Tuesday March 27, 2001

 

Attendees

 

Ahmed Abou-Sayed

Advantek

Brian Odette

First Choice

Marco Brignoli

AGIP

Henrik Ohrt

Maersk

Roberto Cherri

AGIP

John Shaw

Statoil

Jean-Louis Detienne

TFE

Alastair Simpson

Triangle

Marc Hettema

Statoil

Idar Svorstol

Norsk Hydro

Trond Jensen

Phillips

Paul van den Hoek

Shell

Bruce McIninch

Marathon

Dale Walters

Taurus

John McLennan

TerraTek

Karim Zaki

Advantek

Laurence Murray

BP

 

 

 

Overview Session

Marco Brignoli welcomed the participants and outlined safety and organization details.  John McLennan summarized the agenda and the financial status of the Project.

 

The purpose of the meeting was indicated as:

 

Ø      To review ongoing contractor work and to critique planned activities.

Ø      There will be a summation of results from each Task (on Tuesday, 3/27) followed by a Task-by-Task presentation of gaps, future requirements, and recommendations for ongoing work on Wednesday, 3/28.

Ø      Thursday morning, 3/29 is set aside for a Sponsors review, followed by feedback to the Contractors.

Ø      Thursday afternoon will be the start of a workshop on the Toolbox, which will continue through Friday morning for a tentative adjournment before lunch on Friday morning.

 

The Sponsors reviewed the financial status.  The only question raised was what has happened to interest on the money that has been provided.  This question was posed to Bob Siegfried, with GTI, and his response is as follows.

 

John-

I have asked our attorney to review the PWRI sponsor agreements regarding
the accrual of interest on funds collected for the project. His reply is
excerpted below:

Section 4.4-b provides the following: "In the event that more than sixteen
(16) Sponsors join the Project, a proportionate share of the excess
contributions shall be repaid to the Sponsor upon Project Completion or
termination. However, the Steering Committee ... by vote approval may
decide to increase the total budget of the Project as long as this does
not imply major changes in the scope and objective." Because there is no
provision in the contract for the payment of interest on excess funds
collected, normal rules of contract construction will not imply one. The
parties only stipulated upfront that excess contributions will be "repaid"
on a proportionate basis (without reference to interest on such funds).

Thus, while we will cooperate fully in the establishment of a new
administrative arrangement and return excess contributions to the sponsors
as provided in the agreement, we do not plan to pay interest on these
funds.

Regards,

Bob

Robert W. Siegfried, II
Gas Technology Institute
1700 South Mt. Prospect Road
DesPlaines, IL 60018-1804
(847) 768-0969 Fax: (847) 768-0995
Robert.Siegfried@gastechnology.org

Overall Summary of Tasks

John McLennan provided a brief overview of the status of all of the Tasks.  The slides in the presentation are self-explanatory.  There are some remarks that are relevant.

 

  1. For all Tasks, Paul van den Hoek emphasized the importance of providing reports that, for example, discuss what was done when data were analyzed.
  2. Jean-Louis indicated that it was necessary to check the links to the files in the Workshop.
  3. Task 1: In discussions of falloff testing, indicate why don’t commercial software simulators work.  What are their limitations and what is the range of validity?
  4. Jean-Louis Detienne (shortly after the meeting) provided additional information on fracture models being used at TFE (Hydfrac, Diffrac, and Predictif).  These have been added to the model audit document for Task 1 and will be distributed for review soon.
  5. Task 3: Various changes in the Completions Worksheet are being made based on observations by Laurence Murray and Paul van den Hoek.

 

 

Task 1 – Monitoring

John McLennan provided more detail on Task 1, covering Monitoring and Interpretation.  The key elements of the presentation are as follows.

 

  1. John indicated the status of incorporating monitoring methods in the Best Practices document.  The status is:

Ø      Conventional methods completed

Ø      Need to add fiber optics

Ø      Need to finalize PLTs

Ø      Need to improve fractured monitoring

Ø      Need to incorporate specific Best Practices, in addition to the descriptions of the various methods.

Ø      Falloff testing was discussed (Slides 3 and 4).

Ø      Slide 5 describes the importance and analytical methods for determining average pressure.  It is emphasized that even if these evaluations are done with software it is desirable to understand the basic principles and the potential pitfalls.

Ø      Slide 6 outlined the status of the section on Hall plotting.  Basic documentation is done and additional information is being added on some of the moving average concepts that have been recently developed.

Ø      Step rate testing and Hydraulic Impedance testing have been completed.  Sections have been added on pulse testing, interference testing, leakoff testing and micro-hydraulic fracturing.  The section on Drillstem Testing will be brief as it is discussed in detail in numerous public domain references.

Ø      Prior to the meeting, sections were added on tiltmeter as well as microseismic monitoring.

Ø      Real-time methods for evaluating the success of acidizing have been included in the Best Practices.

Ø      The status of the Fracture Modeling Audit was reviewed.  As indicated, some recent additions have been made (TFE models and additional information on two Shell models).

Ø      Because of its relevance, a brief section describing some of the key theoretical aspects of evaluating well performance (with a focus on injectors) has been started and will be completed before June.

Ø      The new look of the Best Practices document was presented.  The philosophy is to navigate on the basis of appropriate well or field activities rather than according to Tasks in the JIP.

 

Tasks 2 and 4

Tasks 2 and 4 cover Matrix Injection and Stimulation/Mitigation, respectively.  Because of commonality in these Tasks, they were presented together by Ahmed Abou-Sayed.  Ahmed prefaced his discussion by showing the contents of the Toolbox, as it currently stands and summarizing the planned modifications (additional modifications were specified by Sponsors as the Meeting went on).

 

Ahmed continued with a presentation on Matrix Injection issues.  Recall that matrix injection has been specified to include injection into non-propagating fractures.  Some of the observations made included the following:

1.     The Reciprocal Injectivity Index –RII, following on from Chevron's use of this, has been adopted for many of the analyses that are carried out. Paul van den Hoek wanted to know why was RII used rather than II.

2.     A number of examples were given as to the influence of reservoir pressure – constant versus variable.  The basis for the analyses were wells from two blocks in the Maersk A field.  Reservoir data are imported into the Toolbox for analysis.  Ahmed committed to allow reservoir pressure to be interpolated between discrete chronological measurements to represent approximations of temporal variation in reservoir pressure.

3.     Ahmed then showed a Phillips' field case in a chalk.  The reservoir pressure was determined from history matching (Figure 1).  Figures 2 through 4 illustrate, fairly dramatically, the pitfalls of not using the correct reservoir pressure.  Figure 2 is a Hall plot.  The two curves are bases on 1) constant reservoir pressure and 2) the reservoir pressure shown in Figure 1.  Certainly constant reservoir pressure assumption, in this case, would be misleading.  Figures 3 and 4 support the observations in Figure 2 on the basis of the variation of the Reciprocal Injectivity Index.  Some actual injectivity improvement is seen if a variable reservoir pressure is incorporated (this may actually be a situation where some fracture growth is occurring?).

 

 

Figure 1.        Reservoir pressure in an example chalk well, chronological.

Figure 2.        Hall plot determined with a constant reservoir pressure and with the reservoir pressure shown in Figure 1.  Certainly the assumption of constant reservoir pressure could lead to costly and unnecessary intervention.

Figure 3.        The variation of the Reciprocal Injectivity Index with date, calculated presuming that the reservoir pressure was constant.  The increasing RII plotted would mistakenly indicate loss of injectivity (see Figure 4).

Figure 4.        The variation of the Reciprocal Injectivity Index with date, calculated using the reservoir pressure from Figure 1.  The increasing RII plotted would seem to indicate relatively constant injectivity (see Figure 3).

 

The next topic discussed was modifications to the Frictional Calculations Program that are planned (additional changes were suggested to Karim Zaki when he presented the Toolbox itself, later in the week - for example, multiple tubing strings and a way to include minor losses - e.g. valves).  Ahmed indicated that they were adding viscosity changes with temperature, density, and viscosity, as well as the influence of solids content on viscosity.  The rationale for this level of sophistication was a point of argument, presuming that roughness will have a large role.  However, it was accepted that any level of improvement in BHP inference was important because the net pressures (either above closure or above the formation pressure) are often quite small.

 

Next a series of slides was shown to demonstrate characteristic differences between matrix and fractured injection.  There was the Elf3 example, under matrix injection – there was one step where there was no propagation of a fracture.  This was followed by a Shell NAM example showing some spontaneous fracture growth events and a BP Prudhoe Bay example.  There was a great deal of consternation because the slides were labeled as indicating that there was stimulation.  NOTE:  The use of the word stimulation.  In the context of these slides stimulation refers to improved performance because of spontaneous fracture growth, as well as hydraulic fracturing or acidizing.

 

Finally, there was emphasis of the concept that fractures are conductive even when they are nominally closed.  Fractures are conductive long before they are reopened.  Ahmed showed a plot from Voegelle et al., 1982, based on large block testing that showed fracture conductivity even with high normal stress, acting across pre-existing fractures.  Roberto Cherri and Ahmed discussed this concept at some length.

 

Ahmed next discussed some of the recent work that has been done on Task 4 - Stimulation.  The premise was a mechanism for sporadic propagation of a fracture after it had been progressively plugged.  Signatures for this were indicated on RII plots.  Figures 5 and 6 are examples of the concept.  In these figures, it was suggested that the upper locus (blue line) is largely controlled by temperature, damage and pore pressure, whereas the lower locus (blue line) is mostly impacted by the in-situ stress conditions.  Figure 7 shows further specifications.  All of the slopes shown can be taken to be diagnostic.  In Figure 7, the rate of change during the plugging phase is a function of the water quality and the amount of recovery in RII after fracturing is governed by the in-situ stress conditions.

 

 

Figure 5.        The variation of the Reciprocal Injectivity Index with time.  RII varies between two loci.  The fracture plugs, the injectivity declines.  The pressure increases until finally some supplementary fracture extension spontaneously occurs and new injection surface area and fracture volume is created.

 

Figure 6.        The variation of the Reciprocal Injectivity Index with time.  RII varies between two loci.  The fracture plugs, width may increase but length growth is impeded and the injectivity declines.  The pressure increases until finally some supplementary fracture extension spontaneously occurs and new injection surface area and fracture volume is created.  This figure goes one step beyond Figure 5.  It is an assertion that when the two blue loci intersect complete plugging occurs and injectivity will be largely lost unless a completely new fracture system is created.

 

 

Figure 7.        It is believed that the slopes in this type of RII plot can be diagnostic.  For example, you will want to stimulate before the two blue loci intersect.  The specific slopes are functions of certain controllable and uncontrollable parameters.

 

It is hypothesized that this type of plot can provide real diagnostic information on when stimulation should be done.  For example:

 

1.     When the lower blue locus (the minimum fracturing pressure) intersects the upper locus, the fracture is completely filled and there will be a dramatic rise in pressure. Figures 8 and 9 are examples.  Figure 10 suggests another presentation of such data (in a step rate type plot).

2.     Stimulation can be planned by establishing the two loci, predicting their intersection and ensure that you stimulate adequately in advance.

3.     If the loci are relatively parallel (the Prudhoe Bay example in Figure 11), explicit intervention for artificial stimulation will be required relatively infrequently.

4.     The lower locus for RII corresponds to matrix injection in a stationary fracture.  Upper bound is fracture injection.  The bottom and top loci can be parallel.

5.     Both lines could reflect matrix injection, one is clean one is dirty.

6.     It is desirable to tune these lines.

7.     If the two lines intersect, you have no tolerance for a drop in pressure.  When the two lines come together, it means you must propagate a fracture for further accommodation of solids.  You must be able to accommodate the fluids.  You need to decide when to stimulate.

8.     The upper line is governed by the stress and the lower line governed by the damage - or is it the other way around as was shown in the slides.  There was some argument that the filter cake should be attached to the bottom line. 

9.    RII for a fractured condition needs to be based on the pressure minus the fracture propagation pressure.

 

The next issue that was addressed was the validity of the PEA-23 correlation outside of domains similar to Prudhoe Bay. How can PEA-23 be extrapolated outside of Prudhoe Bay?  It was argued that this requires a method for predicting frac gradient.  Ahmed speculates that PEA-23 behavior will be seen if propagation occurs before the fracture is too fully filled.

 

An example from the Maersk A field was used for the assessment.  Ahmed then discussed an approach for using these observations as a diagnostic tool.  A baseline RII was defined, indicative of best performance for a well.  Certain other terms were also defined.  These included slopes of the RII plot before and after stimulation and DRIIbefore and DRIIafter (before and after stimulation).  These are the deviation of the RII from the baseline before and after stimulation.  DRII is a measure of damage and the slope of the RII plot is an indication of the rate of damage accumulation.

 

Once you go past the intersection of the two blue loci, you cannot go back to the baseline.  What you create with stimulation may be embedded in the damaged zone? 

 

Figure 8.        Progressive plugging is hypothesized for this well along with break-back as new fracture area is created or accessed.  If you draw bounding loci for this situation they would seem to diverge, suggesting interpretation can be more complicated than envisioned.  On the other hand, Figure 9 tends to support the validity of the locus construction method.

 

 

 

Figure 9.        Without too much imagination, two bounding loci can be drawn.  When they intersect, significant loss in injectivity is indicated and presumably the recourse at this time was stimulation.

Figure 10.      Step rate type presentation showing stationary fracture (blue), propagating fracture (red) and a plugged fracture in yellow.

Figure 11.      This is an example from Prudhoe Bay where relatively little human intervention is required.  The fracture system sporadically extends but the two loci are relatively parallel.  The behavior at the end was attributed to shutdowns.

 

Two examples of looking at the slopes and the DRII are available in the presentation.  An Improvement Ratio was defined as the DRIIbefore/DRIIafter.

 

Note: You need to have enough pressure capacity in your system to be able to handle the events.  It is a discontinuous rather than a continuous process - discontinuously propagating fractures - a self-correcting process.

In the long run, even if you have enough pressure available, you may not be able to stimulate adequately. 

 

There was further discussion of the meaning of the various slopes.  Laurence Murray discussed water quality as a regulator for the "red" slope.  Why does the red line commonly have a constant slope?  If the water quality changes, the slope changes for the red lines.  Laurence believes that you can argue that propagation is not the only mechanism for cleanup and that change in water quality is another one.  The red lines can be extracted from PEA-23.

 

What are some of the products of these concepts?  One example is an extension of PEA-23 with time, the influence of water quality on the frac gradient, and looking at the influence of tortuosity (per Paul van den Hoek).

 

After Paul brought up tortuosity there was additional discussion.  Laurence Murray supported the role of near-wellbore effects (citing three distinct data sets for produced water injection and oriented perforations where pressure loss was systematically determinant).  Jean-Louis Detienne supported the importance and argued for additional consideration of the completion skin.

 

Task 3 – Soft Formations

Dale Walters summarized the status of the Task on Soft Formations.  The relevant components of Dale’s presentation are as follows:

 

  1. Work completed since the December Meeting in Houston is:

 

Ø      Additional data for soft formations has been received since the December 2000 meeting.  This includes two Brage wells and Heidrun pilot data.  An example of the Brage data is shown in Figure 12. 

 

Figure 12.      Injection data from one of the two Brage wells being evaluated currently.  There are also SRT results (Figure 13).

 

Figure 13 shows step rate test data for one of the new wells.  These data will be useful because there is both produced water and seawater injection.

 

Ø      Further improvements have been made to the radial damage spreadsheet analysis tool (Jean-Louis Detienne has suggested that this tool should be called PWRAD) based on Dec 2000 meeting (isolating completion skin).  A User Guide has been written.

 

The completion skin is now input separately, following discussions of the Kerr McGee G field data at the December Meeting in Houston.  Those data have been reanalyzed, assuming a completion skin, Sc, of between 100 and 250.  The skin due to damage is predicted to be smaller but still significant (Figure 14).  Figure 14 shows skin of 600± during later phases of the injection.

 

 

Figure 13.      Step rate test data from one of the Brage wells.  This data set is “nice” because it has both seawater and produced water injection information.

 

 

Figure 14.      PWRAD simulation for a well in Kerr McGee’s G Field, showing the position of the water front, the calculated permeability in the flooded zone and the skin.

 

Figure 15 shows better-sustained performance in an equivalent BP Amoco Field.  Specific reasons for this have not as yet been determined.

 

 

Figure 15.      The injectivity index for two wells in the G field and an equivalent BP Amoco well.

 

Ø      The injectivity index for two wells in the G field and an equivalent BP Amoco well.

Ø      Re-analysis of all data sets with the modified tool and assessment of the differences, with inputted values for the completion skin.  There were no substantial changes in the conclusions.

Ø      A tool entitled WellStress has been developed and is available in the Toolbox. This is a poro- and thermoelastic stress evaluation tool.  A User Guide has been written.

Ø      Corrections and revisions have been made to the reports issued for December meeting

Ø      Evaluation of injectivity of various completions has continued.

 

  1. Dale then described work in progress, scheduled to be finished before the end of June.  This includes the following:

 

Ø      No new data analysis will be undertaken.  Taurus suggested including the Brage and Heidrun pilot data analysis in a project extension?  The SRT will be incorporated in Task 1 (Monitoring).

Ø      A report will be prepared on the completed Soft Formations (SF) data analyses.  This will include the general methodology, a detailed discussion of each data set, findings from the comparative analysis, and, a concise summary.

Ø      Investigate if a PEA 23 type correlation for fracture injection mode is feasible in soft materials.

Ø      Write a User Guide for the Radial Damage Tool

Ø      Summarize Task 3 results.  This would include major learnings, concepts, developments; gaps in understanding; future work that may be needed and Best Practices.

 

  1. Taurus also provided some suggestions for appropriate future developments/efforts.  These included:

 

Ø      Develop a comprehensive Completion Skin Tool for inclusion in the Toolbox - full set of correlations for cased and perforated completions, correlations for screens, liners and excluders …

Ø      Analysis of newly-arrived data sets (Brage, Heidrun pilot)

Ø      Rewrite WellStress library (portability)

Ø      Improvements to PWRAD.

 

More will be said about future efforts later in these minutes.

 

Task 5 – Layered Formations

John McLennan summarized the status of the Task on Layered Formations.

 

Ø      The first two slides in the presentation emphasized that even for matrix injection, analytical methodologies are difficult to use when there is crossflow.  Specifically:

 

ü      Without a simulator, it is difficult to represent interlayer vertical crossflow

ü      However, for horizontal crossflow on shut-in, there is an analytical tool that is available in the ToolBox.  A Users’ Guide is being prepared.

ü      Methodologies for minimizing horizontal crossflow have been encapsulated in Workshop minutes and in the Completions Selection Worksheet.

 

Ø      Several slides were presented summarizing the difficulties of pressure transient interpretation in multi-layered situations.  One of the difficulties is that commingling can mimic a naturally fractured reservoir if there is a fairly large permeability contrast or can appear to be homogeneous if contrast is not as strong.  For example, “If the commingled layers consist of one high permeability layer while all the others are low permeability, then the test will only give the kh of the high permeability layer.”

Ø      The next slide in the presentation summarized some of the methods that are available for analyzing layered formations.  Nearly all methods currently available require intervention and production logging.  Some require physical isolation of individual layers.  These techniques are being summarized.  Maybe the most appropriate method currently available was published by Ehlig-Economides and Joseph (1987).  It requires accurate measurement of pressure and rate in individual zones and two rates are needed.  An example has been developed for this method.  Methods were also developed by Kucuk et al. 1986 (multi-rate analysis, early transient state).  The implication for all of the available protocols is bottomhole measurements.  Flow rate and pressure survey measurements are required with depth for each rate

Ø      Nelson and Economides (1996) presented a similar concept for hydraulically fractured (stimulated) wells.  This is from an SPE publication and will be documented.

 

The fundamental issues in layered formations include:

 

ü      Plugging criteria: For example, “Where do the solids go and how are they distributed between high and low permeability layers?”  Provide a summary of injection plugging relationships (lessons learned and experience factor).

ü      Thermo- and poroelastic components

ü      Pressure signatures (future work will be required)

ü      Remediation and control (partially completed based on Layered Formations Workshop).

 

Surface Systems

Alastair Simpson presented the status of the Surface Systems Module.  Alastair’s presentation is available.  The chronology of development was described and the features of Version 5, for issue in April or May, were characterized.  These features include:

 

ü      Improved operability/user interface

ü      PDF Files (file format and software issues)

ü      More field data, etc.

ü      Best Practices will be included

ü      Improved and tested startup and support/help functions

ü      “Interim” issues should disappear

ü      CD version available April/May 2001

ü      Website version May-June 2001

 

This is an interim issue.  The final CD version (Version 6) will be available in June/July and a web-based version will follow one to two months after this.

 

Wednesday, March 28, 2001

 

Task 6 – Horizontal Wells

Dale Walters described the recent progress in the Horizontal Wells Task, including the horizontal multiple fracturing tool that was first described at the December Meeting in Houston.  Certain decisions on Task 6 were made at that meeting.  These revolved around pulling together the experience from operators and modeling work done in the project to create:

 

Ø      Operational Best Practices

Ø      Modeling Best Practices.

 

Interest was expressed in a multifractured horizontal well injectivity correlation (developed by Kuppe) for incorporation in the Toolbox.

 

The following has been accomplished since December:

 

  1. Additions and revisions to the presentations made at the December 2000 meeting covering “Analysis of Horizontal Well Injectivity – An Example” and “Horizontal Wells: Injectivity Analysis – Survey of Tools and Methods”
  2. Best Practices for modeling Horizontal Wells – first draft completed.  Four main categories have been identified – 1) thermal effects on stresses, 2) thermal effects on PVT, 3) gridding issues, and 4) Fracture coupling.  Examples will be given of the magnitude of these various effects.  The examples are drawn primarily from the Prudhoe Bay simulation work.  The bibliography is still incomplete
  3. Work on the spreadsheet tool for multifractured wells is in progress

 

The multifractured horizontal well tool incorporates multiple fractures of equal dimensions.  The spreadsheet is being developed to be put into the ToolBox (figure 16).  There was a significant amount of disagreement about the value of the tool that was proposed, although the general feeling was that it could be useful depending on the assumptions that are made in the analysis.  The fracture spacing was not adequately described.  The fractures are assumed to be perpendicular to the well.

 

Figure 16.      Schematic view of the layout for the fractured horizontal well tool.

 

Other assumptions and procedures include:

 

Ø      single phase, Darcy flow,

Ø      fully penetrating, infinite conductivity vertical fractures,

Ø      finite (closed) drainage area,

Ø      The injectivity Index is computed after pseudo-steady state flow has been reached, and,

Ø      1, 3, 5 and 7 (odd number) fractures are arranged in a symmetric pattern.

 

Sponsor Presentations

Paul van den Hoek provided case studies from Oman, Thailand, Nigeria and Syria.  The first examples were from line drive waterflooding operations in Oman using horizontals as the injectors.  Figure 17 is a plot of sandface pressure versus rate.

 

Figure 17.      Formation face pressure versus injection rate (operations in Oman).

 

Chronologically grouped data are indicated in Figure 17.  Originally, there was probably some matrix injection.  Subsequently, the flat behavior of the curve in Figure 17 suggested fracturing (either induced or injection into natural fractures.  The mobility in this reservoir is low, since the in-situ oil viscosity is approximately 60 md.  Some skin has developed due to water quality.  This is shown by the higher than initial injection pressure for the period of injection after a long shutdown (refer to the labels on Figure 17).  This could also be due to some pressure buildup in the reservoir.  To maintain matrix conditions, the capacity of a 1-km well was only about 1200 BWPD, which is economically unacceptable.  One of the important observations made was the low slope of the curves if friction is correctly represented.

 

Paul then presented a slide on a NAM, showing similar behavior (Figure 18).

 

Figure 18.      Formation face pressure versus injection rate (NAM example), showing likely fracturing behavior at a stress smaller than was originally assumed as well as flat or slightly negative sloped curves – arising when friction is calculated accurately.

 

The next example that Paul showed was from Thailand (Figure 19) where there seems to be fractured injection from day one.  Looking at the chronological variation in behavior, trends move up and down but remain relatively parallel to each other.

 

Figure 19.      Fractured injector in Thailand.

The reservoir pressure was likely impacting the fracturing pressure.  Jean-Louis Detienne asked if the water quality was unchanged.  Paul indicated that it was constant (clean aquifer water).

 

The next Shell field case was from Nigeria.  This is an oil reservoir and is soft sands.  The behavior shown in Figure 20 is very similar to behavior seen in the Elf 3 soft sands well.  It is evident that there were numerous shut-ins and the II increased after shutdowns.  There were probably backflow operations providing some sort of stimulation.  John Shaw questioned whether the backflowing caused the well to sand up.  Paul indicated that indeed the perforations were eventually covered up.  After a CT clean out, it was found that the II was much lower than before.  Marco Brignoli indicated that this did not surprise him and that the sand had been compacted.  Marco indicated that after liquefaction you could see a substantial reduction in permeability.  Bruce McIninch argued that viscous pills might have been another cause.

 

Some argument was made that this was a situation that was taking fluid above fracturing pressure.  Laurence Murray asked about screens and Paul van den Hoek indicated that it was necessary to fracture the well for adequate capacity.

 

 

 

Figure 20.      Chronological variation of the Injectivity Index and the bottomhole pressure for the Nigerian well.

 

Figure 21.      Variation of calculated sand frac pressure with injection rate for the Nigerian well, suggesting that some form of fracturing had probably occurred.

 

Paul next showed measured and simulator data for an injector in Syria.  This is competent rock and the injection water is river water treated down to 10 mg/l.  Simulations were done with varying solids loading and with varying temperature.  Figure 22 is one match.

 

 

Figure 22.      Variation of injection rate with a superimposed rate profile used in various Shell-proprietary simulations (Figure 23).

 

Figure 23 shows simulations for 10, 75, 95 and 110 mg/l and actual data.  It can be seen that the injection pressure can move up or down by 1,000 psi depending on the solids loading.  Recall that the water was filtered to 10 mg/l - at least at the surface.

 

Figure 23.      Variation of wellhead pressure with time in various Shell-proprietary simulations – TSS was a variable.

 

It appears from the simulations that water quality can indeed be a significant controlling factor on pressure and there are substantial economic tradeoffs.  In this case, the operator preference is not to buy new pumps to handle higher pressures but to clean the water.

 

For some time, Shell has been developing a customized produced water fracturing simulator.  It is desirable to forecast fracture and pressure containment.  The simulations can be used as a tool for optimizing injector strategy.

 

Some checking of the PEA-23 relationship was carried out.  Paul emphasized the need for a SRT and that it is essential to calculate the correct slope.

 

The model used calculates conductivity based on a volume balance.

 

Paul concluded by expressing Shell’s future interests.  These include:

 

  1. Comprehensive analysis of Sponsor field data plus proper reporting of this
  2. Water injection fracture monitoring (falloff, HIT,)
  3. Best Practices for injectors in soft formations (completion, operation)
  4. Impact of contaminant on injectivity for fracced injection (‘beyond PEA-23 relationship’)
  5. Fracture growth/containment at material property boundaries (i.e., delamination, containment even if there is not a strong stress contrast …).

 

Laurence Murray then presented various examples.  The first example related to reduction in injectivity due to water hammer effects.  Unfiltered seawater (almost all < 4 microns) was being injected below fracturing pressure into three cased and perforated zones.  In this example frequent shut-ins have been required and reverse circulation has been used for cleanup.  There were successive reductions in this weak, relatively soft formation (E ~ 500,000 psi) due to perforations being covered.  When the sand fill was cleaned out, the injectivity was reversed.  There was gradual fillup that can be characterized by a partial penetration skin.  The modulus is high enough that there actually has been some thermal fracturing.  Performance during injection is tubing limited.  In addition to partial penetration there is also another skin component due to deviation through the reservoir (since the deviation is small the associated negative skin due to inclination is correspondingly small – see Earlougher, 1977, page 157).  If you are careful, fillup can be used to your advantage.  If the rate of fill up is not too rapid, progressive coverage may allow the well to fracture.  It was argued that you can frac the sand away if you don’t come past the top of the perforations.

 

Laurence’s second example was for cased and perforated platform wells.  All zones were covered (single zone) to avoid crossflow-related problems.  One well is at 75 to 80° and the other discussed was vertical.  One of the concerns here was BaSO4 precipitation.  There were substantial concentrations of water-soluble organics.  To manage this phosphoric acid was used.  Using the PEA-23, for a permeability of 200 md ± there was no real evidence of a substantial impact for the water quality being used (from 10 to 20 or 25 ppm).  Since surface filters (cartridge) were blinding off, filtration was stopped and no real performance difference was perceived.

 

The first well was pressuring up the reservoir.  Perforations were added to the second well and it was a well voidage completion.  Laurence argued that Well 1 was fraced and that the fracture closed down with increasing stress levels due to poroelastic effects.  It was emphasized that this can dramatically impact reliable interpretation of Hall plots.  The Hall plot presumes constant kh.  If the height open to the fracture changes (part of the fracture closes), response would be similar to an increase in skin and the plot would be worthless.

 

There was some consideration to back off on Well 2 and to try and clean up Well 1.  The reservoir is weak (E ~ 100,000 to 200,000 psi), the net to gross ratio is low, and the permeability is low (60 md).  Almost all of the injection is at the top.  Originally this was a frac packed produced and it appears that an injection induced fracture probably occur adjacent to the fracpack, going into an unpropped fracture.

 

The next case study related to falloff interpretation in a weak formation.  Both production and injection tests had been done.  The completion was an openhole WWS.  There were waxing problems associated with cooldown in the wellbore and associated problems with the screens.  The screens plugged during injection.  What is required to match pressure transient data for such a situation?  It was matched by using a damaged zone around the well.  There was a different size of damaged zone depending on whether you are injecting or producing.  Dale Walters pointed out the compaction zone ahead of a fracture.

 

Laurence also showed a BP PWRI Best Practices web-based presentation.  One of the unique features was the ability to viewed threaded comments.

 

At this point there was a discussion of skins and how to use them in PWRI situations.  For example:

 

  1. Filter cake ahs no mechanical strength but causes a pressure drop.  Presence of a filter cake can help fracturing.  The fracture grows through the skin; the rate goes up and eventually performance is tubing limited.  Partial penetration may need to be considered because the cake and/or fracture may develop and evolve (or devolve) preferentially.
  2. It may be desirable to complete the well as a partial penetration (actually this is more accurately called a partial completion) so that you can initiate a fracture and you at least know where it starts.
  3. Another skin scenario is internal cake where the differential pressure can inhibit getting pressure into the formation.

 

The obvious related topic is “How should you complete multi-zone wells?”  There was some sentiment that flow control may be a prerequisite.

 

There was discussion about fracture containment.  Laurence indicated that BP has seen it in some cases (particularly Prudhoe Bay) where there is not substantial stress contrast and it could be related to conduction/convection.  Out-of-zone fracturing may ensue.

 

In some wells where sidetracks have been drilled above the main bore; they have gone into areas that have already been cooled by conduction.  Mechanical control may be required.  Laurence will provide an example of this.

 

Another area of discussion was for soft rock, with injection above fracturing pressure – What role does water quality play?  For example:

 

  1. What are the criteria for screen plugging?
  2. Can hot spots develop?
  3. Hot spots may preferentially develop when you fracture because of high concentrated velocities.
  4. What about hot spots in ESSs?

 

Bruce McIninch recounted their experience with screen testing for West Brae.  They found two things.  One was that friction was not significant and two the testing was beneficial because it allowed them to specify a redesign to the manufacturer.

 

John Shaw described two field trials that are being run by Statoil.  The first is the Statfjord test.  John showed a Hall plot and discussed several monitoring issues.  Protocols on one well have included running seawater at 30°C, followed by a 50/50 seawater/produced water at approximately 55°C.  Injection is into the Upper Brent (permeability is approximately 670 md).  There has been a drop in reservoir pressure and some rate increases during the pilot.  The Hall plot showed a significant increase in injectivity with produced water due to viscosity effects.  They are now switching to 100% produced water (75°C).  Step rate testing has shown a thermal stress affect of 1 bar/°C.  The injected produced water is low in oil but high in solids (up to 250 mg/l).

 

The next pilot was Heidrun.  This is soft, heterogeneous and there is a good amount of clay.  The well is cased and perforated in five zones over 45 m, with permeability ranging from 150 to 350 md.  They have injected seawater (30°C) since November 1999 and the rate has been dropped from 3000 to 2000 m3/day.  Subsequently they switched to produced water at 60°C.  The water quality is extremely poor.  If you evaluate the Hall plot, for the initial seawater part, you see two distinct slopes.  Another seawater injection seemed to show a second breakdown.  A produced water test was then done.  It seems as if a fracture has been established and is taking fluid.  There is not classical matrix behavior (the rate stays up).

 

Idar Svorstol then presented on Norsk Hydro's work on Snorre.  43 wells have been developed so far and a new platform will be installed in the North - with 16 producers and 10 injectors.  Production will start in August 2001.  These wells have intelligent completions.  This entails sliding sleeves and bottomhole gages.  The program also includes cuttings reinjection. 

 

Current Snorre capacity is approximately 40,000 sm3/day.  Some of the issues on Snorre are environmental - the deoxygenation can't handle any more seawater.  In addition, Snorre involves high and low permeability zones and problems might be anticipated for lower water quality situations.

 

Jean-Louis Detienne indicated that he will be getting two or three more field cases but they will not be available soon.  Jean-Louis expressed his opinion on the requirements for the overall PWRI project.  He indicated that it is necessary to process all of the available data and to get the tools into the hands of the Sponsors and to give feedback on the Toolbox.

 

AGIP provided new data to the Contractors for an aquifer injection situation (it is a good case study since one well is fractured and one is probably taking fluid under matrix conditions).

 

Laurence Murray provided another example - relating to openhole gravel packed wells.  Underreaming before gravel packing was designed so that there would be enough pack for fines to accumulate in.  Water hammer cleanup was also discussed.

 

Trond Jensen discussed Ekofisk seawater injection that has been carried out on the flank for 12 years.  Pressure has increased and injectivity has gone down.  This is for clean, deoxygenated seawater.  It is batch treated with biocide, ultraviolet exposure and filtration down to 2 microns.  The Hall plot shows a change in slope.  Trond is uncertain as to whether the change is due to real skin development or to a change in reservoir pressure, although he favors the former.

 

There was further discussion of souring evaluations on pilots and whether backflowing would provide substantive information.  John Shaw felt that it was possible but he was skeptical.

 

Ahmed Abou-Sayed presented several options for development of an economic module for PWRI.  These were:

 

Option 1. A comprehensive model, based on the platform and database provided by Paul van den Hoek.

Option 2. A spreadsheet that is a stripped down version of the Shell model, with comparative costs.

Option 3. A spreadsheet with categories but no comparative cost information.

Option 4. A checklist based on the flowchart presented by Advantek at the December quarterly meeting.

Option 5. Abandon the Task.

 

Sponsors expressed their preference by vote, as indicated below.

 

  1. Marco Brignoli indicated that AGIP's preference would be not to have an in-depth tool.  He opted for Option 3.
  2. Jean-Louis Detienne indicated that he was interested in the flow chart structure that Advantek had presented at the December Meeting.  He indicated a preference for something simple (Option 3).
  3. Idar Svorstol indicated that while he did not have a strong opinion on this particular issue, his preference would be something that is not too detailed and would provide a rough idea of the economics.  His vote was for Option 3.
  4. Henrik Ohrt selected Option 4 in order to safeguard (be sure that previous efforts are available) what has already been done.  He is just interested in a checklist.  He would accept Option 5 but would vote at this stage for Option 4.
  5. John Shaw voted for Option 5, but "could live with Option 3 if necessary."
  6. Laurence Murray opted for Option 4, to have a methodology for just seeing if all of the bases are covered.
  7. Paul van den Hoek strongly supported Option 1 and would not go down below Option 2.  "Anything else is a waste of time."
  8. Trond Jensen indicated his interest in comparative costs but he believes accomplishing this is not practical.  His vote was for Option 4 and Option 3, if absolutely necessary.
  9. Bruce McIninch only wanted a checklist (Option 4).

 

Options 3 and 4 had three votes each.  The decision was made to opt for Option 4.

 

Based on a suggestion from Henrik Ohrt on the previous day, John McLennan tried to outline some of the accomplishments and future requirements for the JIP.  This presentation is available.  This presentation can be summarized as follows.

 

Five Key Findings

 

  1. If reservoir pressure is not reliably considered and if formation face pressure is not reliably represented, any sort of diagnostic or predictive methods are potentially in significant error.  This is a trivial but essential concept.  Tools are available for representing these features in fundamental, continuous injection monitoring evaluations.

 

Consequences of not reliably knowing formation pressure and formation face pressure include inability to plan when treatments should be carried out and to recognize fundamental degradation in injectivity, misinterpretation of pressure transient data, inadequate/inappropriate support, incorrect decisions on the type of intervention that is required, if any, etc.  Many reservoirs are being fractured and it has not been appreciated - except by specialists - sweep considerations, etc.

 

  1. The PEA-23 concepts can be extended/modified to quantify discrete events representing extension/growth of fractures/discontinuities even in unconsolidated formations.  More quantification is required in the next three months using available data.  Data are available for calibrating available models.

 

Consequences:  Design tools and “vision,” stimulation planning and optimization of investment for maintaining injectivity

 

  1. There can be a point of diminishing or no return in stimulation beyond which further “conventional” stimulation will provide progressively less effective results. 

 

Consequences:  Methods are being developed to represent this during the current phase of the project using back-analysis of available injection history.  Stimulation types can be refined – using this and the concepts in Key Finding 2.

 

  1. Near-wellbore skin, particularly completion skin is an under-appreciated feature in comprehending injection programs. 

 

Consequences: It can be huge in soft formations and misrepresent required stimulation/intervention.  It can possibly be manipulated to achieve improved conformance.  It may entail significant perforation fillup in soft formations and its magnitude and impact can vary depending on whether injection is above or below fracture opening/reopening pressure.

 

  1. The complexity of fracture growth in consolidated and in soft formations is such that planar, single fracture, constant height or ellipsoidal fracture models may be inadequate in many cases.  Consider cases where they will be inadequate – in layered reservoirs with significant stress or permeability contrast where fracture or matrix fingering dominates, in weakly consolidated reservoirs, in reservoirs where there are multiple fractures that are created by pressure cycling (does not even need to be batch injection), in untuned (the injection program has not been fine-tuned for optimum performance in a specific reservoir based on historical response) reservoirs with injection just below fracture extension pressures, in naturally fractured reservoirs treated at low (or possibly cycled) injection rates.

 

Consequences:  Cautious or selective use of LEFM (linear elastic fracture mechanics) is essential.  The wellbore and the completion must be reliably represented.  Disposal domain concepts may be a pre-requisite.  Vertical fracture growth must be represented.

 

Five Key Achievements (not prioritized)

 

  1. Availability of Data

 

·         Acquisition of data!

·         Archiving of data!

·         Using data!

·         Although it has been slow, the development of the database and the interpretation of the data are approaching the point where information is being more effectively archived and processed. 

·         Before the end of this phase, more needs to be completed.  Much of the intellectual maturation in this project has come about from looking at trends in and features of available data by a diverse group.

 

  1. Development/Organization and Ultimate Deployment of Analytical Tools

 

          Examples of tools developed include:

 

·         Completions Worksheet

·         PWRAD

·         Multilateral/Multifracture Models

·         Stimulation Selection Tool

·         Thermal and Poroelastic Stress Tool

·         Data Management Capabilities (BHP, plotting …)

·         Others

 

  1. Knowledge Management/Best Practices/Assurance

 

Consolidation of available findings is being accomplished by incorporation into guidelines.  Key issues are being prioritized as Best Practices.  These include supplementary information for design and outline, where possible, procedural limitations.  Methodology has been developed for logically archiving and providing accessibility.

 

  1. Workshops, Exposure and Interaction

 

While it is somewhat more philosophical, Workshops and Quarterly meetings have facilitated:

 

·         Presentation of case studies

·         Shared experience and platform for communication

·         Constructive criticism by a diverse audience

·         Exposure to experience from other disciplines

 

  1. Recognition of Concepts, Methods, Limitations and Future Requirements

 

          Examples include:

 

·         Definition of inadequacies in current testing methodologies (e.g. SRT, falloff in layered formations)

·         Definition of limitations in modeling methodologies

·         Consolidation and outlining of various options for completions and stimulation (e.g., conformance methods, procedures for bringing horizontals on-line, the future role of intelligent completions, ESSs, fiber optics, indirect appreciation of the economics of surface treatment versus stimulation, possible occurrence of a compacted zone …)

 

Main Recommendations

 

Recommendations were presented as to what some of the future or ongoing work might be.  This is a Contractor's perspective and does not necessarily reflect the position of the Sponsors.

 

  1. Maintenance and Updating of Tools, Best Practices, Database

 

This would seem desirable as data has driven much of the project (database).  Even through the duration of the project, opinions and conceptual appreciation have evolved.  Tools developed in the Project were not necessarily envisioned at the outset.  Additional, high quality data are anticipated from upcoming or ongoing pilots and field programs.  This would also help to maintain communication among a diverse group

 

  1. Case Studies

 

·         Each Sponsor will be requested to continue to provide field data.

·         These data may in fact be superior to some of the historical data that have been provided because of improvements in instrumentation, acquisition, and experience/knowledge.

·         Information will be incorporated into the database system and evaluated with specific intent of testing/improving tools that have been developed, conceiving/recommending/developing of new tools, concepts, learnings and identifying physical behavior.

 

  1. Workshops

 

·         Two workshops were suggested – one after six months and one after nine months?

·         Each of these would be designed to look at new technologies and at new and recent developments or events.

·         It was recommended that the first Workshop cover horizontal injectors and that the second Workshop would incorporate important aspects from all of the Tasks in the first phase of the JIP.

·         There would be quarterly meetings, two of which would coincide with the Workshops.

 

  1. Miscellaneous

 

·         Accounting /Contractual

·         Management: Recording and posting of minutes, coordination, interaction with Sponsors on performance issues and interaction with Contractors to improve delivery, etc.

·         Travel: Will need to be preauthorized for four travel sessions (quarterly).

 

  1. New Tool Development

 

The following four areas of Tool development are described adequately in the presentation.

 

·         Diagnostic Methods (e.g. PTA methods)

·         Predictive Model (PWSIM)

·         Completions Skin Tool

Thursday March 29, 2001

 

Sponsors met and reviewed the requirements for proceeding forward to complete the existing Project and evaluate mechanisms for developing additional deliverables with the funding remaining in the Project.

 

Laurence Murray summarized the results of the Sponsors meeting.  Laurence’s remarks are summarized below.

 

Project Administration:

A firm proposal is required.  A new contract is required that gives a flexibility to the work beyond June 30th.  Contracts stop at the end of March.  There is no mechanism for payment.  GTI have given a proposal as an interim solution - at least until the end of June.  This has been rejected because of a monthly cost of approximately $10,000.  The Sponsors are amenable to a small fee to GTI to transfer the contract to another party.

 

Mentors:

To facilitate proper completion of the initial phases of the project, a group of Mentors was agreed upon.  These are:

 

Task

Mentor(s)

Monitoring and Prediction

Laurence Murray and Mark H.

Matrix Injection

Paul van den Hoek and Henrik Ohrt

Soft Formations

Jean-Louis Detienne

Stimulation and Mitigation

Henrik Ohrt and Paul van den Hoek

Layered Formations

Trond Jensen and Idar Svorstol

Horizontals

 

Database

All Sponsors

Surface Systems

John Shaw

Validation

John McLennan

Best Practices

All Sponsors

Web Site

All Sponsors

 

Requirements for Completion:

For all Tasks there are some basic requirements.  Some specific requirements were required for individual Tasks.  For example”

 

Ø      For all Tasks: An essential requirement is final reports, including descriptions of relevant tools with worked examples.

Ø      For the Database, a User Guide is required.  The database needs to operate as a standalone PC version and there needs to be an equivalent web-based version.  This is in effect and Brian Odette will continue to see that it is implemented.

Ø      Best Practices: Some additions need to include an overview of the JIP and a summary of the achievements.  It must also contain guidelines and relevant reports and documents.

Ø      Website: As with the database, there needs to be a web version and a CD version.

Ø      In the minutes, footnotes or descriptors need be added to the viewgraphs and slides.

Ø      For the Validation Task, it is necessary to incorporate field testing results where they are available.  An Opportunities List needs to be developed and circulated. 

Ø      Surface Systems: Final reporting is required, as is a Users’ Guide.  The Surface Systems component of the project needs to be web- and CD-based.  The final report needs to cover some of the technical issues.  What are the key things that you need to be looking out for?  Some sort of guide for using.

Ø      Validation: As field data comes in, evaluate these data and see what is derived and make the results known.  There needs to be feedback on the validity of tools, models, etc.

Ø      Toolbox: The requirements include a finished working version, including a Users’ Guide that incorporates a technical manual for the tools.

Ø      Best Practices: Incorporate a JIP Overview, achievements, guidelines and documents.

Ø      Website – can also run off a CD.  Also, where data have been interpreted, it is important to clarify the analytical methodology. 

 

Future Work:

A firm proposal is required.  There are some minimum requirements, as listed below that need to be accomplished and/or continued before the end of June.

 

  1. Website, toolbox, database guidelines – working finished version by the end of June.
  2. Web site maintenance
  3. Database maintenance

 

Using the oversubscribed funds, the Sponsors will consider proposals for work from July 1, 2001 to June 30, 2002.  Some of the components need to be:

 

  1. Database and Website Maintenance – to maintain the current status and to add information to it.
  2. Upgrade of Toolbox, Best Practices and Existing Guidelines It is important for the Task Mentors to actually study and test to identify bugs, errors, omissions …  Upgrading would entail increasing the functionality or implementing new tool(s).  An example of upgrading the Guidelines would be to indicate if somebody has carried out a new technique and has new experience.  Consolidation into existing Guidelines would be required.

 

Beyond these basic Tasks for continuation, there is motivation to carry out certain additional or new Tasks/Developments.  These could include:

 

  1. Pressure Transient Analysis: Starting from the Monitoring Workshop in Denver, it was evident that improved pressure transient analysis methods are required for single and multiple layer scenarios.
  2. Predictive Model: During the first phase of the Project, comprehensive predictive model development was discouraged in order to focus on evaluation of data.  There is some feeling that a more comprehensive tool could be proposed.  Laurence Murray indicated that he supports some sort of standalone tool that everyone (including their partners) can use, possibly a module that can be put into a commercial simulator.  There was discussion relating the costs and scope of such a model.  For example, Paul can den Hoek felt that models were going to cost hundreds of thousands of dollars to couple to reservoir simulators.  In any submitted proposal, it was stressed that it was important to identify the functionality of the proposed model (e.g. local grid refinement?, wellbore management functions?, …).
  3. Completions Skin Tool: One of the issues that has been strongly identified during the course of the Project has been the importance of specific completion characteristics.  One possible component for future work was identified as a tool for characterizing completion behavior, as has been done to a certain extent for perforations.

 

During the in-camera session, the Sponsors prepared a prioritized list of desirable components for future work.  This was not passed onto to the Sponsors.  Laurence Murray indicated that this would be circulated to Sponsors (by the middle of April) so that a Sponsor position could be developed on how to usefully spend the remaining money.

 

Date for the Next Meeting:

The meeting was scheduled for the last week of June, in Houston.  The dates would be Monday, June 25 through Thursday, June 28.  All participants are requested to plan on attending the complete session.  A hosting organization will need to be identified.

 

Prior to the meeting, it is desirable for Sponsors to review the current Project and, in particular, for Task Mentors to dig into some of the details.  Any Sponsor review should be given back to the Contractors by the first of June.