Datapages, Inc.Print this page

The Hit Cube: A Stochastic Pseudo Well Matching Tool for Seismic Prediction of Rock and Fluid Properties


Ayeni, Gboyega1, Arnaud Huck2 (1) Stanford University, Stanford, CA (2) dGB Earth Sciences BV, 7511 JM Enschede, Netherlands


The Hit Cube assigns spatial positions to stochastically generated pseudo wells with the aim of predicting rock and fluid properties throughout a seismic cube. In this method two groups of pseudo-wells are generated through Monte Carlo simulation: the Hit-targets (e.g. gas-filled reservoirs) and the False-Hits (e.g. brine-filled reservoirs). After accounting for fluid effects, synthetic seismic traces are generated for all pseudo wells. The real and synthetic seismic traces are then scaled to similar amplitude ranges. The Hit Cube algorithm matches the synthetic traces with real traces at every sample position throughout the seismic volume. If the match (similarity or correlation-coefficient) between the two traces is above a specified threshold, rock properties defined by both are considered the same and a ‘Hit' exists. Outputs from the algorithm include cubes of Hits (time-thickness of the Hit-target), Scores (sum of similarities/correlation-coefficients) and Winner wells (wells with highest similarities/correlation-coefficient) at each sample position. Probability cubes are obtained by dividing the Hit and Score Cubes of the Hit-targets by those of the False-Hits. Using this workflow and algorithm, reservoir presence and distribution within a 3D migrated seismic dataset [Area: 20x14.4Km2; TWT: 1400-2800ms] from a deltaic setting was investigated. The distribution of gas-filled channel sands throughout the cube was then predicted using 2 groups of 100 stochastically simulated pseudo wells representing gas-filled (Hits) and brine-filled (False-Hits) sands. The predicted distribution defines distributary channel-systems known to be present in the area while water saturation logs in the three real wells confirmed accurate prediction of hydrocarbon presence.


AAPG Search and Discover Article #90063©2007 AAPG Annual Convention, Long Beach, California