Abstract

The borehole image provides high resolution circumferential resistivity image data that can bed used to describe surface area conductivity and from that derive distributions of pore throat, permeability and porosity. This paper presents some of the approach that can be used to derive this information within the Carbonate sequence of these basins and presents a couple of examples showing how the information can help target fracture zones.

The brittleness index was computed using a conventional log analysis approach and increasing values tend to be associated with increasing fracture complexity, particularly when fracturing the formation by artificial means.

Although strictly from shale gas work, the Rickman et al., 2008, SPE 115258) paper has some interesting observations on brittleness which is reproduced in the Figures 1.1 and 1.2 These figures show applying brittle-ness one can estimate the type of fracture system. By combining with the grain size and pore throat work intervals where the fractures will access more of the pore structure can be identified.

The changes in rock fabric are captured in image logs as variation in conductive and resistive image texture proportion. This in turn is captured by variation in the BHI porosity distribution histogram in the form of changes in amplitude, width and skewness of the porosity map. A special technique was adopted to capture the above changes in porosity histograms and converted into a single heterogeneity index curve. This curve can be utilised to predict petrophysical facies changes in the reservoir and stacking patterns. As a result, the BHI petrophysical properties are used to identify small-scale changes in the rocks fabric and subsequently generating an enhanced porosity log and permeability index log. This highlights the degree of reservoir property variability and uncertainties in the conventional core plug measurements.

To further enhance and capture variation in surface area conductivity and heterogeneity, the BHI conductivity histogram limits are set within the conductivity distribution to eliminate edge effects. These effects include, but not limited to, borehole fluid (conductive) when pads not in contact with the borehole wall and cemented bands or tar bands (resistive) features measured by the tool (Figure 1.3). This approach will filter out any anomalies in the data and enhance conductive variability of the interval logged. These data are presented as textural histograms. The bi-modal distribution (separation or widen in histogram) suggests that the reservoir is poorly sorted and/or not uniformly packed. The skewness of the histograms to the right suggests an increase in relative particle size. The conductiance from the tool is actually reflecting the surface area conductivity of the particles which can be grains, crystals or other particles. For the purpose of brevity particle size will be used to synonymously refer to either crystals and associated surface area conductivity or grains. The results have been extensively compared core lithology descriptions and photographs and there is a good match with the trends observed in core.

This is an approach to create histograms over an interval of 1 to 3 inches that take advantage of the high resolution of the BHI data and construct histograms of the resistivity or conductivity data directly. This approach has been shown in literature (Newberry, 2004) to provide a measure of sorting and potentially particle size. The particle size analysis is obtained if the textural differences (e.g. between sandstone, Heterolithic Sand, Silt and shale) are identified from image-core integration phase. Sorting Index depends on the relative spread of the resistivity and it is not in absolute value.

The workflow that enables the computation of particle size and pore throat size is illustrated in Figure 1.4.

Effectively the computation is commenced from a map of the conductivity, which is shown in Figure 1.4.

A normalized textural map is produced and illustrated in Figure 1.5 as previously discussed.

The textural map is rescaled to produce a particle size distribution as approximated from surface area conductivity in the recognized sieve PHI scaling (Figure 1.3).