When core data, such as mineralogy, porosity, total organic carbon, and elastic properties, are available from a large multiwell database, it is possible to extract meaningful statistical relationships between the data. These relationships can then be applied in other wells with limited or no core data. X-ray diffraction (XRD) mineralogical data is unique in that it can readily be transformed to synthetic wireline geochemical elemental data, such as the weight fractions, aluminum, silicon, and calcium. This transformation establishes a relationship between the mineralogical data and the synthetic wireline geochemical data, which can be applied to "real wireline" geochemical data to select an appropriate set of minerals for prediction from the wireline geochemical data. Statistical optimization based approaches are generally used to perform this prediction of mineralogy from wireline data. The most difficult aspect of predicting mineralogy from wireline geochemical data is the selection of the proper mineral model. The proposed approach assists the analyst in the selection of the proper model. In addition, this paper shows how core XRD mineralogical data can be used to establish constraints between the predicted mineralogy, such as percent clay fraction illite greater than 60%. An accurate prediction of mineralogy from wireline geochemical data can be transformed to an accurate prediction of grain density, enabling an accurate prediction of total porosity (corrected for kerogen effects). The workflow described in this paper provides an effective approach for establishing and predicting mineralogy, grain density, and porosity from wireline geochemical data. This paper presents results for a Haynesville shale example. Finally, the proposed approach helps to maximize the value of the core data for shale gas plays.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractOne of the main applications of MRILab, the recently introduced NMR fluid analyzer, is the estimation of mudfiltrate contamination during the pump-out phase, prior to securing high quality fluid samples for PVT analysis. The NMR fluid analyzer utilizes T 1 saturation recovery measurements for contamination analysis. The main objective of this paper is to discuss the principles behind recently developed T 1 -based novel methodologies that yield accurate and robust contamination estimates. The proposed algorithms were used to estimate the contamination levels in an offshore well drilled with OBM. Comparisons with laboratory assays on nine fluid samples show excellent agreement, thus proving the concept of using an NMR fluid analyzer for contamination analysis in sampling applications. A secondary objective of the paper is to illustrate the benefits of integrating fluid analyzer data with open-hole and NMR logs. Such integration leads to an overall improved interpretation, particularly in terms of hydrocarbon fluid properties.

The Eagle Ford Shale hydrocarbon-fluid properties depend on the source rock maturity and, within the formation, occur in varying degrees of gas, gas condensate, and oil. Using conventional logs and pyrolysis data, several log-core regressions, such as delta log R, density, and uranium, can be derived to predict total organic carbon (TOC). The TOC can be used in conjunction with geochemical elemental measurements for a more accurate assessment of the formation kerogen and mineralogy, as well as hydrocarbon volumes. Nuclear magnetic resonance (NMR) porosity measures an apparent total porosity in the organic shale plays, measuring only the fluids present and excludes the kerogen. The complex refractive index method (CRIM) in conjunction with the mineralogy log data can be used to compute accurate dielectric porosities, which exclude both kerogen and hydrocarbon. Integrating the core TOC, predicted TOC, mineral analysis, NMR, and dielectric information, a final verification of the kerogen volume, hydrocarbon content, and mineral analysis can be assessed. This paper will describe the integration of conventional logs, a geochemical log, an NMR log, and dielectric to predict TOC, kerogen volume, and hydrocarbon volume, as well as, total porosity and mineralogy. The data is compared to the actual core data from three Eagle Ford wells, and it will be shown how the proposed approach will eliminate some coring operations. Finally, it will be shown how these interpretation results can be rolled up to make decisions on where to drill the lateral.

A new geochemical logging tool has been developed to provide measurements of elements important to the mineralogical analysis of rock formations and geochemical stratigraphy. The measurement system uses a chemical americium-beryllium neutron source to introduce moderate energy neutrons into the surrounding environment. These neutrons lose energy through scattering reactions and are eventually absorbed by materials in the borehole, the formation and the tool itself. The scattering and absorption reactions cause various elements among the atoms in the surrounding materials to emit gamma rays which have characteristic energies. Emitted gamma ray spectra are detected and recorded using a large bismuth germanate scintillation detector coupled to a high-efficiency photomultiplier and pulse-height analyzer.The recorded 256-channel spectra are analyzed to derive individual contributions, or elemental yields, arising from neutron capture reactions involving hydrogen, silicon, calcium, iron, sulfur, potassium, titanium, gadolinium, magnesium, aluminum and chlorine. The spectral analysis is performed with a constrained, weighted least-squares solver to ensure physically meaningful solutions are obtained for the elemental yields. Additional signals from inelastic neutron reactions with oxygen, carbon, silicon, calcium, and iron are treated separately in the spectral fitting process.The paper discusses several subjects including basic operation of the tool, derivation of characteristic elemental spectral responses, processing of the recorded spectra to obtain elemental yields and the conversion to elemental weight fractions for downstream formation mineralogy evaluations. Example logs from initial field tests are presented to demonstrate tool performance under various borehole conditions and lithologic environments including sandstone and carbonate reservoirs as well as organic shales.

Accurate quantification of total organic carbon (TOC) is an important step in evaluating log data in organic-rich reservoirs. The literature describes many log-based approaches for predicting TOC that have been introduced over the years, including the use of uranium content or GR linear regression, bulk density, the DeltaLogR approach, neural network approach, and a response equation-based method using sonic, density, and resistivity logs. All of the approaches require core-to-log calibration for validation. Each of these techniques involves assumptions for them to be valid, and, in a given instance, it is possible some techniques will not produce reliable results. However, good log-based TOC quantifications can be achieved by taking the median average of TOC estimates from several indicators. Many shale reservoirs contain 10 wt% pyrite and total organic carbon (TOC), which translates to 7% pyrite and 20% kerogen by volume. High volumetric percentages of pyrite and kerogen significantly affect the rock grain density. In low-porosity shale reservoirs, each 0.02 g/cm3 error in grain density produces approximately 1 p.u. error in porosity. Pyrite is commonly present in organic-rich shale intervals of shale gas formations because of the reducing conditions that enhanced organic matter preservation, and it may play a role in decreased resistivity response if the volume is sufficient. Consequently, in shale reservoirs, any method of predicting TOC using resistivity logs, such as DeltaLogR, should also consider the presence of pyrite. Similarly, TOC predictions based on bulk-density logs may also be sensitive to elevated pyrite concentrations. The link between pyrite presence and the depositional environment for many organic-rich shale reservoirs suggests that pyrite and sulfur may be useful TOC indicators in some situations. This paper examines the possible application of pyrite and sulfur for predicting TOC in shale reservoirs, such as in the Haynesville shale reservoir, but results should be applicable to many other shale reservoirs. An interesting result is that, although it may be possible to calibrate a TOC-based pyrite indicator for individual wells, the calibration is not universally applicable.