FEM modelling was made using a literature example.
In the last step, the actual sand/shale ratio is estimated, based on the inverted SP and the resistivity log. This further enhancement is in particular effective in hydrocarbon-saturated sands. The application of the method is demonstrated with a large field composed of stacked sandstone reservoirs.
KeywordsReservoir Characterisation Inverse Filter Spontaneous Potential Shale Content Thin Sand
Unable to display preview. Download preview PDF.
- Doll, H.G. (1948) The SP log: Theoretical analysis and principles of interpretation. Petr Trans AIME (JPT) 179:146–185.Google Scholar
- Doll, H.G. (1950) The SP log in shaly sands. Petr Trans AIME (JPT) 189:205–214.Google Scholar
- Pirson, S.J. (1963) Handbook of Well Log Analysis, Prentice Hall, Englewood Cliffs.Google Scholar
- Pirson, S.J. Electrical Logging, a Study of the SP Curve. Oil and Gas Journal, Oct 1947.Google Scholar
- Schlumberger Log Interpretation Principles (1972), p 7–12.Google Scholar
- Waxman, M.H. and Smits, L.J. Electrical conductivity in oil bearing shaly sands. SPE Journal, June 1968: 107–122.Google Scholar
- Smits, L.J. SP log interpretation in shaly sands. SPE Journal, June 1968:123–136.Google Scholar
- Schlumberger Chartbook (1979), Chart SP-3 and SP-4.Google Scholar
- Tabanou, J.R., et al. (1988) SP deconvolution and quantitative interpretation in shaly sands. Log Analyst 29/5: 322–343.Google Scholar
- Taherian, M.R., et al. Spontaneous potential: Laboratory experiments and modeling results. 33. Ann Symp SPWLA (1992): paper E. Also in Log Analyst 36/5 (1995): 34–47.Google Scholar
- Poupon, A., Loy, M.E. and Tixier, M.P. Contribution to electrical log interpretation in shaly sands. Petr Trans AIME (JPT) 201:138–145.Google Scholar
- Segesman, F. and Tixier, M.P. (1959) Some effects of invasion on the SP curve. Petr Trans AIME (JPT) 216:138–146.Google Scholar
- [13a]Dakhnov, V.N. and Keller, G.V. (1962) Geophysical Well Logging, p.323–337. Quarterly Summer School of the Colorado School of Mines.Google Scholar
- [13b]Minh, C.M. et al. (2000) Integration of NMR with other openhole logs for improved formation evaluation. SPE Reservoir Eval. & Eng. 3/6: 509–516.Google Scholar
- Arbogast, J.S. and Franklin, M.H. (1999) Artificial neural networks and high-speed resistivity modelling. Hart’s Petr Eng, May and June 1999.Google Scholar
- Poupon, A. and Leveaux, J. Evaluation of Water Saturation in Shaly Formations. Trans SPWLA Symp 1971.Google Scholar