SPE Reservoir Evaluation & Engineering
Volume 11, Number 2, April 2008, 233-237

Summary

This paper presents a case study of a North Sea appraisal well in which a vertical fluid-composition variation, missed by a conventional pressure-gradient-analysis method, was observed in situ in real time by a new fluid-composition analyzer using visible near-infrared (NIR) spectroscopy. For optimal oil production, assessing the spatial variation of fluid properties is as vital as assessing the spatial variation of formation properties.

Conventional wireline triple-combination measurements showed that the interval of interest was uniform and free of noticeable impermeable layers. A resistivity log showed an approximate oil/water contact (OWC). Wireline pressure testing identified three different pressure gradients corresponding to gas, oil, and water, all in hydraulic communication. However, the pressure testing did not indicate a gradient in hydrocarbon composition. Fluid was sampled and analyzed in real time by a wireline fluid-sampling-analyzing tool string that included the fluid-composition analyzer. This tool analyzes petroleum fluid and gives concentrations for four group compositions (C₁, C₂–C₅, C₆₊, and CO₂), gas/oil ratio (GOR), and qualitative information regarding heavy-end content and stock-tank crude density. The analyzer showed that the hydrocarbon fluid in an oil-bearing zone was not vertically homogeneous but, instead, had a vertical variation. The samples captured by the wireline sampling tool were sent to a laboratory for compositional analysis that confirmed the variation determined by the downhole analysis. Both results identified the heterogeneity of hydrocarbon fluid in the interval.

This paper also briefly covers the measurement principle of the analyzer and discusses the impacts and benefits brought about by the new technology. The concept of flexible fluid sampling is particularly important because it enables operators to make sampling decisions on the basis of real-time fluid-analysis results rather than a predetermined job plan.

Introduction

It is known that some oil reservoirs show a fluid compositional variation across relatively short vertical intervals. Such reservoirs are of great interest for reservoir engineers and petrophysicists because a proper assessment of the formation-fluid gradients is critical to optimum hydrocarbon production. There are several different mechanisms that create fluid compositional gradients. Fluid gradients can be caused by gravitation, thermal gradients, biodegradation, water washing, multiple reservoir charges, and leaky seals. Because it is difficult to predict the existence of fluid gradients a priori, it is prudent to determine the magnitude of these gradients by actual measurements. Current wireline formation evaluation is inadequate to determine the magnitude of fluid compositional gradients. Even multiple sampling with subsequent laboratory analysis is somewhat risky because a variation of fluid properties measured in separate sample bottles might be caused by differing levels of oil-based-mud (OBM)-filtrate contamination or to some degree of nonrepresentative sampling. In addition, it is often difficult in practice to justify the extra cost of taking multiple samples in a small interval without some indication of fluid variations. It is much more preferable to perform the sample analysis in situ so that the subsequent sampling program can be optimized in real time by comparing observations to predictions.

Visible-to-near-infrared (VIS/NIR) absorption spectroscopy is widely used to assist wireline fluid sampling today. Identification of gas, oil, and water is now well established (Smits et al. 1995). Problematic OBM contamination is quantified during sampling jobs using buildup curves of spectral data (Mullins et al. 2000b; Fadnes et al. 2001). Recent advances have enabled analysis of live fluid properties in situ. For example, in-situ GOR measurement by NIR spectroscopy has been established (Mullins et al. 2001) and is now commercially available (Dong 2003). A recent study showed the feasibility of downhole fluid-composition analysis conceptually (Fujisawa et al. 2002). The authors built prediction models using principal-components regression to various hydrocarbon spectra measured at high-pressure and high-temperature conditions typical for oil and gas reservoirs. The fluid-composition analyzer built on the basis of this principle estimates the concentrations of C₁, C₂–C₅, C₆₊, and CO₂ in the flowline fluid and its GOR. The first field application of the tool allowed accurate downhole fluid characterization of a complex miscible-flood program (Fujisawa et al. 2003). This downhole-fluid-analysis (DFA) technique was also found useful for detecting formation compartmentalization (Mullins et al. 2004).

In this paper, we report a case study that identified a fluid compositional gradient by use of a fluid-composition analyzer (Fujisawa et al. 2003; Mullins 2004). Fluids in the flowline of a wireline sampling tool are analyzed in terms of the concentrations of C₁, C₂–C₅, C₆₊, CO₂, and water. With this compositional analysis, GOR can be estimated also. Here, we show that the compositional analysis is far more sensitive than the pressure-gradient method for detecting slight changes in reservoir fluid. Results demonstrated the advantages of real-time DFA and led to a modified sampling job that validated those measurements. Four sampling points in the liquid column below a gas-cap zone were used to establish a large GOR gradient in a 30-m column of oil. Sample acquisition with subsequent laboratory analysis also confirmed a compositional gradient.

© 2008. Society of Petroleum Engineers

View full textPDF ( 1,253 KB )

History

• Original manuscript received: 11 May 2004
• Meeting paper published: 26 September 2004
• Revised manuscript received: 23 September 2007
• Manuscript approved: 2 October 2007
• Version of record: 25 April 2008