document preview

Abstract

Since the introduction of the dipole shear sonic measurements a large number of compressional and shear data sets have been acquired, especially in shaly sands. Correlations between these measurements are now well established and used routinely. The Castagna mudrock and sandstone lines on the Vp and Vs crossplot are popular among geophysicists while trends of Vp/Vs versus Δt proposed by Brie et al. (1995) are popular among petrophysicists. As expected, a comparison of these correlations shows that they are quite similar in saturated and dry conditions. However, the Vp/Vs trends can simulate different pore fluid types or saturation conditions because the fluid effect is accounted for through the Gassmann equation.

Using published correlations for adiabatic elastic moduli with temperature and pressure, we have generated slowness charts for water solutions and hydrocarbons and used them to refine the Vp/Vs trends. In particular we generated different trends for sands impregnated with fresh water, salt-saturated water and live oil. We show that the Vp/Vs trend for live oil is quite distinct from that in water.

An application of these correlations is to convert shear to compressional. As shear slowness is little sensitive to fluid type, the obtained compressional slowness, Δt_wet, simulates water-filled shaly sands. A comparison of Δt_wet with measured Δt provides a quicklook indicator of hydrocarbons. Various examples are presented and discussed. We show that all correlations give good results in sands with porosity larger than 20%. Between 10 and 20% porosity, correlations must be adjusted to obtain reliable results. Below 10% porosity, the technique becomes unusable.

The Δt_wet estimate has other interesting applications in the domain of rock mechanics and stress evaluation because it is unaffected by fluid content. An example shows how the technique was applied to improve hydraulic-fracture design.

Introduction

Dipole shear sonic measurements are now widely used in the industry and provide invaluable data for seismic interpretation, formation evaluation and rock mechanics applications. The Vp/Vs versus Δt crossplot introduced in 1995 (Brie et al., [1]), is a valuable tool for quicklook evaluation of the data and has gained wide acceptance. However, this plot only described the effect of water and gas in shaly sands. To go further a detailed description of the properties of reservoir fluids with temperature and pressure is needed and the effect of natural oil need to be added. Also the response of sonic logs in carbonate reservoirs has been studied and can be updated.

Various quicklook techniques have been proposed for hydrocarbon detection from the sonic log. Williams (1990, [2]) used the Vp/Vs versus Δts crossplot to identify hydrocarbon bearing zones; Shiuma et al. (1997, [3]) used the Vp/Vs versus Δt crossplot fitted with straight line. We will propose a technique based on the reconstruction of the wet sand Δt from the shear slowness, which has the advantage of working on a wider range of cases and especially at low porosity. This technique was applied with success in the Nakhla field (Wintershall Libya).

Rock mechanics evaluation uses sonic data and is often disturbed when hydrocarbon effect is present on the log. In this case using a wet Δt instead of the measured value would make the evaluation more accurate and the predictions more reliable. Recent cases of hydro-frac design in the Nakhla field have shown this need and will be discussed.

Vp/Vs Correlations

Geophysicists often use the Castagna sand and mud rock line on the Vp versus Vs crossplot (Castagna et al., 1985, [4]). An additional line on this plot, linking the origin to the sand point would also represent gas-bearing sands. These lines are based on simple equations, making them very useful for quicklook evaluation and quality control; however they lack resolution and flexibility for petrophysical applications.