The porosity of a reservoir and its water saturation fraction have long been at the heart of wireline and MWD log based formation evaluation. In many formations, however, the complexity of the mineralogy and the properties of the formation's conductivity have only permitted qualitative answers. This has even been true with evaluations based on multi-log data suites. At the 1996 SPE Conference a new NMR logging process, the MRIL®* C/TP system, was introduced. C/TP provides a measure of total porosity independent of mineralogy and dependent only on the hydrogen density of the fluids themselves. This represents a significant advantage over the conventional density, neutron, and acoustic 'porosity' logs, which mostly respond to rock properties. Furthermore, the C/TP amplitude can be subdivided into pore size groups associated with clay mineral bound water, capillary bound water and pores free to accumulate hydrocarbon (FFI). Usage of the MRIL has proven the benefit of a borehole-centralized log, a feature that makes it largely independent of borehole size and shape. It has also become recognized as the first porosity log that truly measures the pore system and not some inference of porosity based on the rock matrix. Thus the log analyst is provided with a wireline log that is truly a porosity log which provides a quantitative measure of total and effective porosity. To investigate the accuracy and reliability of these methods and ideas, well bore data are assembled and compared to conventional log and core results. A full description of these comparisons and processes are included to showthe benefit of a true porosity measurement that is insensitive to matrix mineralogy and only depend on pore size and fluid contents,the advantage of a porosity measurement largely independent of borehole condition. Introduction Formation porosity is essential information to explore and exploit hydrocarbon reservoirs successfully. Until today, a combination of well logs (i.e., neutron porosity and density and, in fewer cases, sonic) is used to obtain porosity information. However, bad data quality or, logs missing at all, often affect the accuracy of the porosity results. Another potential source of inaccuracy is the lithology-dependence of the models used to interpret the individual logs and their combinations. Since its commercial introduction in 1991,1,2 NUMAR's MRIL tool has progressively made in-roads into these issues. It began by providing the first well log porosity measure independent of mineralogy, though limited to measurement of non-clay pore sizes.3,4 Also provided by the MRIL was the ability to separate the porosity associated with capillary bound water from the movable fluid volume, commonly referred to as the FFI pore volume.2 Later, further improvements in logging system design opened the possibility to determine fluid volume, fluid type and, to estimate diffusion properties directly from MRIL data.5 In 1996, the ability to measure the small, clay size, pores previously missed was added, opening the way to formation evaluation based on gradient-field NMR.6 To quantify the capability of the MRIL's porosity methods, and to compare the gradient-field NMR based formation evaluation with the conventional approach, a series of studies of core and log data were made.

The first Magnetic Resonance While-Drilling (MRWD) tools have been built and field tested. The hardware was designed to provide data compatible with the Magnetic Resonance Imaging Log (MRIL®) wireline tool in terms of processing and interpretation. The paper discusses the theory of robust MRIL measurements in the drilling environment and reviews the theory of T1 measurements. We report on the results of a field test in the Gulf of Mexico. The MRWD device logged the entire TD run from casing to TD and provided additional hydrocarbon-typing data in wiping mode over the target zones. Wireline MRIL data and conventional logs were used to verify the MRWD measurements. Introduction We previously reported on our initial field experiences with the Magnetic Resonance While-Drilling (MRWD) tool in a recent transaction paper.1 The present paper focuses on the petrophysical deliverables of the MRWD measurement and their robustness in the drilling environment. Furthermore, we report on a field test conducted in the Gulf of Mexico in March 2000. The test was run with an experimental prototype version (EX) and we were able to compare the results with wireline data from the MRIL-Prime tool.2 The comparison indicates satisfactory performance both during drilling and wiping operations. The primary application areas of MRWD will be high-cost offshore exploration and development wells, potentially in deep water. Typical bit sizes for the final TD run are 8–1/2 to 10–5/8 in. High rates of penetration (ROP) and long bit runs are common, aided by poorly cemented sediments, oil-based mud systems and advanced bit technologies. In this context, open-hole logging operations are expensive in terms of added rig time and risky in terms of hole stability. Additional problems are highly laminated reservoirs (e.g. turbidites) that frustrate traditional log interpretation because electrical and nuclear measurements may show little or no contrast identifying reservoir rocks. In these situations, breakthroughs in reservoir characterization have been achieved by computing hydrocarbon volumes directly from MRIL data, circumventing the problems with conductivity/resistivity.3 Real-time MRIL data from an LWD device will result in faster, better and cheaper evaluation and exploitation of these reserves. Logging-while-drilling (LWD) tools are expected to replicate as close as possible the functions of their wireline counterparts. In particular, the expectations for the LWD version of an MRIL device are:To provide a measure of rock porosity that is lithology-independent and that does not require radioactive sources;To collect a spectrum of NMR relaxation times that is suitable for input to lithology models that estimate bound fluid volumes, free fluid volumes and rock permeability;To enable fluid typing by exploiting the inherent differences in T1 and/or T2 (longitudinal and transversal relaxation times) between the water, oil and gas phases;To withstand the shock, vibration and erosion associated with drilling;Not to interfere with the drilling process; andNot to interfere with any other LWD or MWD measurements. Tool Hardware and Sensor Physics Fig. 1 shows the basic tool layout conforming to these requirements. The tool has two main sections: the sensor and the electronics package. The total length of the EX version tool of 42 ft is driven by the inclusion of extra electronics and will be reduced in the forthcoming commercial tool. The tool diameter is under-gauge at about 7–1/4 in. All compressional, torsional and tensional stress ratings exceed those of standard 4½ IF API connections. To meet the sensor-to-sensor non-interference requirement, a distance of 45 ft to the MWD directional package is recommended.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractThe porosity of a reservoir and its water saturation fraction have long been at the heart of wireline and MWD log based formation evaluation. In many formations, however, the complexity of the mineralogy and the properties of the formation's conductivity have only permitted qualitative answers. This has even been true with evaluations based on multi-log data suites.At the 1996 SPE Conference a new NMR logging process, the MRIL ®• C/TP system, was introduced. C/TP provides a measure of total porosity independent of mineralogy and dependent only on the hydrogen density of the fluids themselves.This represents a significant advantage over the conventional density, neutron, and acoustic 'porosity' logs, which mostly respond to rock properties. Furthermore, the C/TP amplitude can be subdivided into pore size groups associated with clay mineral bound water, capillary bound water and pores free to accumulate hydrocarbon (FFI).Usage of the MRIL has proven the benefit of a boreholecentralized log, a feature that makes it largely independent of borehole size and shape. It has also become recognized as the first porosity log that truly measures the pore system and not some inference of porosity based on the rock matrix.Thus the log analyst is provided with a wireline log that is truly a porosity log which provides a quantitative measure of total and effective porosity.To investigate the accuracy and reliability of these methods and ideas, well bore data are assembled and compared to conventional log and core results. A full description of these comparisons and processes are included to show • MRIL is a registered trademark of NUMAR Corp.• the benefit of a true porosity measurement that is insensitive to matrix mineralogy and only depend on pore size and fluid contents, • the advantage of a porosity measurement largely independent of borehole condition.

A recently developed nuclear magnetic resonance (NMR) downhole fluid analyzer provides in-situ fluid characteristics at reservoir PVT conditions. The device is part of the Reservoir Description Tool (RDT*), a modular formation sampler and tester. This paper presents the theory and the experimental evidence that led to the use of NMR as a means to assess the amount of mud filtrate contained in the pumped formation fluid. In particular, the problem of differentiating oil base mud filtrate from connate hydrocarbons is difficult and has not been addressed by conventional measurements. NMR relaxometry uses the inherent differences in T1 time constant values and distributions to make these distinctions. The T1 data is available in real time and can be used to determine the optimum pump-out rates and durations while sampling reservoir fluids. Introduction Advances in wireline formation fluid sampling allow the recovery of representative hydrocarbon samples. It has become imperative to obtain high-quality samples with minimum mud filtrate content, while simultaneously minimizing the pumpout time required per station. Traditionally, contrasts in resistivity and/or dielectric properties between connate fluids and contaminants have been exploited. These methods fail if neither phase contains water. The problem of differentiating oil base mud (OBM) filtrates from crude hydrocarbons has received considerable attention. So far, only optical analysis methods have been described to address the problem. It has long been known in NMR wireline logging that crude oils and oils introduced from the borehole exhibit different nuclear relaxation time behavior.1 This is readily understood given that the filtrates typically have lower viscosity and therefore slower relaxation rates. The mixture of hydrocarbon chains of different lengths and mobility in crudes gives rise to a spectrum of relaxation times, spanning 1–2 orders of magnitude. Filtrates, on the other hand, are characterized by a single T1 relaxation time. With the advent of a downhole NMR analysis system2 it has become possible to exploit the differences in relaxation behavior in a systematic and quantitative fashion. To this end, we studied the NMR characteristics of a range of crude oils, of typical mud base oils and of their mixtures. We found that with reasonable measurement times, contaminant levels down to 10% can be detected. The downhole NMR fluid analyzer fulfills the signal-to-noise and measurement speed requirements to perform oil-v.-filtrate differentiation as part of an overall NMR fluid analysis. The Downhole NMR Fluid Analyzer The analyzer is part of the RDT (Reservoir Description Tool).3 It continuously analyzes in real time the fluids being pumped from the formation. The main measurement is relaxometry by hydrogen nuclear magnetic resonance. Protons are being brought in magnetic resonance by a combination of a static 1,000-gauss magnetic field and radiofrequency (r.f.) pulses tuned to 4.2 MHz. There are a number of measurement outputs:Hydrogen index. This is the hydrogen density compared to a reference fluid. The reference fluid is water at room temperature and atmospheric conditions. With few exceptions, oils have hydrogen indices between 0.8 and 1.Relaxation time T1. The bulk relaxation time constant reflects the mobility of the fluid's molecular components. The higher the mobility, the higher the macroscopic T1, and vice versa. Molecular mobility entails both in-place reorientation and Brownian motion.

Summary The first Magnetic-Resonance-Image Logging-While-Drilling (MRI-LWD*) tools have been built and field tested. The hardware was designed to provide data that are compatible with the Magnetic Resonance Imaging Log (MRIL®) wireline tool in terms of processing and interpretation. This paper discusses the theory of robust MRIL measurements in the drilling environment and reviews the theory of 7 relaxation time measurements. We report on the results of a field test in the Gulf of Mexico. The MRI-LWD device logged the entire total depth (TD) run from casing to TD and provided additional hydrocarbon-typing data in wiping mode over the target zones. Wireline MRIL data and conventional logs were used to verify the MRI-LWD measurements.