A New Multiband Generation of NMR Logging Tools
- M.G. Prammer (NUMAR) | J. Bouton (NUMAR) | E.D. Drack (NUMAR) | M.N. Miller (NUMAR) | R.N. Chandler (NUMAR)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Evaluation & Engineering
- Publication Date
- February 2001
- Document Type
- Journal Paper
- 59 - 63
- 2001. Society of Petroleum Engineers
- 4.1.9 Heavy Oil Upgrading, 5.6.1 Open hole/cased hole log analysis, 4.1.2 Separation and Treating, 5.2 Reservoir Fluid Dynamics, 5.1 Reservoir Characterisation, 6.1.5 Human Resources, Competence and Training, 1.11 Drilling Fluids and Materials, 4.1.5 Processing Equipment
- 2 in the last 30 days
- 284 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 10.00|
|SPE Non-Member Price:||USD 30.00|
This paper describes the hardware and operation of a new generation of nuclear magnetic resonance (NMR) logging tools. In the past, NMR required the logging engineer to consider the T1 relaxation times of the reservoir fluids likely to be encountered. Actual, or simply assumed, long T1's translated into slow logging speeds. The new tool generation overcomes this limitation. The key feature is that nine sensitive volumes are polarized in parallel and are read out in rapid sequence. A new sonde design speeds up the polarization process by a factor of 2. Each volume contributes equally to the result and can support identical measurements for rapid stacking and fast logging, and each can be used for individual, simultaneous measurements.
Laboratory data and field-test results are presented to demonstrate both the relative simplicity of operation and the improvement in data quality. Logging speeds typically can be upgraded by a factor of 4, while data for total porosity determination and fluid typing are acquired in a single logging pass.
Over the past few years, log analysts have become familiar with the potential and the limitations of NMR logging. Basically, an NMR tool reports the total number of hydrogen atoms that are in the liquid or gaseous state. As such, NMR is a lithology-independent porosity tool as long as the hydrogen index of the fluids can be estimated.
The commercial use of modern pulsed-NMR tools (NUMAR's MRIL1,***, and Schlumberger's CMR,2,**** brought two surprises:
The near-borehole zone, which was assumed to be flushed, can contain substantial amounts of native hydrocarbons, both oil and gas.
The T1 relaxation times of hydrocarbons (connate fluids and filtrate from oil-based muds) under reservoir conditions are substantially longer than previously assumed.
The consequence of these findings was that NMR began to be used as a hydrocarbon-detection and reservoir-quantification tool, at the expense of logging speed and wellsite efficiency.3,4
From the theory of nuclear spin relaxation in liquids by Bloembergen, Purcell, and Pound5 follows the proportionality of bulk relaxation time and self-diffusion coefficient: T1~D. The Stokes relationship between viscosity and correlation time stipulates that D~T/µ; therefore, we can expect that T1~T/µ over a certain range of temperatures. We conducted measurements of T1 and T2 in the 30 to 150°C range on oils used for oil-based mud synthesis.***** Some of our results are listed in Table 1. These oils are typically type C16/C18, with hydrogen indices close to that of water. The absence of longer chains or aromatics suggests short correlation times and long T1 relaxation times for Larmor frequencies in the low-MHz range. For all samples investigated, T1=T2. In general, our data confirm the expected temperature dependency of T1. Certain oils, however, including Oil B and Oil C in Table 1, show a sharp discontinuity at some point between 110 and 150°C. We have confirmed that no chemical change takes place in the oil because the original T1 can be restored by cooling the sample to room temperature and exposing it to the atmosphere. The most likely explanation is dissolved oxygen that becomes volatile above 110°C. Paramagnetic oxygen is a potent relaxation agent even at low concentrations, and its disappearance at high temperatures causes an additional increase in T1.
The surface interaction, which is responsible for rapid relaxation/polarization in the water phase, is inefficient for oil, even in cases where rock analysis would classify the rock as oil-wet. Gas is another example of high T1's (4-5 sec and more) caused by weak internal relaxation and nonexistent interaction with the rock surface.
T1 affects data acquisition and logging speed in a direct fashion:
The hydrogen atoms must be exposed to the polarizing magnetic field for a multiple of T1. A factor of 3 is considered minimum. Fig. 1 illustrates exponential polarization curves for T1's of 1 sec, 2 sec, and 4 sec. Note that 95% polarization is reached only after 12 sec for fluids with T1=4 sec.
The measurement itself is contaminated by thermal noise and must be repeated a few times to bring the influence of this noise down to acceptable levels. After each measurement, a full wait time (tw) of at least 3× T1 is required.
Assuming 8 repeats and T1=4 sec, we find that the wait times required for a single measurement add up to 8×4×3=96 sec. If a vertical resolution of 3 ft is acceptable, the NMR tool cannot move faster than 3×60/96˜2 ft/min. A speed limit of 120 ft/hr makes it impractical to deploy NMR on a routine basis over large openhole intervals.
An undesirable option is to forego full polarization. This mode is faster but results in data that are substantially harder to interpret in a quantitative fashion. Furthermore, this mode defeats the unique capability of NMR to detect hydrocarbons independent of resistivity contrast. It is highly desirable to use an NMR tool that is virtually free of T1 effects. Current NMR applications such as total and effective porosities, pore-size distribution, permeability modeling, hydrocarbon typing, and gas detection require that all hydrogen components are equally visible; i.e., even the slowest T1 component should be fully polarized. Furthermore, these applications should run at logging speeds of 1,000-1,500 ft/hr. Lastly, a higher level of automation should reduce the amount of job planning and setup procedures required today.
These requirements are met by the newest generation of MRIL tools, MRIL-Prime. The T1 problem is solved by using a large number of measurement volumes in parallel and by employing a new prepolarization scheme.
New Tool Features
The key feature of the new MRIL tool is the ability to rapidly polarize and to read out many identical measurement volumes. The scheme is illustrated in Fig. 2. There are nine tightly packed cylindrical shells, each 24 in. tall and each containing on average 750 mL. The tool electronics can rapidly switch back and forth between volumes by changing the operating frequency over a wide range. The magnetic field gradient translates lower operating frequencies into resonance conditions that occur farther away from the tool. The gradient is circularly symmetric, resulting in resonance shells around the tool. These shells are labeled A (innermost, diameter 14 in.) to J (outermost, diameter 16.5 in.). In an 8-in. borehole, these diameters correspond to a depth of investigation between 3 and 4 in. The individual volumes are completely separated such that concurrent measurements do not influence each other. The approximate field strength, resonance frequency, and magnetic field gradient for each measurement volume is listed in Table 2.
|File Size||235 KB||Number of Pages||5|