EXPLORATION Raster logs may be basis for a geologic workstation

Scott L. Montgomery
Petroleum Consultant
Seattle

Use of computers in exploration geology has expanded during the past decade to include many areas of data collection and analysis.

Computerization is now viewed in the petroleum industry as essential to such activities as: mapping and contouring; core analysis; use of production and test data; mineralogical studies; communications; and more. This has led to increased efficiency, better reliability and standardization of data, and progress in interpretation.

Despite such advances, however, the most fundamental data type employed in petroleum geology-well logs-are still predominantly used in traditional, hardcopy (paper) format. This fact, if unchanged, places a strict upper limit on future advances in geologic work related to computer technologies.

Industry personnel have increasingly recognized the gap that exists between the geologic "paper mill" and the seismic workstation. Since most exploration, sooner or later, returns to questions of geological interpretation, this gap has important implications for companies trying to increase efficiency and cut costs.

A universal need

At present, for example, generating seismic profiles from processed 3D data requires a simple point-and-click operation and only minutes to complete.

By contrast, geologic cross sections employing paper logs require a cumbersome series of manual steps including:

1. Retrieval of logs from log library, or ordering them from a provider;

2. Picking tops, marking interval(s) of interest;

3. Copying of each individual log;

4. Cutting and pasting of log intervals onto a chosen datum;

5. Sending the rough section to drafting for finishing;

6. Final proofing and correcting of finished draft.

This process commonly takes from 1-2 days to as much as a week for completion, depending on the number of wells and degree of detail involved. This assumes a lack of any complicating factors-logs checked out, missing, or otherwise not immediately available; different scales for different wells; the need to add new wells or replace old ones; and so forth.

The capability to display logs, pick tops, correlate, produce both rough and finished cross sections on-screen, in rapid succession, would clearly mark a major advance over existing manual procedures.

A geologist's ability, for example, to produce ten or more cross sections in the time now required for one section in paper format would obviously have a profound impact on productivity. This underscores the need for a universally accessible workstation for the average geologist.

Such a workstation would put into digital form, and make available to various forms of manipulation, all major sources of data used in geological interpretation: core data, DST data, mud log data, completion information, production data, and above all, well logs. The benefits of a workstation for the average geologist have been widely recognized in the industry and have helped kindle efforts to create digital representations of well logs.

On-screen well logs

As outlined recently in these pages by Cisco,¹ two approaches exist for creating computerized screen versions of well logs: vector logs and raster logs.

Vector logs consist of fully digitized log data, with every point on the log curve given its digital equivalent. Raster logs, in contrast, are scanned versions of paper logs; they consist of visual images available for digital use.

Until now, vector logs have been the only means capable of creating a workstation environment for the geologist. Being entirely digital, they permit on-screen display capabilities for individual and multiple logs, plus the entire array of standard log calculations. They thus present a nearly complete online data tool for geologic analysis.

Vector logs, however, are time-consuming and expensive to produce, commonly running $100-1,000/log.1 This places them beyond the reach of most companies as a regular resource. Moreover, most industry applications require log analyses for selected intervals only, thus much of the digitized data for each log is unnecessary and cost ineffective. Also, digitized versions of older logs are often less practical, since contemporary analysis cannot be performed with the relevant data.

For these reasons, vector logs are likely to remain an important element of, but not comprise the universal basis for, a geologic workstation environment.

Raster logs, on the other hand, cannot be used for digital calculations, but are well suited for the full range of visual analyses employed in log work. This includes:

1. Picking tops and intervals of interest;

2. Making inter-well correlations;

3. Delineating and tracing signatures for specific facies, flow units, etc.;

4. Making basic estimates of lithology and porosity; and

5. Producing cross sections.

A considerable advantage to raster logs vis vis vector logs is their cost: at around $15/log they are equal in price to paper logs currently in use throughout the industry. Furthermore, since vector logs are often priced by the foot, this cost differential increases considerably for deeper wells.

A second possible limitation to raster logs is that their quality depends upon the original paper version used for scanning.

Raster logs benefit from offering the average geologist a form with which he or she is already entirely familiar. As a result, learning to work with these images on-screen requires a minimum of adjustment, particularly if the relevant software is user-friendly. On the whole, therefore, the low cost, broad use, and familiarity of raster logs makes them the obvious choice for a semi-universal replacement of paper logs.

Current raster displays

Recent advances in raster software technology make such replacement possible.

A significant new tool, known as "intelligent" raster technology, is now available for geologists in the industry and goes a good distance toward creating a universally accessible workstation environment.

Developed during the past year by Interpretive Imaging, Denver, this new technology is being actively employed in a number of projects by geologists working for independent and major operators. Such projects include reservoir and prospect studies in various provinces of the Rocky Mountain, Midcontinent, and Gulf Coast regions.

Raster logs by Interpretive Imaging are depth-calibrated and come with standardized tops provided by the company's Well Log Delivery System (WLDS). Wells chosen for display appear simultaneous in a data list and on a map in a cross section line.

Well order, and thus orientation of the section line, can be changed with a single click. Where available, a log set can come complete with mud log data, scout ticket information, core photos, etc.

On-screen logs have the following display capabilities. Either single or multiple logs can be placed on-screen at any specified vertical scale, with multiple logs at various interwell spacings, including distance-proportional spacing. Entire logs can be displayed and scrolled; they can also be enlarged for close-up viewing, or reduced. Alternately, portions of logs showing selected intervals (reservoir zones, flow units, facies intervals, etc.) may be shown, correlated, and hung from a chosen datum.

Cross sections produced at one scale can be changed instantly to a different desired scale. WLDS software allows selected intervals to be demarcated and will automatically enter these into the tops database. This database, meanwhile, is formatted for use in most mapping software packages now widely employed in the industry. Additional software is available to perform raster-to-vector conversions of intervals for detailed log analysis.

Cross sections can be hung from any previously designated datum. New sections may be produced in a matter of minutes or seconds by choosing a different datum. Stratigraphic cross sections can be converted to structural sections and vice versa.

The technology is designed for use on a normal PC (personal computer) running Windows95 or Windows NT. It therefore requires no specialized equipment, as is the case with many seismic workstations.

There are limitations to consider. While cross sections constructed from existing WLDS technology are suited to both on-screen and hardcopy interpretation, the software has limited drawing capabilities and produces sections that are preliminary in comparison to those created by final-stage drafting (Figs. 1-2). Future software releases by the company, due this year, will address this specifically.

Raster logs

Raster log technology is currently in use for reservoir evaluation, facies correlation, pay zone assessment, and prospect generation in a number of provinces, including Piceance basin; San Juan basin; Denver basin; Paradox basin; Hugoton embayment; East Texas basin; Midland basin; and Delaware basin.

San Juan basin

Amoco has applied the technology to several ongoing projects in the San Juan basin.

One such project involves a reservoir study of Late Cretaceous formations, notably the Dakota, Mesaverde, and Pictured Cliffs intervals, each containing multiple pay zones. In the central basin area, well density is 450-500 wells per 36 sq mile township, making identification and correlation of pay sands for possible pay-add opportunities a time-consuming process.

Raster logs have permitted this work to be done on-screen, with rapid picking and correlating of pay zones, both in dip and strike directions. Depth intervals (thus thicknesses) for each pay zone are automatically entered into the well tops data base. Selective vectorization is done to determine porosity, oil/water saturations, and other reservoir parameters.

When complete, the entire data base will be exported to a reservoir simulation package for detailed evaluation. Use of raster logs for the visual analysis portion of the project has resulted in an estimated productivity increase of at least 7-10 times (i.e. 7-10 logs can be processed in the time previously required for one log).

Hugoton embayment

Exploration for new and overlooked reserves in Lower Pennsylvanian (Morrow) fluvial deposits of the Hugoton embayment has benefited from use of raster technology to identify and analyze channel facies.

Such work benefits considerably from an ability to make multiple cross sections in a short time.

As shown in Fig. 1 [50369 bytes], regional stratigraphic correlations are first established to help locate a potential valley system (Fig. 1A). Re-hanging the section, using the top of the valley system as datum (Fig. 1B), both confirms the presence of channel sands and delineates depositional relationships, which in this case involve changes in the degree of channel incision and thickness of overbank deposits.

A final structural section is produced to observe present-day relationships (Fig. 1C). Each of the latter two cross sections was created in less than 5 min. with the original stratigraphic section built in less than a half-hour. This has allowed a large number of sections to be constructed, yielding significantly improved understanding of channel occurrence, orientation, thickness, and so forth.

West Texas example

Raster logs were also used to examine the Permian (Guadalupian) Yates sandstone for possible infill gas drilling in giant Kermit field, Winkler County, in the Midland basin-Central Basin platform area.

Operated by Parker and Parsley Petroleum Co. of Midland, Tex., Kermit field is a large oil pool with a significant gas cap and scattered gas production. Raster technology was used to investigate possible correlations between gas production and 1) structure; and 2) maximum sandstone porosity distribution.

Fig. 2 [47230 bytes] shows the screen display of 42 gas wells included in the study. A cross section line is indicated for the first four wells that appear in the Display List box (lower right), and a printout version of the relevant on-screen section itself is given in Fig. 3 [35354 bytes] Datum is top of the Yates, and zones with log porosities of 20% or greater are marked.

As shown, wells 1 and 3 have the thickest net sandstone (30-33 ft) with porosities in this range. Comparison of this section with the map of Fig. 4 [123513 bytes], displaying Yates structure and gas EUR (estimated ultimate reserves), suggests that no correlation exists between reserves and either structure or net thickness of high porosity sand (compare, especially wells 2 and 3).

The raster log cross section of Fig. 3 [35354 bytes], produced in less than 30 min., makes it clear that high porosity sand zones have localized development and poor lateral continuity, suggesting that depositional/diagenetic factors are determining controls on gas occurrence and reservoir quality.

Benefits

As demonstrated by the previous examples, "intelligent" raster logs permit the geologist to do several things rapidly:

1. Identify and correlate zones of interest among a large number of wells;

2. Construct multiple cross sections with different orientations;

3. Re-hang individual sections on a range of separate horizons, thus allowing for quick evaluation and mapping of different pay zones.

Producing more cross sections in less time has the inevitable effect of adding geologic information previously inaccessible due to time and economic constraints. Such data, in turn, can support superior geologic and thus engineering models of reservoirs or pay zones.

Based on a survey of geologists currently employing "intelligent" raster technology daily, estimated productivity increases range from 5-20 times, depending on the specific project and activity (Table) [22034 bytes]. Such estimates, while qualitative, derive from comparisons of the average time spent performing each activity using paper logs vs. raster logs. A commonly cited benefit to raster technology, essential to increased efficiency, is its elimination of "drudgery work," thus freeing the geologist to generate more information and perform more interpretations.

Other cost benefits result directly from the digital format of raster logs. Such benefits include reductions in paper use, drafting requirements, and technical support to manage well log data. Current needs with regard to paper log storage space, preservation, and replacement are abolished. More important from the geologists' standpoint, all logs for a particular basin, play, or field are instantly available at all times, either on CD (compact disk) or a company computer network. This permits simultaneous use of individual logs and log sets by different personnel. Increased control over data availability, plus enhanced log display and cross section building, are also cited as sources of increased interest with regard to previously routine geological tasks.

Summary

Software technology for raster well logs has advanced to the point where a geologic workstation environment is now largely acessible to the average geologist.

Such an environment, currently in use for a range of projects in the U.S., permits rapid retrieval, display, and on-screen interpretation of logs, the building and modification of cross sections in a much-abbreviated time, and the exporting of all relevant data to existing mapping or engineering software programs.

This new computer technology supports the full spectrum of visual log analyses. It does so, moreover, in a far more convenient and time-saving format than is possible with either paper or microfiche logs. Documented productivity benefits are substantial and extend to many aspects of geological work.

It appears that the time has arrived to exchange the existing "paper mill" office for the first true version of a geologic workstation, based on raster logs.

Acknowledgments

Gratitude is expressed to the following companies and individuals who contributed data to the present study: Gerald Craig and Charles Bartberger, Amoco, Robert Johnson, Parker and Parsley; Bill Ross, Interpretive Imaging, Inc.

Reference

1. Cisco, S.L, Raster images offer low-cost well log preservation, OGJ, Dec. 16, 1996, pp. 40-43.

The Author

Scott L. Montgomery is a geologist, author, and translator residing in Seattle. He is the author of the quarterly monograph series Petroleum Frontiers, published by Petroleum Information Corp., and books and articles on the history of science. His current research interests include frontier plays and technologies and various topics in the history of geology and astronomy. He has a BA degree in English from Knox College and an MS in geological sciences from Cornell University.

Cuba

MacDonald Oil Exploration Ltd., Calgary, is processing seismic, aeromagnetic, gravity, and soil geochem data acquired on 9,900 sq km Block 22 on southern Cuba's west coast.

Target is oil in the Los Angeles sub-basin generally south and west of Vertientes.

The concession, covering parts of Camaguey, Las Tunas, and Granma pro- vinces, is Cuba's largest onshore hydrocarbon exploration block. It is west of and partly contiguous with a concession on which affiliate MacDonald Mines has been working an open pit gold deposit for about 3 years.

Egypt

Two Houston independents plan significant drilling in Egypt this year.

Apache Corp. plans to drill 62 wells, two thirds of them exploratory, and Seagull Energy Corp. will drill 40, 23 of them exploratory.

Apache, with interests in more than 27 million acres in Egypt, plans to spend $150-200 million in the country in 1997. Seagull holds interests in 9.3 million acres in Egypt and plans to spend $70 million this year.

Sudan

A group led by Arakis Energy Corp.'s State Petroleum Corp. unit plans a 3D seismic survey this year over the El Toor structure, where it found Cretaceous Bentiu and overlying Upper Aradeiba oil pay at an appraisal well (OGJ, June 10, 1996, p. 74).

The El Toor 50a-14-688 has a production capability of 4,500 b/d of oil on a 10% drawdown from Bentiu Unit I, the primary target. It also encountered 6.5 m of net pay in Upper Aradeiba sand and is evaluating the areal extent and productivity of this zone.

El Toor is 40 km southeast of Greater Heglig oil field (see map, OGJ, Feb. 24, 1997, p. 37).

The Toma South 88C well further confirmed the Toma South structure, discovered in 1996 by the 10,000 b/d-plus capacity Toma South 21d well. The 88C well, 6 km northwest of the discovery has a gross 73 m pay section with 45 m of net oil pay.

Two rigs are drilling, and two more are arriving, one of which has a 5,000 m depth capacity. Three more seismic crews are arriving.

Ukraine

Canadian companies are formulating a plan to further develop giant Dolina oil field in the western Ukraine Carpathians starting this summer.

Odyssey Petroleum Corp., Calgary, signed an option agreement with Trident (88) Exploration Ltd., Calgary, to acquire a 60% working interest in a proposed joint venture with the Ukrainian joint stock company Ukrnafta to develop the field, 520 km west of Kiev.

The field, with 264.5 million bbl cumulative production, yields 4,000 b/d which is pumped to the Drohobych refinery 48 km north. Dolina was discovered in 1950 and placed on full production in 1959.

Odyssey said the field had 1.234 billion bbl of original oil in place in four formations. It believes a further 235 million bbl can be recovered with western technology and equipment.

Most of the remaining oil will come from the upper zone at 1,560-3,000 m. Net pay zones at Dolina average 170 m thick. The companies project a production increase to about 40,000 b/d.

California

Tri-Valley Oil & Gas Co., Bakersfield, is completing its 1-15 Pimentel well as a 2.5 MMcfd gas discovery from two zones.

The well, in the city of Tracy, is the first commercial Upper Cretaceous Lathrop test on the high side of the Tracy thrust fault. Tri-Valley plans development wells and a 3D seismic survey to further define a deeper, larger objective.

Louisiana

A group led by Yuma Petroleum Co., Houston, is starting production from and plans development drilling near a discovery in Main Pass Block 2 off St. Bernard Parish.

The 1 State Lease 14564 well flowed 451 b/d of oil and 100 Mcfd of gas from a Miocene sand interval at 7,221-30 ft. TD is 9,030 ft. The well is on the Raccoon Island prospect, derived from 3D seismic data, about 3 miles northeast of Main Pass Block 11 field.

Working interest partners include Yuma, Shell Western E&P Inc., Enron Capital & Trade Resources, and Unexco.

Mississippi

Roundtree & Associates Inc., Jackson, spudded a 10,300 ft well near a Cretaceous Hosston discovery on the Moselle salt dome.

The 1 Gladdis Knight 32-14, in 32-7n-13w, Jones County, flowed 312 b/d of 37° gravity oil on a 24/64 in. choke from two Hosston reservoirs at 14,080-300 ft. It cut a combined 527 ft of net oil pay in Midway, Eutaw, Tuscaloosa, Paluxy, Rodessa, and Hoss- ton, reports Box Energy Corp., Dallas, with an 80% interest.

Texas

Tesoro paid $3.1 million for a 40% interest in an exploratory joint venture with Kaiser-Francis that includes 23 prospects and more than 47,000 gross acres under lease. This gives Tesoro about 11,000 net acres.

The areas are centered in Anderson, Cherokee, Henderson, Leon, Nacogdoches, Smith, and Van Zandt counties. Tesoro said it is eager to apply its geophysical expertise that has proved successful in the Gulf Coast region.

Utah

Harken Southwest Corp., Irving, Tex., and the Utah Geological Survey are drilling a horizontal well in Mule field in San Juan County in the Paradox basin.

The Mule 31 K-1 well, part of a 6 year, $5 million project funded by the U.S. Department of Energy, is to have a 1,000 ft horizontal leg in a reefal mound. Operators usually drill one to five vertical wells to develop such mounds in the area.

Wyoming

Louisiana Land & Exploration Co. hiked to 1 tcf its estimate of gross proved and probable reserves in the Madden field Mississippian Madison reservoir in the Wind River basin.

This is based on results of the 4-36 Big Horn extension well in Fremont County 1,000 ft downdip from the crest of a subsurface structure that has a minimum 23 sq mile areal extent. Logs indicated a 260 ft gross hydrocarbon column with no water level and excellent reservoir quality, the company said. The well was to be cased and tested during the next 6-8 weeks.

TD is 24,600 ft in Madison, and the logs are similar to those of two existing Madison wells.

LL&E's 1-5 and 2-3 Big Horn wells are producing gas from Madison to the Lost Cabin gas processing plant at its capacity of 52 MMcfd. Sales average 34 MMcfd after carbon dioxide and hydrogen sulfide removal. The plant has treated more than 28 bcf of gas from the two wells since completion in 1995.

Working interests in 4-36 Big Horn are LL&E 38%, W.A. Moncrief Jr. 36%, North Central Oil Corp. 14%, and BHP Petroleum (Americas) Inc. 12%.