Stochastic detection of transgressive and regressive sequences from well-log data using discrete wavelet transform: first results

@inproceedings{RUNKJRM100,
 abstract = { This article proposes a new multi-scale method to detect geological stratigraphic sequences from
well-logs in silico-clastic coastal reservoirs. This method uses multi-scale signal analysis (discrete
wavelet transform) to compute the probability density of placing maximum 
ooding and maximum
regressive surfaces as a function of depth. It then recursively decomposes the studied stratigraphic
section into sub-intervals where the analysis is repeated. This process generates a succession of
transgressive and regressive sequences, that may be used in automatic or expert-based well corre-
lations.
Currently, the first step of the transgressive and regressive sequences detection is mainly based
on the density of probability to find maximum 
ooding surfaces or a maximum regressive surfaces
regarding Gamma Ray, and assumes maximum regressive surfaces to be associated to locally low
Gamma Ray values and maximum 
ooding surfaces to locally high Gamma Ray values and the
second step is the stochastic simulation of surfaces and the sequence detection. The application of
this method to a Jurassic North Sea data provided by Equinor shows that results are consistent
with model hypotheses and allows us to establish some future research directions. },
 author = { Baville, Paul AND Caumon, Guillaume AND Le Solleuz, A. AND Géraud, Yves },
 booktitle = { 2018 Ring Meeting },
 publisher = { ASGA },
 title = { Stochastic detection of transgressive and regressive sequences from well-log data using discrete wavelet transform: first results },
 year = { 2018 }
}