2018 Spring Workshop

20MHz NMR Core Analysis Value

Harry Xie (Core Lab)

Because of its great sensitivity to solid organics and other advantages, high frequency (20MHz) NMR core analysis has been applied extensively to unconventional mudstones and gained market acceptance in the oil and gas industry. The high frequency NMR correlation time plots, e.g. the T1-T2 Mapping techniques, have the sensitivity and resolution in the relaxation time space to detect and quantify all liquids and solids that contain hydrogen in mudstones. The state of art high frequency NMR core analysis techniques are utilized in two ways in order to study petrophysical and petrochemical properties of unconventional core samples: (1) static measurements to quantify liquid porosities, liquid saturations, and solid/immobile organics; (2) dynamic measurements under reservoir conditions to study changes of each constituent due to temperature, pressure and fluid flow in the sample. This talk presents the technical advantages of high frequency NMR compared to conventional low frequency (2 MHz or lower) NMR. Typical NMR 2-D maps from various shale formations will be shown with emphasis on quantifying water and organics together with comparisons with traditional Dean Stark analysis and Rock-Eval pyrolysis. The dynamic NMR measurements reveal that solid organics in the unconventional formations play a key role in oil and gas generation and production.

Reservoir Performance in Unconventional Reservoirs, the EOR Requirement

Derek Beckett/Ted Griffin (Core Lab)

Reservoir performance evaluation in liquids producing shale reservoirs is, to say the least, unconventional.  So little is yet fully understood and conventional wisdom may be sufficiently misleading to be inappropriate.  As an example, diffusion coefficients in nano-scale pores may be remarkably pessimistic and consideration need be given to in situ, experimental investigation.  Experimentation to determine recovery under primary depletion requires restoration of pore pressure and effective stress without catastrophic consequence, establishing reservoir temperature and restoring the saturation state with live reservoir fluids before performance testing can commence.  In permeable reservoirs this reconstruction is trivial.

Additional, incremental recovery under gas cycling following primary depletion can be economically significant and simulation of this process in the laboratory can provide recovery characteristics.  The testing regimen normally includes reservoir fluid phase behavior, condensation/vaporization & miscibility study followed by injection/production profiling.  This presentation includes EOR (gas cycling) recovery improvement, diffusion considerations, permeability stress dependence manifested during production/injection phases and visualization of pore space and fluid distribution using high frequency NMR and microCT scanning.    

Lessons Learned in Permian Core Analysis: Comparison Between Retort, GRI, and Routine Methodologies

Aiden Blount (Shell)

In today's competitive cost environment, core acquisition and analysis is too often dismissed as unaffordable. This forces petrophysicists to make every dollar count in core evaluation. Tough choices have to be made-many people chase the lowest bid, least expensive methodologies, reduced oversight, and less sampling. In this paper, insights will be shared from a comprehensive round-robin study directly comparing the results of the most common techniques (GRI/Retort/RCA) used by major vendors. Understanding differences in techniques early in an evaluation process can help efficiently direct technical spending.

As with many comparison studies, this project started with the reconciliation of analysis sourced from different laboratories using different methodologies.

There was a significant business driver to this work as we noticed differences in measured porosity and fluid saturations that contribute to significant differences, approximately 25%, in hydrocarbon pore volume among vendors using alternative techniques. These differences directly impact log calibration objectives as well as estimations of hydrocarbons in place.

We began to ask a series of simple questions: Should we use crushed samples or routine core plugs? What is the impact of analytical technique on the results? What role does lithology and organic content play in the results from different analytical techniques? What is the role of sample size? What is the variability between vendors for identical procedures? If there is variability, what is the apparent cause?

A set of 10 twin samples of Permian Bone Spring formation from the Delaware Basin in Texas and New Mexico was evaluated using a variety of laboratory­ derived measurements, including X-ray diffraction (XRD), total organic carbon (TOC)/RockEval, retort, and Dean­ Stark/Gas Research Institute (GRI) protocol analyses from two labs and RCA from one lab. These 10 samples were selected to represent varying lithofacies with a range of organic, mineralogical, and water/oil content. The level of oversight at each data source was also tracked.

Through detailed analysis of the raw data from these measurements, we address the questions above. With these results, we hope to (1) maximize every dollar spent in core analysis, (2) focus oversight where it is truly required, and
(3) accurately and consistently evaluate the core analysis in the Permian play for fast and value-driven business decisions.

GRI Porosity Issues; Integration of GRI with NMR

Mark Sutcliffe (Council Oaks)

TBA

Strategies for Estimating Lithofacies, RQ, and HC in Non-Cored Wells Using Cored Wells

Eric Eslinger (eGAMLS Inc)

Main Points

1. Probabilistic multi-well MVCA (multi-variate cluster analysis) is used to discriminate, classify, and correlate rock types (~lithofacies) into electroclasses.
2.  The clustering variables typically used are standard well logs (RHOB, NPHI, GR, + PEF, + log RT …).
3.  Wells used in a multi-well MVCA can include both cored and non-cored wells.
4.  Mixed type clustering variables (e.g., well logs plus core plug data) can be used.  When this is done, an essential pre-clustering task is an accurate, variable core-to-log depth correction. Also, when this is done, a by-product (in addition to the actual clustering results) is predictions for missing or nulled core data within which the predicted results are inherently upscaled due to the differing scales of the log and core data.
5.  The MVCA procedure requires that all samples be assigned an initial starting classification.  This process is termed “initialization”. Unlike a back propagation neural net where only a few samples are selected for initialization, all samples - except those that are nulled - are used for initialization and all samples are used during the iterative MVCA. This initialization process is important as the goal is to begin the assignments (viz., the a priori classification) as close to the perceived “answer” as possible. Initialization is termed “unsupervised” when there is little upfront information on the actual rock classification (e.g., there is no whole core data to permit a preliminary sedimentologic description). Initialization is termed “supervised” when there is some core data that can be used to develop a preliminary sedimentologic description.
6. Ideally, the best initialization procedure is to use a whole core sedimentologic description for initialization of the MVCA.  There are special challenges (to be discussed) when doing this, particularly for thin-bedded rock sequences.
7.  At conclusion of a MVCA that has used unsupervised initialization, an “expert system” is used to assign the resulting electroclasses to likely rock types (e.g., clastics or carbonates, and their likely sub-types (Ss, Sltstn, Sh, Ls, Ds …).  This classification can be updated when additional information (e.g., additional wells or core data) is obtained.
8.  A mineralogy-based forward-modeling routine permits determination of mean mineralogy of each electroclass, and the specific mineralogy of each sample for each electroclass is determined using the probability assignments from the MVCA. During this modeling process, the number of minerals used is not restricted to the number of well logs used during clustering. The modeling process generates profiles for TPOR, EPOR, Swt, Swe, plus additional parameters using a mineralogy petrophysical properties "look-up" table plus a series of empirical equations with outputs constrained by selected logical criteria.  A model is developed for the “best” cored well, and the inputs saved and then applied to non-cored wells.
9. In a multi-well MVCA, the model electroclasses translate and correlate among wells.  Additional wells can be added to the multi-well Project as they are drilled.
10.  More detailed (enhanced) modeling is possible when a downhole elemental spectroscopy log is available along with sufficient and accurate XRD (or FTIR) core data such that the stoichiometric conversion of elemental data to mineralogic data can be calibrated using the core mineralogy data.  Such analysis results can be translated to non-cored wells included in the multi-well MVCA.
11.  A less robust but much faster multi-well "quick look" analysis for electroclass-based TPOR and Swt can be implemented. This procedure is also dependent on the results of a multi-well probabilistic MVCA. Ideally, this procedure can be calibrated using results from the more robust mineralogy-based forward-modeling analysis.

Fluid Typing with Core NMR

Ravinath Viswanathan (Schlumberger)

TBA

Automated Cuttings Analysis for Risk Evaluation and Reservoir Characterization

Don Hall (Schlumberger)

Cuttings are an underutilized resource in the petroleum industry. They are an unavoidable byproduct of the drilling process, are available from the entire penetrated section, and are relatively inexpensive to collect and archive. As exploration focus changes, cuttings from legacy wells can be obtained from whatever stratigraphic section is of current interest, in contrast to core. Given the relatively high percentage of horizontal drilling in resource plays, and the difficulties in obtaining cores or comprehensive logs from laterals, it is important to collect as much information as possible from the rock material that is inevitably produced.

Automated instrumentation has been developed for characterization of rock-fluid systems whereby large numbers of cuttings or other rock samples are conveyed sequentially through a series of analyzers including: high-resolution ultraviolet and visible light photography, trapped fluid analysis by direct quadrupole mass spectrometry, X-ray fluorescence analysis, and diffuse-reflectance infrared Fourier-transform spectroscopy (DRIFTS). All analyses are performed on the same 0.5g rock sample, thereby preserving interrelationships between rock and fluid, and allowing for multiple data sets to be collected on limited sample volumes. The entire multi-stage analytical cycle lasts 5 days, including automated interpretation and report generation, and the current scalable capacity is more than 2000 samples/day.  This allows databases to be efficiently and cost-effectively generated using fresh or archived rock material.

Although cuttings are the main feedstock for the process, other rock material can be used, including the frequently discarded residue from the Scratch or TSI instrument, which produces continuous strength index measurements by cutting a groove along the surface of core samples with a specialized apparatus.  This introduces a high resolution geomechanical component to the workflow.  Additionally, the automated data backbone can be used as a screening tool to optimize additional work, including petrology, mineralogy, geochemistry and detailed geomechanics.

Automation of this process has significantly increased consistency and interpretive certainty, and reduced turnaround time as compared to manual approaches.  More importantly, demonstrable and measurable value has been established with this process on both the conventional and unconventional side, and for both exploration risk assessment and reservoir characterization. The relatively quick turnaround time for these analyses bridges the time and information gap between near-real time analyses conducted at well site, and longer-term rock-based analyses completed in the lab.

Impact of Experimental Studies on Unconventional Reservoir Mechanisms

Richard Rosen (RL Rosen Consulting)

Novel apparatuses have been developed to measure permeability using steady- and unsteady- state methods on nano-Darcy (nD) shale (source rock) using intact cylindrical samples returned to isostatic effective reservoir stress. The steady-state method uses a high pressure dual pump system using supercritical fluids. High pressure supercritical fluids have low viscosity and low compressibility. The effect of low viscosity fluid results in measureable flow rates and the effect of low compressibility fluid minimize unsteady-state transients thereby reducing the amount of time required to achieve steady-state equilibrium. Specially designed and configured pump systems, seals and sleeves reduce leak rates to allow Darcy flow and permeability determination below 1 nD. The unsteady-state method is based upon standard designs but is optimized for small pore volume. In this report we present a summary of over 200 such permeability measurements. Permeability is observed to be dependent on geologic parameters, such as, texture and composition. Stress dependence, with hysteresis, is observed for samples with and without fractures as is rate dependent skin (Forchheimer). An interpretation model where matrix storage feeds a progressively larger fracture network provides a logical basis for a dual-porosity reservoir simulation model. This dual-porosity model is used to understand the influence of reservoir production parameters, such as choke management.

An additional observed effect is possibly related to pore collapse and disconnection. Pores associated with organic matter are softer than the surrounding mineral matrix. If these pores have a sufficiently small throat diameter, it is not hard to envision that they easily compact and close under increased effective stress as the result of reservoir depletion. Therefore, organic pore systems can become isolated unlike those of a sponge where fluids remain in pressure communication at all times. The implication of such pore isolation phenomena is that fluid material balance is not preserved during production and can contribute to large production decline rates.

Moving from Qualitative to Quantitative Geomechanical Models

Claudia Amorocho (Weatherford)

Geomechanics solutions offered in the industry cover a wide spectrum, going from single well,  log-based, uncalibrated dynamic elastic properties calculations to fully calibrated field studies. Understanding the differences among them is essential to use the results correctly.  Basic products are considered useful to determine rock quality and define terms such as “fracability” and “fracture potential”, however the values obtained from such simple models should be used cautiously when applied to different geomechanical models. On the other end of the spectrum, full geomechanics studies are focused on integrating log, core, and subsurface data to obtain calibrated results.  This sort of integrated study allows us to move from a qualitative indication of “fracability” to a calibrated estimation of the expected fracture geometry and ultimately hydrocarbon production.  This presentation will review the stages in a standard geomechanical model and show how the integration of core data, well logs and field measurements can allow us to move from qualitative indicators to quantitative models.

Part 1: Results from a Comparative Study of Core Analysis Procedures in Unconventional Tight Oil Rocks

Neil Fishman (PetroLogic Solutions, LLC)

Having previously encountered difficulty in obtaining good quality core saturation analysis data using routine core analysis (RCA) from low porosity/permeability rocks in the Bakken and Three Forks Formations, North Dakota, a multi-lab comparative study was undertaken to search for ways to improve core analytical measurement approaches. This study was conducted to understand and evaluate the various core analytical approaches, methods and techniques being utilized by commercial service laboratories on unconventional tight oil reservoirs, particularly rocks containing hyper-saline brines. The primary goal of this study was to determine which methods would be most suitable and accurate for use on Bakken/Three Forks rocks.

This study evaluated and compared four different core analytical saturation approaches using core plugs and crushed rock analytical approaches. The four techniques conducted by the three labs were:

1. Dean Stark routine core analysis (RCA) strictly following protocols outlined in API RP-40
2. Crushed rock analysis (CRA) broadly following the techniques outlined in GRI- 95/0496
3. Crushed rock Retort using the individual lab protocols
4. Nuclear Magnetic Resonance (NMR) on core plug used in Dean Stark RCA method at native state and post-cleaning conditions

Part 1 of this presentation will present the results of the first three methods made by each of the labs. Results indicate that Sw determinations were higher from CRA compared to RCA from all labs. In addition, permeability measurements are generally higher from RCA compared to CRA, although problems that developed in core plugs may indeed be responsible for some of the measured permeability.  These results have led to a new set of protocols and best practices for use on Bakken/Three Forks cores. Part 2 of the presentation will cover the results of the NMR measurements.

Part 2: Results from a Comparative Study of Nuclear Magnetic Resonance (NMR) Core Analysis Measurements in Unconventional Tight Oil Rocks

Gary Simpson (consultant)

Included in the comparative study of unconventional tight oil core saturation measurement techniques presented in Part 1 of this presentation, laboratory Nuclear Magnetic Resonance (NMR) measurements techniques were evaluated. Each of the three core analysis service companies participated in this study were instructed to make native-state NMR measurements on each of the routine core plugs prior to undergoing Dean Stark cleaning processes and then repeat the measurements on the plug following the completed Dean Stark cleaning processes.

Part 2 of the presentation will present the remarkable results the data revealed in the comparative study using NMR measurements. The main result coming from the study confirmed that industry standard Dean Stark RCA cleaning processes did not clean all fluids from these unconventional tight oil rocks. The data also revealed that standard interpretation T2 cut-offs typically used in NMR logs analysis to calculate clay bound water (CBW) and free-fluid were incorrect. Lastly, the data revealed that in the case of the Bakken/Three Forks petroleum system that NMR log and core data must have a large hydrogen index correction due to hyper-salinity of formation waters.

Finally, the comparative study has broadened our understanding of errors which can occur in core measurements associated with unconventional tight oil core samples. These understandings are allowing the development of new core analysis protocols that could potentially reduce the overall error of petrophysical measurements in unconventional tight oil cores.

Visualization and Quantification of 3D Pore Networks using Confocal Laser Scanning Microscopy (CLSM)

Michael Hoffman (AIM GeoAnalytics)

Imaging small micropores from core samples and cuttings is a key building block for accurately evaluating the volumetric pore space of a rock sample. Standard petrographic techniques have limited value for micropore quantification due to the interference of multiple crystallographic layers and limited optical resolution. High-resolution imaging techniques, such as SEM and micro-CT scanning have played a pivotal role in the characterization of micropores, but theses imaging techniques are destructive and/or they cannot easily obtain a wide field of view of a sample, and thus often only image a very limited number of pore types in a sample.

This study focuses on recent advances in the application of confocal laser microscopy (CLSM) to image and characterize geological samples with complex microscale porosity networks. In the life sciences, CLSM is an established light microscopy technique for imaging fluorescently labeled specimens with three dimensional structure. In the geosciences CLSM has recently been applied as a promising method for generating high-resolution 3D datasets of pore networks in conventional and unconventional reservoirs. CLSM has the unique capability of producing optical sections of the specimen, a non-destructive imaging approach which uses light, rather than a physical method, such as a focused ion beam milling, to section the sample. These optical sections are captured at a submicron scale down to the resolution limit of CLSM of ~300 nm in the vertical direction and ~ 200 nm in its horizontal direction. This high-resolution 3D imaging technique allows for the quantification of (micro-) pore shapes and associated pore throat sizes over a wide field of view in complex porous samples. One major shortcoming of CLSM is the resolution limit of ~200 nm in the horizontal that prevents the 3D imaging of pores below that size, and can lead to diffusion of the signal in some areas. However, CLSM is a promising tool that provides a fast and inexpensive analysis of pore networks and can help to bridge the resolution gap between standard petrographic microscopy and higher-resolution techniques. This study discusses the technical and practical aspects of CLSM with application to tight oil and gas reservoirs.