No abstract

Colombia's Oil production is around 1.1 million barrels per day (bpd), where 57% is from Heavy Oil fields. Current oil recovery is in the 15% to 17% range; with targets to increase it up to 50% through different methods. A typical reservoir exhibits hydrocarbon viscosity variations in hundred of centipoises, with formation water salinity typically below 5000 ppm and heterogeneities driving a complex fluids distribution. Since the low amount of salt in these environments prevents low frequency conductive devices for contrasting water versus hydrocarbons, where additionally, resistivity profiles are ambiguous to assess fluids mobility in the reservoirs. In this context, the incorporation of additional physics of measurements opens a new perspective in the reservoir evaluation in Llanos basin, by reducing uncertainties and helping in the initial reservoir characterization. The new generation of wireline measurements supporting the present job is represented by multifrequency dielectric propagation, radial magnetic resonance and dynamic testers, in addition to the conventional triple combo logs. Since good-oil bearing rocks (high porosities and permeabilities, very clean sands, high oil saturations) do not guarantee oil production (very high water cut is likewise common), the identification of movable oil and free water volumes in low salinities is mandatory. Understanding its distribution across sands is also a critical factor in heavy oil environments. As a resistivity and salinity-independent reservoir evaluation approach, the combination of dielectric dispersion and radial magnetic resonance, provides a valuable sensitivity for the evaluation of displaced oil, free and irreducible water, viscosity and rock quality variations. Dielectric Dispersion is the variation of relative permittivity and conductivity versus frequency, enabling pore fluids determination. With a dielectric analysis at two depths of investigation, integrated with NMR-based diffusion mesurements, a direct identification of movable oil under filtrate invasion conditions and free water presence is achieved. The correlations encountered between the dielectric dispersion is encouraging; whereas a better understanding on the movable oil occurrence and estimation of the fluid to be moved during production is achieved. Discussions with case studies in Llanos Basin are presented in this paper.

Important hydrocarbon accumulations occur in tight rocks in Colombian areas. Those tight reservoirs consist of clean sandstones with matrix porosities in the 3% to 4% range, relatively complex mineralogy and naturally fractured. The success of achieving a representative formation evaluation relies on obtaining accurate porosity, oil, gas, water saturations, natural fractures detection and good estimates on reservoir permeability. Resistivity-based approaches are difficult to apply since reservoir conductivity is not only influenced by fluid type, but also by salinity (typically low in our reservoirs), variable tortuosity (mostly high in the matrix and very low in fractures) and very high formation resistivity (above 1,000 ohms.m). In addition, a combination of low pores volumes and a matrix not properly assessed, leads to high errors in the porosity determination with conventional logs (in a 3 – 4 p.u. reservoir, the porosity error computation can be as high as 50%). Uncertainties in porosity estimates also translates to uncertainties during saturation assessment. Further challenges are found when attempting the saturation computation from resistivity logs. The tight sands are drilled with Oil Based Muds, creating a logging environment where only induction logs are possible. However, since the resistivity range in these rocks is above 1000 ohm.m range, the induction measurements are out of range in many of the target zones. Alternative formation evaluation methods for assessing fluids saturations, like magnetic resonance, sigma and carbon-oxygen logs cannot be applied below 10 porosity units; whereas dielectric measurements strongly depend on accurate porosity computations for deriving the hydrocarbon volume. Some of these reservoirs, are also deep (in the 17,000 ft range) and close to foothills, where wellbore stability issues and narrow mud weight windows used for drilling, translates into higher risks for open-hole logging via logging while drilling or wireline conveyance, all of it detrimental to data acquisition in open hole. Therefore, the case studies presented in this paper were assessed in cased hole conditions. In this paper, we present a solution that cover tight matrix and natural fractures assessment, at a level not previously achieved. At the tight matrix level, we carry out advanced nuclear spectroscopy with a new pulsed neutron device, that carry out simultaneous time domain and energy domain measurements. A new resistivity and salinity independent methodology for obtaining Gas saturation from a new measurement in the industry known as "Fast Neutron Cross Section" (FNXS), oil saturation from the total organic carbon (TOC) log, mineral volumes solved from formation elemental concentrations from energy domain, and porosity from hydrogen index obtained from the spectroscopy time domain, is presented. At natural fracture level, we make use of a Borehole Acoustic Reflection Service for deep natural fracture detection and spatial orientation analysis, done at cased hole conditions. The main advantages of the new method for obtaining porosity, mineralogy, salinity-independent hydrocarbon saturation in tight matrix and natural fracture assessment behind casing are: 1) conversion of dry weight total carbon to oil saturation, and fast neutron cross section to gas saturation done through a simultaneous inversion by solving matrix-porosity-fluids volumes into an elemental analysis, proven to work at low porosities rocks; 2) independency of salinity and reservoir tortuosity effects; 3) clay and/or other lithology effects is quantified and taken into account; 4) faster logging speeds and improve tools combinability in bigger holes while ensuring full reservoir assessment in small holes; 5) operational time reduction. The spectroscopy logging is carried out in single acquisition pass at 150 to 350-feet per hour (ft/hr), whereas sonic acquisition is done at 400 ft/hr in a single pass as well.

Major heavy oil accumulations are found in the tertiary sandstones of the Lagunillas Formation of the Costanero Bolivar Field, located in Lake Maracaibo, Venezuela. The hydrocarbons are found in poorly consolidated shaly-deltaic sands, at depths of around 1200 ft to 2350 ft. The oil ranges from 10 to 18 degrees API, with viscosities ranging from 400 cP to 10000 cP. The formation water salinity is below 5000 ppm and variable within the reservoir, after years of injection of fresh water and steam to increase recovery. Consequently there is today no correlation between water cut and resistivity and the differenciation between oil and water with conventional petrophysical techniques is inaccurate. Conventional log analysis has limited potential since the resistivity shows identical values in both oil and water bearing levels. Deciding on a completion strategy from an inaccurate saturation computation is a major challenge. Additionally, the free water presence reduces the net pay and rapidly increases the probability of water production in this high oil viscosity environment. Therefore an accurate assessment of free water and oil viscosity is a critical factor in the economics of the field. The present work incorporates dielectric and molecular diffusion measurements, showing significant progress in detecting free water from oil and defining the most prospective intervals. Movable oil and fresh water are clearly identified using dielectric polarization at multiple frequencies. The dielectric measurement provides the water-filled porosity, while the magnetic resonance identifies the irreducible versus free water within that volume. This allows predicting the likelihood of producing hydrocarbon or water in areas with high oil saturation. In conclusion, the integration of dielectric polarization and diffusion information at multiple depths into the reservoir enable to distinguish oil from free and bound water and to estimate the oil viscosity, a result impossible to obtain with conventional logs in these environments. This integrated methodology allows accurate reservoir characterization and definition of the production potential of these heavy oil sands, leading to improved completion decisions. The development campaign in Lagunillas sands now has a new workable technique to reduce uncertainties and to optimize heavy oil production.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractAdvanced nuclear magnetic resonance (NMR) methods for fluids identification and reservoir evaluation were developed recently. These techniques, focused on identifying and characterizing fluids in the reservoirs, were mostly applied in clastic and carbonate reservoirs. In such environments, fluid characterization with NMR logs is relatively simple when the fluid responds to standard models. However, exceptions to NMR models such as restricted diffusion produce variations in model behavior and cause erroneous answers. Experience shows that restricted diffusion occurs mostly in carbonates with water trapped in small pores and in short T2(ms) domains (bound water) in clastic formations. This effect reduces water diffusibility and produces NMR signals occurring in oil diffusion domains, with the consequent incorrect computation of liquid hydrocarbons.The purpose of this study is the characterization of the restricted diffusion effect from rock samples measured in the laboratory with an NMR logging tool, as well as from field examples in diverse lithologies. It is expected that such analysis will help in reservoir petrophysical characterization and will avoid misinterpretations in fluid typing, when performing fluids identification based on NMR principles.Since restricted diffusion phenomena have been observed in medium to long transverse relaxation times T2(ms), above T2 cutoff for free fluids (in moderate to high porosity environments), some possible explanations are proposed to understand such phenomena and the implications on reservoir evaluation both in exploratory and production wells.