Webinar Q&A - Jorge Costa Gomes Answers Your Questions

There are many uncertainties pertaining to wireline log interpretation. To compute Sw, one has to predict the Archie parameters ‘m’ and ‘n’. A lot of the times interpreters use values of 2 for both. In reality these values vary with respect to pore fabrics (tortuosity) / lithofacies / wettability and should be corrected. Porosity from logs is measured between 1000 cm3 and 10,000 cm3 of formation (i.e. up to over 1000 times the core plug volume) (Rider, Malcolm H. 2008). Measuring points also vary between i.e. 0.3 – 1 ft intervals therefore there are a lot of uncertainties inherent to log interpretations.

The range of porosity varies between 10-35% throughout the model. The vertical shift between layers is associated with the geological characteristics found in and around the Bab Basin for the Upper Kharaib Member. Layers with maximum porosity and permeability are commonly found in the upper part of the formation associated with grain supported limestone fabrics. The reservoir is subdivided into 15 zones with 8 thin baffle zones (dense / stylolite intervals) having roughly a Net-to-Gross of 95%.

Q4 2020.

This is a synthetic model based on real data (but without any seismic cube). Our model has sonic and density log readings and therefore the acoustic impedances of the layers logged could potentially be calculated and a synthetic seismic petro-elastic model constructed. This however falls out of the scope of this work.

Schlumberger Petrel© and CMG IMEX©

The diagenetic overprints (improving or degrading petrophysical properties / dynamic performance) is manipulated by selecting a better or worse representative drainage capillary pressure curve per rock type.

Mercury Injection Capillary Pressure (MICP) uses mercury which is a non-wetting. The wettability characteristics of our rock types capture the full wettability envelope (oil – water wet characteristics). Rock types with large mean pore throat size distributions were assigned low irreducible water saturations and more oil-wet characteristics and vice versa.

The Early Aptian (Shu’aiba) is characterized by a long-term rise in sea level that had a profound influence on platform architecture and facies distribution. During the Late Aptian (Shu’aiba) there was a fall in sea level in which the intra-shelf basin became filled by a mixture of prograding pure carbonate wedges and organic-rich and/or argillaceous basinal deposits. The clinoforms prograde eastward (basinward). In Oman, in the Al Huqf sections much of the Lower Shu’aiba Platform has been eroded/not deposited. Please refer to “Barremian–Aptian Stratigraphy and Hydrocarbon Habitat of the Eastern Arabian Plate – Volume I & II” for further information (Buchem, Al-Husseini et al. 2010).

The capillary pressure driven reservoir rock typing technique used was initially distributed by computing the mean pore throat size distribution of the reservoir. That then enabled us to reverse engineer the facies distributions by grouping pore throat size ranges (via histogram) into geological facies. No seismic data or deterministic regional facies trend maps were used – however the overall petrophysical trends (via stratigraphic modified Lorenz plots (SMLP)) capture the generalized large-scale heterogeneities found in the region. Moreover, this synthetic case study used 100% core data giving us a strong understanding of the distribution of petrophysical properties.

This is a synthetic model based on real data (but without any seismic cube).

The first initiative of this work was to create a benchmark of a typical carbonate aggradational parasequence (Upper Kharaib Member). We have yet to model the progradation (Shu’aiba) and retrogradation characteristics of carbonate formations.

Our model currently only models single carbonate matrix porosity and permeability characteristics as there is enough geological uncertainty in all aspects leading to their construction. Fractures are currently out of the scope of this work but most defiantly will be incorporated in the future. Multiple well configurations / perforations scenarios were tested but we decided to keep all the wells vertical in a 5-spot pattern for simplicity.

No, single matrix porosity and permeability values only for now.

The synthetic representation of the aggradational parasequence modelled captures the large permeability contrast between B-Upper and B-Lower of the Upper Kharaib Member. Basinward the trends in petrophysical properties varies significantly – we tried to capture the large-scale heterogeneity found in and around the Bab Basin into one single geological model for reservoir simulation studies.

This is a synthetic model based on real data (but without any seismic cube).

The recovery factor is low in part due to the reservoir being very large (16 Billion bbl OOIP) and we only produced the synthetic production data for the first 25 years. A lot of residual oil saturation after waterflood (Sorw) is present for rock types with more oil-wet characteristics (causing trapped oil). Waterflooding is currently the only mechanism being used but EOR could improve the RF further. Moreover, the producer wells are currently 2 km apart and with local grid refinement more infill wells could be added to recover more oil. The producers were constrained to produce to a minimum bottom hole pressure of 300 psi above the bubble point pressure (which is quite conservative). By reducing the bottom hole pressure constraint this will in turn increase the RF / plateau of wells. In addition, horizontal wells could be drilled to increase the reservoir contact and improve the RF. In terms of producer wells, skin factors have an immediate effect on oil production and recovery (positively or negatively).

The logs used in this study were Gamma Ray (GR), Density (RHOB), Sonic (DT), Resistivity (RT) & Neutron Porosity (NPHI). However, the use of Compensated Neutron Log (CNL), Compensated Density Log (FDC), Formation Micro-Imager (FMI), Nuclear Magnetic Resonance (NMR) and Microresistivity logs (MSFL), just to name a few, could aid in better characterising carbonate reservoirs at large.

The most impactful geological uncertainty that would influence the recovery of oil in carbonate reservoirs I would say starts with the stratigraphic framework construction, followed by all aspects of petrophysical property modelling (along with diagenetic overprints). The manner in which one conducts his reservoir rock typing and all the dynamic characteristics of those rock types (Pc, kr, Swirr, Sorw etc.) all have an impact on the oil recovery behaviour. Furthermore, the selection of saturation height functions used per rock type will also impact on the OOIP. In terms of producer wells, skin factors have an immediate effect on oil production and recovery (positively or negatively). By accounting for geological uncertainty in reservoir modelling and by reconstructing multiple geological models (using the same dataset) using different modelling techniques and geologic interpretations one can then compare the production behaviour of our synthetic production data (from our undisclosed ‘truth case’) with the multiple geological releases which will be made public. We have yet to run history matching studies between the multiple ensembles of models to see the hierarchical effect geological uncertainty has on i.e. oil recovery factor, rates and pressures. We have yet to model dual porosity characteristics and fracture behaviours (expected to create even more realistic flow behaviour and oil recovery behavioural changes).

No vertical interference tests were used.

Schlumberger Petrel© and CMG IMEX©

The model was constructed using 100% core data therefore giving us very strong confidence in the distribution of petrophysical properties. The reverse engineering of the facies distribution started by modelling the mean pore throat size distribution and then grouping sets of distributions into lithological facies. The internal architecture is modelled by knowing the thickness of each geological zone along with the mean porosity and permeability values – SMLP plots. The SMLP characteristics mimic realistic flow unit architectures typical in aggradational carbonate parasequences.

About this webinar

This talk presents a new open access carbonate reservoir case study that uniquely considers the major uncertainties inherent to carbonate reservoirs using one of the most prolific aggradational parasequence carbonate formation set in the U.A.E; the Late Barremian Upper Kharaib Mb. as an analogue. A wide range of interpretational scenarios and geomodelling techniques were used to capture the main components of its reservoir architecture, stratal geometries, facies, pore systems, diagenetic overprints and wettability variations across its shelf-to-basin profile.

The novelty of this work has been to create semi-synthetic open access carbonate reservoir models which will enable the geoscience and reservoir engineering community to analyse, study and test a number of cases related to new numerical algorithms for reservoir characterisation, reservoir simulation, uncertainty quantification, robust optimization and machine learning.

The value of this study is also to expose a model and a dataset to the reservoir simulation engineers so they can explore the impact of different fluid flow physics on sweep and recovery across multiple carbonate reservoir architectures with diverse lateral and vertical rock and fluid complexities – all of which can be history-matched against a ‘truth case’.

Jorge Costa Gomes

Jorge Costa Gomes
Heriot-Watt

Presenter Bio

Jorge holds a BSc in Geology & Petroleum Geology from the University of Aberdeen, UK and an MSc in Petroleum Engineering from the Petroleum Institute (Khalifa University), U.A.E. Jorge worked as a research assistant in the research labs of The Petroleum Institute as well as working as a Reservoir Geologist for Abu Dhabi Marine Operating Company (ADMA-OPCO). Jorge is currently pursuing a PhD in Petroleum Engineering at Heriot-Watt University, UK.

Date & Time

Monday, 1 June 2020 |
9 am GMT / 12 pm KSA

COST

Free of charge.

PRESENTER

Jorge Costa Gomes

PhD Candidate
Heriot-Watt University