Estimation of static and dynamic petrophysical properties of multi-layer hydrocarbon reservoirs is crucial for the assessment of storage and flow capacities, compartmentalization, and for best primary or enhanced recovery practices. Interactive numerical simulation to reproduce field logs and core data is a reliable procedure to estimate static and dynamic petrophysical properties of complex rock formations. Previously, Voss et al. (2009) introduced the concept of Common Stratigraphic Framework (CSF) to construct and cross-validate multi-layer static/dynamic petrophysical models by invoking the interactive, numerical simulation of well logs both before and after invasion. This thesis documents the successful implementation of the CSF concept to examine and quantify the effects of mud-filtrate invasion on apparent resistivity, nuclear, and magnetic resonance logs acquired in San Martin, Cashiriari and Pagoreni gas fields in Camisea, Peru. Conventional petrophysical interpretation methods yield abnormally high estimates of water saturation in some of the reservoir units that produce gas with null water influx. This anomalous behavior is due to relatively low values of deep apparent electrical resistivity, and has otherwise been attributed to the presence of clay-coating grains and/or electrically conductive grain minerals. On the other hand, electrical resistivity logs exhibit substantial invasion effects as evidenced by the separation of apparent resistivity logs (both LWD and wireline) with multiple radial lengths of investigation. In extreme cases, apparent resistivity logs “stack” because of very deep invasion. We diagnose and quantify invasion effects on resistivity and nuclear logs with interactive numerical modeling before and after invasion. The assimilation of such effects in the interpretation consistently decreases previous estimates of water saturation to those of irreducible water saturation inferred from core data. It is shown that capillary pressure effects are responsible for the difference in separation of resistivity curves in some of the reservoir units. The final multi-layer CSF is in agreement with gas production measurements and permits reliable flow predictions to assist in reservoir engineering and production studies. / text

Nuclear magnetic resonance (NMR) has become an effective borehole measurement option to assess petrophysical and fluid properties of porous and permeable rocks. In the case of fluid typing, two-dimensional (2D) NMR interpretation techniques have advantages over conventional one-dimensional (1D) interpretation as they provide additional discriminatory information about saturating fluids and their properties. However, often there is ambiguity as to whether fluids detected with NMR measurements are mobile or residual. In some instances, rapid vertical variations of rock properties (e.g. across thinly-bedded formations) can make it difficult to separate NMR fluid signatures from those due to pore-size distributions. There are also cases where conventional fluid identification methods based on resistivity and nuclear logs indicate dominant presence of water while NMR measurements indicate presence of water, hydrocarbon, and mud filtrate. In such cases, it is important to ascertain whether existing hydrocarbons are residual or mobile. The radial lengths of investigation of resistivity, nuclear, and NMR measurements are very different, with NMR measurements being the shallowest sensing. Even in the case of several radial zones of NMR response attributed to different acquisition frequencies and DC magnetic field gradients, the measured signal originates from a fairly shallow radial zone compared to that of nuclear and resistivity logs. Depending on drilling mud being used and the radial extent of mud-filtrate invasion, the NMR response of virgin reservoir fluids can be masked by mud filtrate because of fluid displacement and mixing. In order to separate those effects, it is important to reconcile NMR measurements with electrical and nuclear logs for improved assessment of porosity and mobile hydrocarbon saturation. Previously, Voss et al. (2009) and Gandhi et al. (2010) introduced the concept of Common Stratigraphic Framework (CSF) to construct and validate multi-layer static and dynamic petrophysical models based on the numerical simulation of well logs. In this thesis, the concept of CSF is implemented to reconcile 2D NMR interpretations with multi-layer static and dynamic petrophysical models. It is found that quantifying the exact radial zone of response and corresponding fluid saturations can only be accomplished with studies of mud-filtrate invasion that honor available resistivity and nuclear logs. This thesis indicates that the two interpretation methods complement each other and when applied in conjunction, improve and refine the overall petrophysical understanding of permeable rock formations. Examples of successful application include field data acquired in thinly-bedded gas formations invaded with water-base mud, where bed-boundary effects are significant and residual hydrocarbon saturation is relatively high. In such cases, numerical simulation of mud-filtrate invasion and well logs acquired after invasion enables reliable interpretations of petrophysical and fluid properties. The interpretation procedure introduced in this thesis also provides an explicit way to determine the uncertainty of petrophysical and fluid interpretations. / text