Minutes PWRI Meeting
Edinburgh - June 3-4, 1999
Thursday, June 3, 1999
David Davies opened the meeting. Discussion of certain administrative issues was deferred until all Sponsors were present. The attendees are indicated in Appendix A.
Alastair Simpson summarized the available data. Specifics are provided in Appendix B. More data are anticipated. Sponsors were polled on additional data availability.
Ron Harisch presented data that had been provided by ARCO. DE&S had previously performed simulations representing fracture flexure and growth in response to temperature variations accompanying seawater and produced water injection. Basic data are available or can be acquired. Some issues were raised.
Ghassem was concerned about Young's Modulus and the sensitivity to thermal fracturing. What sort of moduli were there and was thermal fracturing a significant issue after all.
Jean-Louis and Nick K. argued that the thermal distribution in an open fracture should be more uniform. John McLennan disputed this. Maybe width profiles or something would help. They were concerned about the encapsulated cooler zones.
Nick Koutsabeloulis presented simulations of BP's Y-05 Prudhoe Bay well. This is an horizontal well over the tar mat. Previous simulations by V.I.P.S. had been carried out to determine the evolution of "fractured" zones from three perforated intervals in this well. The simulations shown were undertaken to indicate the influence of rate and completion configuration (layout of the perforated intervals) on the nature of the "fractures" created. Some details of the constitutive models used were presented. Sponsors indicated that clarification of the mechanics of the modelling would be helpful. Nick indicated that a formal report would be forthcoming.
Jean-Louis Detienne presented information from onshore case studies (carbonates?) for both thermal and "forced" fracturing. One layered case was quite significant. It strongly demonstrated the preferential occurrence of thermal fracturing in the lower permeability zones. It was also demonstrated that Elf has been able to successfully use radial flow models derived from PEA-23 findings. The one situation shown could not tolerate extended radial flow because the injectivity decline was too rapid.
An open session summarized some of the issues that had been brought up. These included:
Ron Harisch presented simulations that DE&S had performed for Maersk in the Dan Field (MFB-07). The well had seawater injection followed by produced water injection. The Hall plot showed a significant change in slope and Ron will investigate this further. The strong message from this presentation was that accurate interpretation of cumulative injection data as well as short-term testing information was critically impacted by the reservoir pressure. Two inferred scenarios were compared highlighting the errors that could occur. Temperature and reservoir pressure data are required. Contractors will use public domain software for temperature forecasting bottomhole.
John McLennan presented a preliminary evaluation of Gorm data. The focus was to delineate the available data, to indicate what sort of additional, supporting information would be desirable and to emphasize the importance of bottomhole, chalkface data. The minimum data set information had indicated treating was below fracturing pressure. When the available surface data were corrected to the best approximation of bottomhole data, there was strong evidence that some of these wells were fracturing. The other important aspect of these wells was the repeated saw-tooth behavior as plugging apparently occurred with time (a progressive loss in injectivity) and was temporarily restored with stimulation operations. Kitt Ravnkilde presented further information on the stimulation activities used (subsequently). Chronological stimulation and water quality data were solicited. Additional data will cover water quality, seawater injection, production history, wells from the west side possibly, voidage plots, completions diagrams … and certain reports. Contractors were advised that the data from N34A is probably unacceptable because of bypass that had been caused by an earthquake.
John McLennan presented data from PanCanadian's Rosemary Field. The purpose of the analyses carried out were to indicate the importance of using the best possible information for inferring bottomhole and sandface pressure and to begin the exercises in evaluating data to see if additional information can be extracted from testing data. Also presented were preliminary guidelines for conducting step rate testing for PWRI purposes. One of the recommendations had been that if bottomhole gauges are run they should be used in tandem. Doug Ashford indicated that PCP's extensive experience was that they only had gauge malfunction or inaccuracy in one of a hundred or so cases and that they would generally not run dual bombs. It was pointed out by others that the failure rate was often much higher in certain more extreme reservoir conditions (>100°C was cited) and the cost of failure outweighed the cost of duplication. The guidelines will be revised to acknowledge deferral to local experience and costs. One of the step rate tests shown displayed a negative slope in the post-fracture regime. It was argued that this could be due to a number of reasons. Christine Valvatne pointed out that a cause for this could be incorrect friction pressure calculations in the fracturing domain. This could very well be the case since the friction pressure calculations that were used were strictly for radial flow and these had been applied to the fracturing regime. The contractors will fix this error.
John McLennan presented a re-evaluation of the PanCanadian Countess data. At the meeting in The Hague, preliminary evaluation of these data had indicated very high treatment pressures with respect to the vertical and horizontal in-situ stresses. Since that meeting the contractors have developed a consistent methodology for accounting for tubular and perforation friction pressure losses. Gravel pack will be added. This was discussed further by Ron Haresch. The PanCanadian data now appears to show "sub-critical" fracture propagation where sandface pressures "dance" around the minimum horizontal stress (inferred from the Rosemary Field) - sporadic propagation is followed by plugging periods and the cycle is repeated? The maximum temperature difference may be 30°C. Water is stored in transfer tanks before reinjection and could reflect significant seasonal temperature fluctuations. Contractors will assess if this is what is shown in the pressure fluctuations.
Ron Haresch summarized the methodologies used for the friction calculations used by the contractors. Standard methodologies are used in the tubulars. Friction pressure in the perforations is from standard service company methodologies. The issues of open perforations, the extent of perforation damage and how this impacts friction and injectivity were highlighted. The extreme difficulties in evaluating friction factor were outlined. Depending on the particular tubulars, Statoil may default to 0.0015, Conoco to 0.0006, Shell would be similar to Statoil and Conoco, and Maersk's big problem in Gorm is progressive reduction in ID. Maersk is planning step rate testing before and after an upcoming scale removal operation.
Friday, June 4, 1999
Jean-Louis Detienne presented case study data, showing some possible characteristic behavior where pressure-rate plots showed a negative slope after fracturing had occurred. It was indicated that there could be a number of reasons for this - including reservoir support, height growth etc. As was pointed out by Christine Valvatne the previous day, this could also be due to incorrect frictional relationships after fracturing (at least for the data shown the previous day by McLennan).
After Jean-Louis' presentation J. McLennan presented an administrative overview. The key issue discussed was the participation of AdvanTek. John Shaw indicated his disappointment in the lack of proactivity of various participants and indicated his disappointment that this had taken three months to be resolved. The participation issue was voted on and it was agreed that AdvanTek should participate.
The issue of mentors was discussed. It was agreed that the contractors would interact with individual Sponsors to reach agreement on mentoring (refer to Action Items at the end of the minutes).
Doug Ashford indicated that he had been very disturbed by the previous morning and was somewhat more relieved when he saw more concrete events in the afternoon (refer to Action Items).
In polling the Sponsors, there was a reiteration of the Sponsors desire, as had been expressed the previous day by Ghassem Gheissary that the data be consolidated and appraised overall.
The issue of PEA-23 data availability was brought up again.
All of the Sponsors indicated that stronger project management was essential. John McLennan agreed to rectify this on the contractors part (refer to Action Items). A clarification of the role and responsibility of GRI was requested.
Co-Chairpersons for the Sponsors were proposed as John Shaw and Laurence Murray. John indicated he would speak to Laurence on this issue.
There was further discussion of additional data that was anticipated. Jean-Louis will provide case study data (some sanitizing may be required). Brage field data and Shell data were confirmed.
John McLennan agreed to provide John Shaw with comments on the Statfjord pilot test.
Jean-Louis discussed the issue of benchmarking and his willingness to run his models in conjunction with the contractors. There is also apparently Shell data from Eider. The models were discussed further (in conjunction with Task 3). When do the various models work and when do they not? What are the benchmarking cases to be run? Christine Valvatne stressed that the emphasis needs to be on using the appropriate available tools when it is appropriate (refer to Action Items). Jean-Louis emphasized that by benchmarking, he did not mean model development. He did not want to develop a reference model. Instead, he stressed the importance of delineating the appropriate available tools. He indicated he was anxious to see Nick Koutsabeloulis' report so he could more fully understand the fundamental basis of Nick's model. He also indicated that Sponsors could run a selected number of simulations using their proprietary codes on selected datasets with feedback to their colleagues and the contractors. Can BPOPE be run?
Anne Skjarstein requested that a status report be provided (refer to Action Items).
Bill Beckie indicated that PanCanadian would be providing more detailed information in the next two to three months. This data would be appropriated into the database. Data from the Weyburn and Winter fields would be provided.
Ghassem Gheissary indicated that modelling work was available. Alastair will continue with follow up.
Kitt Ravnkilde and Jim Hardy discussed the available Maersk data (data requirements were further clarified with Kitt after the meeting). Maersk's focus is on water quality. More Gorm and Dan data will be provided. The Dan data will compliment the data that DE&S has already been working on.
Kitt presented data on Gorm stimulation programs.
Doug Ashford suggested that there should not be additional discussion on submitting additional data and that a cutoff date should be set for the end of August. After vigorous debate, it seemed like the impression was that Sponsors should endeavor to provide data prior to this time so that it could be incorporated in the data processing before the next meeting but that if and when additional data comes in it will be incorporated into the database, processed and contractor conclusions/observations/recommendations/etc would be updated, if appropriate, - based on this new data. The message was that there must be ongoing population of the database but every effort should be made to promptly supply existing information.
Birgitte Schilling emphasized the need for data processing to provide usable deliverables.
The strong interest in Soft Sands was reiterated with an emphasis on Standardization of Methodologies for this and for other topics. The issue of Sponsor focal points was brought up again.
John Shaw stressed to contractors again the fact that he had not received specific technical queries in the time period since the meeting in March and he emphasized that the contractors need to interact more specifically with the sponsors. Several others indicated that it was important to provide specific questions to facilitate answers, opinions or review by Sponsor specialists.
The viscosity dependence of step rate tests was discussed.
John McLennan showed some Snorre data. With clarification from John Shaw it seemed evident that unlike the Eider data, viscosity was the dominant thermally regulated controlling parameter. Stress-related thermal effects were perceived to be second order. Mechanical properties information (such as Young's Modulus) would be desirable. Did Anne have this? Is Snorre data available?
"With soft sands, the real issue is to keep it open and conductive."
Some questions were raised on fracture diagnosis on soft sands. The concern for this came with the diagnosis of fractures in the Snorre pilot where S2 has a permeability of 1200 md. What are the limits for fracture detection? Can some guidelines be provided on HIT testing? Is there an issue of formation variability and the fracturing only occurring in more indurated lithologies? John S. emphasized that Heidrun was quite variable. Statoil and Saga have HIT data, as have BP (Laurence Murray has already provided some). It was indicated that a Pinnacle representative had said that the resolution of HIT data is poor for features with lengths beyond about 10 m. This observation needs to be confirmed. It was further indicated that PDO has done HIT testing and that short fractures were detected and that the results agreed fairly well with the models. HIT testing has also been performed on Forties. Dendritic fractures were apparently indicated. This will be confirmed from the information that BP has provided. A fracture or something similar is detected. Discussion switched to Statfjord where fracturing in selected intervals may be occurring? Statoil, Saga, PDO and BP Amoco have HIT data.
Alastair Simpson brokered a discussion of standardization. The contractors will organize their methodologies of storing and accessing data. Full stop. No further discussion. The preliminary guidelines that John M. prepared for step rate testing will be revised and sent out to all Sponsors for review by themselves and their specialist colleagues. Observations by Doug Ashford and Christine Valvatne on gauge reliability will be incorporated.
Ghassem indicated a potential Shell pilot in Nigeria where a major issue is soft sand.
Other discussion focused on:
Key issues that were stressed were:
Action Items
The following are some of the "agreed" upon action items, either as part of or in addition to the proposed deliverables.
Appendix A Attendees
Appendix B Available Data