Drilling-Fluid Behavior During Reservoir-Formation Drilling and Completion

Topics: Drilling fluids
Fig. 1—Example of the change-mapping technique applied to a core sample. The after-test image (a) shows a drilling mudcake attached at the top, which is not present in the before-test image (b). Image (b) is subtracted from image (a), leaving behind the change (c).

You have access to this full article to experience the outstanding content available to SPE members and JPT subscribers.

To ensure continued access to JPT's content, please Sign In, JOIN SPE, or Subscribe to JPT

A reservoir-conditions coreflood study was undertaken to assist with design of drilling and completion fluids for a Norwegian field. Multiple fluids were tested, and the lowest permeability alterations did not correlate with the lowest drilling-fluid-filtrate-loss volumes. This paper will examine the factors that contributed to alterations in the core samples.

Introduction

A range of measurements are made during reservoir-condition studies, with typical metrics of the performance of a fluid or sequence including the following:

  • Permeability measurements are made at initial reservoir conditions and then again at various points throughout the study.
  • Filtrate-loss volumes are used to compare the performance of various fluid types, including bridging design (drilling fluids), activation of crosslinked gels (kill pills), and breakthrough time/rate (treatment fluids).
  • Production/injection rates or differential pressures can give some broad indications of cleanup but are generally prone to artifacts or misinterpretation caused by multiple mechanisms occurring simultaneously within samples.
  • Visual observations can give an excellent overview of external features of the samples such as drilling-mudcake cleanup and sanding or sample failure or fracturing. However, they do not show what changes have occurred within samples and cannot visualize changes at a microscopic level, and both are generally key to understanding results.

These metrics are unfortunately subject to a number of factors that make interpretation difficult and therefore add risks to the decision-making process. In order to reduce these risks, a number of interpretive techniques are used. These include scanning electron microscopy (SEM), thin sections, and computed-tomography (CT) scanning.

In order to overcome some of the limitations posed by existing techniques, a micro-CT change-mapping technique was developed to show the distribution of alterations within samples at selected points in a study.

Do Filtrate Loss Volumes Tell Us How a Drilling Fluid Is Performing?

In terms of aiding operational decisions, the remaining mudcake attachment after a period of production or injection is most relevant in maximizing hydrocarbon recovery. The cleanup of drilling mudcakes will be influenced by a range of factors. An approach that allows a holistic view of the changes related to drilling fluid, taking into account as many relevant factors as possible, is therefore desirable.

It can be argued that it is not optimal to have a “perfect” filter cake that is immediately formed with no filtrate loss. A well would be vulnerable to gas influx if no filtrate were to enter the formation and the filter cake were to become a membrane dividing a gas-bearing zone from an oil-bearing zone without gas. The optimal solution may be to drill with a fluid such that some filtrate is allowed to enter the formation without creating any permeability reduction.

It is proposed that the operational sequence should be simulated as closely as possible under reservoir conditions, and the results augmented by visualization techniques such as micro-CT scanning to understand the behavior of operational fluids.

Micro-CT Change Mapping

To visualize the changes within the samples caused by the operational fluids and production, a micro-CT change-mapping technique was used. The process generates a visualization of the differences within the core samples at various points in the study. A high-resolution micro-CT scanner was used to take a snapshot of the samples at specific points in the study, allowing visualization of features as small as 15 µm in size. Data sets are loaded into modeling software and aligned, and each point of intersection in the data is compared to assess the change between the two selected scans. An example of the technique used on a single 2D slice can be seen in Fig. 1 above; this is applied to each of the more than 3,000 scans per sample, giving a 3D distribution of detectable changes.

Case Study: A North Sea Oil Field

To aid in selection of drilling fluids for production drilling, a matrix of 12 tests examined three different drilling-fluid formulations. These three drilling fluids were applied to two formations. The tests included evaluation of the drilling fluids alone and in the full operational sequence using openhole completion with standalone screens. All tests were conducted under reservoir conditions.

Representative core samples were supplied for the two formations—Paleocene sandstones of the Heimdal formation (Lithology 1) and injected Hermod sandstones (Lithology 2). These were moderately to poorly consolidated, and gas permeabilities ranged between 2 and 3.5 darcies. The core samples were prepared to initial oil-leg saturation by use of synthetic formation brine and a blended refined oil that was matched to the reservoir viscosity. The baseline permeability to oil was then measured at reservoir conditions, and this value was used as a benchmark for any changes that occurred during the test sequences.

Two phases of testing were carried out:

  • Phase 1 examined the various drilling fluids only. In these tests, the drilling fluids flowed dynamically past the wellbore face for 48 hours at 88.4-bar (Lithology 1) or 97.7-bar (Lithology 2) overbalance pressures. Following the dynamic drilling-fluid application, the core samples were exposed to the drilling fluid (at the same overbalance pressures) statically for a further 48 hours. Filtrate-loss volumes were recorded throughout the 96-hour application period. Production was then simulated by flowing matched-viscosity mineral oil at increasing differential pressures. Permeability measurements after drilling-mudcake removal and spin down gave an indication of the effect of the drilling-mudcake attachment and fluid trapped within the sample.
  • Phase 2 expanded the scope to the reservoir operational sequence. After drilling-fluid application, displacement/screen-running fluid was applied dynamically at the same overbalance pressure as the drilling fluids. A cutout of the 250-µm wire-wrap screen that will be used in the field was inserted, and the samples then underwent production and the various permeability measurements described in Phase 1.

In both phases, visualizations were carried out in order to allow an understanding of the results. This consisted of SEM analysis and micro-CT scanning. For the SEM analysis, an untested offcut of each sample was examined alongside subsamples from various locations in the tested sample. Micro-CT change maps were produced, with scans made at the initial saturation, after the production phase of the study, and again at the endpoint after spin down.

Filtrate-loss volumes were less than approximately 2.5 mL in all samples, well within the expected range for oil-based mud, with samples at approximately 2.5-darcy permeability. While all drilling fluids could be considered to have good filtrate-loss control, the combination of Formulation A with Lithology 1 was clearly different, being approximately 50% higher than the other 10 samples.

Phase 1 permeability data showed alterations in permeability in all samples after drilling-fluid application and drawdown, ranging from –24 to –44%. After removing any external drilling mudcake, permeability improved in all samples (by 4–14%), showing that some barrier to flow remained after production, and further improved (by 3.5–12.5%) after spin down, showing that mobile fluid was trapped (retained) within the samples. It was noted that the sample with higher filtrate-loss volume did not show the highest reduction in permeability overall, nor did it show significantly higher improvements when the external mudcake was removed. This suggested that, after production, the higher filtrate-loss volumes had a significantly negative effect.

Phase 2 permeability data showed trends very similar to those of Phase 1. Generally, the alterations in permeability were slightly higher than those seen in Phase 1, as were the improvements after removing the external mudcake and spinning the samples down. This was most evident after spin down, which showed larger improvements and therefore more mobile fluid retained after production.

The screens were examined under a microscope after offloading; this confirmed that the improvements after removing the external mudcakes and screens were attributable to the attachment of the drilling mudcakes rather than the screens.

The micro-CT change maps provided good insight into the permeability results, showing that most of the change was at the operational fluid-cake attachment and within the first few millimeters of the wellbore.

The combination of test data and visual and interpretive observations offered added insight for these formations, tested under reservoir conditions, with candidate operational fluids. The study confirmed that it was not the volume of filtrate loss, depth of invasion, or the amount of cleanup that was key. Instead, it was the nature of the attachment of the drilling mudcake to the core and how it responds to production.

Conclusions

The North Sea case study provides an excellent example of conclusions that can be drawn with the correct data. A thorough examination of three different drilling-fluid formulations revealed the following:

  • There was no direct connection between oil-based drilling-fluid-filtrate loss to the formation and the formation damage caused by the drilling fluid.
  • The major formation-damage cause is more likely to be the drilling-fluid filter cake’s ability to stick to the formation and whether it can be removed during production.
  • The key zone for permeability alteration in the samples included the first few pores in from the wellbore, regardless of the volume of filtrate loss or thickness of remnant drilling-fluid filter cake.
  • Results were different when looking at the drilling fluid alone and as part of the operational sequence. Studies should ensure that these variables are covered.

Wherever possible, an approach of simulating the operational sequence at reservoir conditions and then gathering the necessary information to interpret the data should be taken to minimize operational risk as much as possible.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 185889, “The Nature of Drilling-Fluid Invasion, Cleanup, and Retention During Reservoir-Formation Drilling and Completion,” by Justin Green, Ian Patey, and Leigh Wright, Corex; Luca Carazza, Aker BP; and Arild Saasen, University of Stavanger, prepared for the 2017 SPE Bergen One Day Seminar, Bergen, Norway, 5 April. The paper has not been peer reviewed.

Drilling-Fluid Behavior During Reservoir-Formation Drilling and Completion

01 February 2018

Volume: 70 | Issue: 2

STAY CONNECTED

Don't miss the latest content delivered to your email box weekly. Sign up for the JPT newsletter.