Reservoir Applications of Dipmeter Logs
- Jean-Pierre Delhomme (Etudes et Productions Schlumberger) | Thierry Pilenko (Services Techniques Schlumberger) | Etienne Cheruvier (Schlumberger Technical Services) | Richard Cull (Schlumberger Technical Services)
- Document ID
- Society of Petroleum Engineers
- Journal of Petroleum Technology
- Publication Date
- February 1988
- Document Type
- Journal Paper
- 180 - 186
- 1988. Society of Petroleum Engineers
- 5.6.1 Open hole/cased hole log analysis, 1.14 Casing and Cementing, 2.4.3 Sand/Solids Control, 1.6.9 Coring, Fishing, 2.2.2 Perforating, 4.3.4 Scale, 1.2.3 Rock properties
- 1 in the last 30 days
- 274 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 12.00|
|SPE Non-Member Price:||USD 35.00|
Summary. Reservoir quality indicators can be derived from the analysis of high-resolution dipmeter records. Changes in the aspect of the dipmeter curves respond primarily to changes in rock texture and sedimentary structure; their analysis can be instrumental in describing the petrophysical and reservoir properties of the formations. A four-step methodology has been set up: (1) extract dipmeter curve attributes, (2) combine dipmeter-derived data with other logs, (3) compare with core data over key intervals, and (4) extrapolate to other intervals and other wells. This approach is illustrated here by three examples.
Dipmeter tools are run primarily to get values of the structural dip. This information allows a better definition of the reservoir geometry and a more precise location of faults. The dipmeter conductivity curves, however, carry other information of importance for reservoir engineers. They show the textural changes, the thin shale beds, and the open fractures that control the flow within the reservoirs. The role of the dipmeter data in fracture detection has been described. We will concentrate here on sonic nonclassical uses of dipmeter data in shaly sandstone reservoirs. Such fine features as shale streaks can be seen on the highresolution dipmeter microconductivity profiles, enabling one to distinguish between laminated and dispersed clays and even to determine the thickness of thin, intercalated shale and sandstone layers. All this reservoir-oriented information can be derived from dipmeter logs and is now synthetized and carried by the so-called synthetic logs. The SYNDIP(TM) technique is the logical continuation of the algorithms that have been developed for fine-scale dip computation during the past decade. Since the early developments, numerous applications to reservoir studies have shown how the synthetic logs enable a better integration of dipmeter data with other logs and with core data. The emphasis here will be on such methodological aspects, and on examples in rock-structure identification and reservoir-unit delineation, including well-to-well correlation issues.
Dipmeter Synthetic Logs
There are two approaches to dip computation: interval correlation and feature matching. They share the same basic principle--to compute dips from depth shifts between dipmeter curves--but they differ in the way shifts are computed. Only feature matching allows the extraction of precise information about reservoir quality. In short, a bed boundary cutting across the borehole causes conductivity changes to be recorded at different depths by the eight (or four) dipmeter buttons; the depth differences or shifts between curves depend on the dip magnitude, the direction of the bedding surface, and the deviation of the borehole. Dip computation techniques based on statistical correlations assume the consistency of such shifts over fixed-length (typically 1-m [4-ft] -long) intervals; they produce dip results that are representative of intervals. On the contrary, the techniques based on fine-scale feature matching aim at determining the dip of each individual bed boundary. In such programs as GEODIP(TM) or DUALDIP(TM), each individual feature of a curve (i.e., a peak or a trough bounded by inflection points) is considered as the signature of a geologic object (layer, lens, pebble, etc.). Therefore, the features of each dipmeter curve are first extracted and then linked, when possible, to the corresponding features of the other curves by a correlation line; the dip computation only follows. In this second approach, all the information is available for computing such synthetic logs as curve activity and density of correlation, high-resolution resistivity profile, or sand/shale thickness logs. Fig. 1 illustrates how the activity of the dipmeter curves and their degree of similarity over a given interval are first characterized. Breakpoints in the dipmeter curves are defined as the inflection points where the absolute value of the curve derivative is above a given threshold. From the apparent thickness on each curve between such breaks (denoted as ATBR), the frequency of breaks (FBR) is derived as an index of curve activity. Not all existing breaks correspond to features that cross the borehole and are connected by a correlation line; the density of correlation lines (DCL) is computed as a curve-similarity index from the apparent thickness (ATCL) between consecutive correlation lines. One can identify the rock structure (or bedding type) over any interval by comparing dipmeter curve activity and density of correlation. This is illustrated in Fig. 1. The hatched, dark-gray area corresponds to correlated activity: a high curve activity (FBR) with a high DCL is the signature of laminations. The area with a bubblelike shading corresponds to uncorrelated activity: in front of hetero-geneities, FBR is high but DCL is low. Note that the reverse, a high DCL with a low FBR, is simply nonsensical. The number of correlated features cannot exceed the number of features. Where both FBR and DCL are less than three per meter [one per foot], the interval can be seen as massive. A high-resolution resistivity profile is also reconstructed from the dipmeter records, it indicates the value of resistivity read by the dipmeter in each individual layer. The eight button readings are first compared over intervals and equalized: possible biases related to imperfect pad contact or to lateral variation of mudcake thickness are detected and eliminated. Then, because the total amount of electrical current (Emex) sent into the formation varies vs. depth, the data are corrected for these Emex variations and calibrated. Last, within each bed, the eight resistivity profiles are compared along the direction of bedding, and statistics (i.e, median, maximum, and minimum) are computed at each level. The median value can be considered as representative of the overall formation resistivity: when the maximum or minimum value differs significantly from the median, this highlights the presence of a resistive (e.g., pebble, cemented nodules, or shell fragment) or conductive (e.g., clay lens) anomaly. In Fig. 2, the median resistivity curve is presented as a solid heavy curve in a logarithmic scale, ranging from 0.2 to 2,000 omegam. The two light curves are the minimum and maximum resistivities. Note the sharpness of the curves. No blurring is introduced at bed boundaries, and this holds true even in the case of high apparent dips because the statistics are calculated along the dipping planes. Thus, full advantage is taken of having eight directional measurements and a high vertical resolution simultaneously.
|File Size||469 KB||Number of Pages||7|