Reservoir Characterization

Heterogeneity quantification, fine-scale layering derived from image logs and cores

Fine-scale layering results in an improved static geological model that, in turn, improves reservoir simulation in a large oil field offshore Abu Dhabi.

Hamad Bu Al-Rougha and Hesham Shebl, ZADCO; and Sandeep Chakravorty, Schlumberger

Most carbonate reservoirs offshore Abu Dhabi are characterized by complex textural heterogeneity that can be detected by high-resolution borehole image logs. This textural heterogeneity gives rise to extreme permeability variation that is the controlling factor in reservoir production. A study conducted by ZADCO and Schlumberger illustrates an approach to improve carbonate reservoir characterization of an offshore Abu Dhabi oil field, by quantifying fine-scale heterogeneity using borehole image logs and core measurements.

The reservoir under study ranges from mud-supported to grain-supported limestones with variable degrees of grain sorting and cementation. Detailed core analysis indicates the presence of small-scale, fining-upward cycles of 0.5 – 5 ft thick. The base of each cycle is coarse-grained sediments with high permeability that can directly affect field development plans. An attempt has been made to detect this cyclicity from image logs to aid correlations in uncored wells.

The study illustrates a methodology to quantify vuggy and cemented textures of the carbonate rocks at a resolution that is closest to actual core scale, and to use image logs to better identify stratigraphic surfaces and vertical staking patterns in the carbonate reservoirs. The fine-scale layering identification resulted in an improved static geological model that improves reservoir simulation, because it is constrained by better stratigraphic architecture of the reservoir and, therefore, its matrix-determined fluid flow structure.

LOCATION/ SCOPE OF STUDY

The study area is 65 km (40 mi) offshore Abu Dhabi, in one of the largest producing fields, covering more than 1,600 sq km. More than 1,000 wells have been drilled into the field’s Lower Cretaceous reservoir. Variable production and injection performances indicate extreme permeability variations. The study illustrates an approach to improve carbonate reservoir characterization by quantifying small-scale heterogeneity, using borehole image logs and core measurements with a view to minimizing coring in future wells.

DATA SET

The study is based on detailed analysis of six wells, four from the eastern area and two from the western area, Fig. 1. All have excellent core coverage over the entire reservoir.

Fig. 1. Field location and wells under study.

The core examination concentrated on recording key lithology changes, including coarse grainstones, rudist accumulations, stylolite locations, dolomite bands and dolomitic-cemented grainstones. This was followed by a thin-section study for reservoir quality and digenetic petrographic studies on different reservoir rock types. Additionally, permeability and porosity were measured from core plugs sampled per foot. The core analysis was later loaded into a workstation to be integrated with borehole electrical image logs. Results were validated by dynamic measurement and simulation.

RESERVOIR GEOLOGY

The Lower Cretaceous reservoir was deposited within a regionally extensive, low-energy, shallow-water carbonate ramp on the Arabian cratonic margin, Fig. 2. Despite widespread lateral lithofacies continuity, reservoir quality varies from superior in the field’s eastern part to inferior (more cemented) in the western area. This regional variability is locally complicated by cyclical, minor (low-amplitude) eustatic cycles in an overall, relative sea level rise. This resulted in deposition of a series of parasequence sets in a major, high stand systems tract (HST). The upward shallowing HST is capped by deepening-upward, transgressive system tract (TST) shales.

Fig. 2. Carbonates were deposited on a ramp with a wide facies belt. Carbonate ramps act as a good record of relative sea level change, since even minor fluctuations cause significant facies belt migration and marked para-sequence stacking patterns in the reservoir.

The best reservoir intervals occur in the uppermost parts of HSTs. Extreme variations in reservoir productivity result from wide-ranging differences in facies, texture, fauna and flora. The complex lateral and vertical stacking pattern has implications for well trajectory optimization, perforation and completion strategies, and well injection and production planning.

PARASEQUENCE CYCLICITY

Cyclicity patterns in the study area occur at a variety of scales, some of which are appropriate for reservoir layering, while others are either too fine or too gross-scaled. Consequently, the definition of sequence, stratigraphic-based reservoir layers involves a two-step process: 1) Identification of the smallest-scale shallowing- and deepening-upward cycles in the cores and image logs; and 2) Consolidation of these small-scale cycles into geological reservoir layers that can be confidently correlated throughout the field, Fig. 3.

Fig. 3. Schematic representation of TH fining-upward grainstone stacking pattern and conceptual permeability profile.

Recognition of cycles and cycle boundaries is based on the following criteria:

• Predictable vertical facies stacking patterns (e.g., coarsening or fining-upward trends)
• Marked changes of grain size across irregular, burrowed surfaces (firm grounds)
• Lithofacies changes across burrowed and incipiently cemented surfaces (hard grounds).

Core-based shallowing and/or deepening-upward cycles can be matched to the image logs to define the smallest scale cycles resolvable on the wireline logs. High-resolution logs can often identify subtle changes in rock fabric and reservoir properties that are not resolved on conventional logs.

Furthermore, the relative number of cycles will vary, depending upon the position of each well on the field structure. Wells drilled on the structure’s flanks are likely to contain more cycles than those drilled on the crest, due to changes in relative sea level more profoundly affecting the middle to outer ramp. On the inner ramp, lithofacies variation in response to sea level fluctuations might be expected to be more modest.

IMAGE LOGS METHODOLOGY

The estimation of reservoir layering using conventional wireline logs is possible, although low vertical resolution limits their utility for characterizing fine-scale heterogeneity that exists in most carbonate reservoirs. For example, conventional wireline logs have vertical resolution capabilities (2 – 4 ft or greater) that commonly are much more gross compared to the study area’s scale of geological heterogeneity. Any geological heterogeneity below the vertical resolution of a conventional wireline tool has gone unrecognized in uncored wells. Ideally, borehole image logs are well suited for such a purpose.

For this study’s purpose, the logging tool’s vertical resolution is more important than the depth of investigation. In this application, small-scale geological heterogeneity in the invaded wellbore region of the formation – reflected in microresistivity measurements from image logs – provides valuable information on reservoir heterogeneity and, ultimately, on reservoir quality.

Heterogeneity quantification and fine-scale cyclicity identification from image logs is a two-step process.

Image scaling. Despite providing great borehole coverage and high resolution, the measurements of all resistivity-imaging tools are relative. Consequently, a layer’s true resistivity cannot be quantified without special processing techniques. To perform quantitative analysis, the image log button response is transformed to a measurement resembling the formation resistivity as measured by a true resistivity tool. This is achieved through scaling, a process that calibrates the button response using a cross-plot of the average image button response against shallow resistivity, Fig. 4.

Fig. 4. Cross-plot of average button response from FMI* (Fullbore Formation MicroImager) against focused resistivity log. The output SRES curve (green color, right-hand track) is a calibrated log generated from the cross-plot. The overlap of SRES and an external resistivity (HLLS; blue in color, same track) shows the nature of correlation. This relationship is used to scale images for mud effects and perform quantitative analysis.

In this study, shallow, high-resolution lateral log resistivity is used for image scaling as follows:

• Computation of average current from 192 electrodes
• Depth matching of shallow lateral log with the computed average current
• Computation of scaling parameters of the average current, cross-plotted against theoretical current derived from a lateral log
• Computation of the scaled image array and the high-resolution curve from the scaling transformation, which should match the resistivity curve used in the calibration process.

Reservoir heterogeneity. The scaled image is then contoured by applying a threshold on resistive and conductive anomalies frequency histograms to capture reservoir heterogeneity. Several iterations are performed in the study wells to optimize parameters used for extraction of formation heterogeneities. Each iteration’s results are validated by visual examination of images and cores, to see whether all heterogeneities are identified and contoured. Conductive events are further classified into three types – large patches, connected spots and isolated (within the wellbore) spots. Resistive events are classified into resistive patches and resistive spots, Fig. 5.

Fig. 5. (A) Snapshot illustrating BorTex* texture classification principle. Once background conductivity is known, thresholds are set at various conductivity ratios. The medium- and high-contrast thresholds find conductive and resistive inclusions. (B) FMI log shows BorTex interpretation drawn onto the images and comparison with core. Conductive inclusions are essentially a representation of vugs and pore space (permeable areas). Resistive inclusions represent cemented or tight areas.

Results of heterogeneity contour analysis are converted into continuous depth-indexed channels or curves to get the proportion and size of each type of formation heterogeneity. Such curves can be averaged over any window length and output at any sampling rate. In this study, they are averaged over a 6-in. window and output at 0.2 in. to capture fine-scale, core textural variability.

To detect such a fine, complex sedimentary hierarchy, a cumulative conductivity curve (conductive inclusion proportion, the sum of conductive isolated spots, conductive connected spots, and conductive patches) and a cumulative resistivity curve (resistive inclusion proportion, the sum of resistive spots and patches) were extracted from image logs. Since conductive and resistive heterogeneities in the borehole image represent vuggy and cemented zones, respectively, an overlay of the image conductive inclusion proportion and resistive inclusion proportion curves was used to detect fine-scale cyclicity.

Electrical image data showed repetition of very fine conductive and resistive layers within reservoir units, which agreed with core examinations. Each conductive (grainstone) cycle is capped by a high-resistivity (cemented packstone) layer, Fig. 6. The resistive layers represent hard ground that can be used in assisting interwell correlations in the absence of core data. Major erosional surfaces are characterized by a sharp change in conductive inclusion and resistive inclusion types of proportion curves.

Fig. 6. Prediction of cyclicity from FMI images and comparison with cores.

DISCUSSIONS/ OBSERVATIONS

The major objective for the fine-scale layer mapping effort has been to understand complex permeability heterogeneity. Core and thin-section observations show that the highest core permeability is related to grain size, sorting and degree of cementation, and linked to lithotype distribution.

Very good correspondence of grainstone occurrences with higher image conductivity and core plug permeability peaks was observed. Several discrete grainstone/ higher permeability levels were identified and successfully correlated, despite significant lateral variations in both thickness (<1 to 5 ft) and permeability. Each grainstone layer tends to fine upward; the upper section is generally packstone or fine grainstone with lower permeability, Fig. 7. In many wells, the permeability profile is very subdued, especially downflank and to the west. Nevertheless, it has been possible to use even these subtle permeability data for correlation.

Fig. 7. Intermediate cyclicity seen on the core permeability and correlatable with FMI-derived, true stratigraphic thickness of gross resistivity and conductivity changes.

Plotting the image conductivity profile, along with core sedimentary cycles, shows higher permeability values near the base of the grainstone cycles and gradual reductions toward the top (cemented packstone facies). This relationship may not be clear, if core plug measurements are random and permeability is derived from low-resolution logs. The highest permeability layers compare well with the pattern of the well test permeability and saturation profile, suggesting a strong primary lithological (grainstone) control on well productivity.

Borehole imaging was not able to resolve cycles with high confidence in areas where cycles are very thin, and/or the contact between grainstone and packstone facies is characterized by gradual texture changes. This limitation is occasionally seen in the crestal area, where the grainstone facies are either poorly sorted (due to bioturbation) and/or overlaying packstone facies are not well cemented.

CONCLUSIONS

The image-derived curves successfully identified small-scale, fining-upward cycles seen on cores. The base of each such cycle represents coarse-grained sediments with high permeability, capped by cemented fine-grained sediments. This can lead to recognition of fine-scale cyclicity even in the non-cored wells, to map the areal extent of these facies, and, hence, their field-wise distribution. High-resolution image logs are the only means to achieve this objective.

Recognition of high-permeability grainstone facies is crucial to understanding the water injection performance of such a mature field. 

THE AUTHORS
	Hamad Bu Al-Rougha is a team leader with Zakum Development Company (ZADCO) in Abu Dhabi. He graduated with a BS degree in geology from U.A.E. University in 1990, joining ZADCO that same year as an operations geologist. Mr. Al-Rougha was later assigned to BP’s Sunbury Research Center on the Reservoir Characterization Study of Upper Thamama reservoirs, concentrating most of his work on cretaceous carbonate in the U.A.E.
	Hesham Shebl is a sedimentologist with ZADCO in the Geological Services and Support Division. Prior to joining ZADCO in 2000, he was with Core Laboratories as a senior petrographer/ sedimentologist from 1993 to 2000. Mr. Shebl received his BS degree from Ain Shames University, Cairo, in 1980 and an MS degree from King Fahd University of Petroleum and Minerals, Dhahran, in 1989. He is a member of AAPG, SPE and SCA, and a director member of Emirates Society of Geoscience.
	Sandeep Chakravorty is a senior geologist with Schlumberger in Abu Dhabi. He joined Schlumberger in 1992 after graduating from Indian Institute of Technology with a master’s degree in applied geology. His main fields of specialization are borehole imaging, and structural and sedimentological interpretation of clastics and carbonates. His current area of interest is carbonate reservoir characterization. Mr. Chakravorty is a member of AAPG and the Emirates Society of Geoscience.