Extraction of permeability from time‐lapse seismic data

Abstract

A B S T R A C TTwo formulae are developed for estimating horizontal permeability directly from maps of 4D seismic signatures. The choice of the formula used depends on whether the seismic is dominated by changes of pressure or saturation. However, pressure derived from time-lapse seismic, or seismic amplitudes controlled predominantly by pressure are to be preferred for optimal 'illumination' of the reservoir. The permeability is predicted to be dependent on porosity but weighted by a 4D term related to the magnitude and spatial gradient of the 4D signature. Tests performed on model-based synthetic seismic data affirm the validity and accuracy of this approach. Application to field data from the UK continental shelf reveals a large-scale permeability variation similar to the existing simulation model, but with additional fine-scale detail. The technique thus has the potential of providing extra information with which to update the simulation model. The resultant permeability estimates have been successfully ground-truthed against the results of two well tests. As non-repeatable noise in the time-lapse seismic data diminishes with improved 4D-related acquisition, it will become increasingly possible to make robust permeability estimates using this approach. I N T R O D U C T I O NIt is now widely acknowledged that time-lapse seismic data, if interpreted appropriately with due consideration of reservoir production and injection history, can provide invaluable information with which to optimize hydrocarbon recovery. This information includes, for example, the location of hydraulically conductive faults, no flow barriers or high permeability pathways, the identification of isolated compartments possibly containing bypassed oil, the position of injected fluid fronts or contacts, and general field-wide pressure and saturation changes. To date, most of these features are commonly inferred by detecting regions of brightening or dimming on difference sections or maps, and are thus essentially semiquantitative and can provide ambiguous interpretations. This style of 4D analysis is useful for reservoir management purposes, but in general may not offer sufficient accuracy to update the flow simulation model and hence to reduce uncertainty in reservoir performance predictions. A more detailed and in-depth level * E-mail: Colin.MacBeth@pet.hw.ac.uk of quantitative analysis is required, one that delivers accurate numerical estimates of absolute permeability variations, transmissibility factors, and reservoir connectivity at a variety of scales. To address this general challenge, we start by focusing here on estimating one of these parameters, i.e. reservoir permeability, which is the key flow property that controls strategies for well completion and production.The usual starting point for determining the permeability field for the flow simulation model is the core plug or welllog data. The interwell space is then populated by extrapolating between wells using an appropriate set of geostatistics, and this final reservoir model i...

Reservoir characterization based upon the onset of time-lapse amplitude changes

“…Estimating reservoir properties such as permeability, based on onset times, contrasts with previous efforts that relied upon timelapse amplitude, traveltime, and waveform changes (Huang et al, 1998;Vasco et al, 2004;Vasco, 2004a;Dong and Oliver, 2005;MacBeth and Al-Maskeri, 2006;Dadashpour et al, 2008Dadashpour et al, , 2009Feng and Mannseth, 2010;Rey et al, 2012). Here, we demonstrate that onset times, extracted from seismic time-lapse reflection data recorded in a few surveys, can provide a reliable basis for reservoir characterization.…”