1 Perforating: An Overview€¦ · less optimal historical methods in perforating is a hybrid intelligent system integrating a fuzzy expert-system and an artificial neural-network

Perforating Deep Wells ... Smartly!

Mazen Kanj and Jean-Claude Roegiers

Rock Mechanics InstituteThe University of Oklahoma

Abstract

Perforating is probably the most important of all completion functions incased holes. The process involves many human/equipment constraints andits effectiveness is influenced by numerous system environment factors.

A methodology capable of resolving the difficulties associated with theless optimal historical methods in perforating is a hybrid intelligent systemintegrating a fuzzy expert-system and an artificial neural-network. Both Ex-pert System and Neural Network technologies have developed to the pointthat the advantages of each can be combined into more powerful (Hybrid)systems. The goal of this paper is to demonstrate the practicality of makingavailable such a system. The system is aimed at creating an improved perfo-ration design/completion methodology, identifying and treating perforations'damage, and forecasting and projecting actual deliverability from the per-forations. As proposed, the system considers the effects of various environ-mental factors, existing/anticipated wellbore conditions, human/equipmentconstraints, and near-wellbore reservoir anisotropy, laminations, and naturalfractures on the perforating design and analysis processes. Perforation isthe last "hurdle" in the process of successfully producing a well. Optimizingits design/completion and solving its problems will, undoubtedly, result in asubstantial economic impact in terms of production cost and efficiency.

Transactions on Ecology and the Environment vol 30, © 1999 WIT Press, www.witpress.com, ISSN 1743-3541

46 Computer Methods in Water Resources IV

1 Perforating: An Overview

The main objective of perforating is usually to establish communication be-tween the wellbore and reservoir fluids. Adequate communication with thereservoir, as well as proper isolation between various zones, at pay depthis essential to evaluate and optimize the production of hydrocarbons. Theprocess is widely used and adopted, and affects the majority of wells drilledworldwide.

Before selecting a perforating technique, one must fully understand cur-rent and long term completion objectives as well as well operating conditions.The perforation selection process is indeed affected, among others, by existingand anticipated pressure environment in the well [1], need for and type of wellstimulation programs [2], need for and type of sand exclusion measures ([3] -[6]), purpose of the completion itself (production or injection), and whetherthis is an original completion or a workover [7]. In addition, numerous envi-ronmental factors (formation compressive strength ([8] - [10]), cement sheaththickness [10], casing grade strength [11], geothermal effects [12], etc.) shouldbe considered while designing and completing any perforating job. Althoughone doesn't have any control over these factors, one should be aware of theirinfluence on the perforating operation.

On top of being almost irreversible, the process is (accordingly) delicateand is (probably) the most important function in cased hole completions.Perforations represent the last portion of reservoir inflow and directly affectwell productivity ([13] - [15]). The determination of proper perforation typeand technique will undoubtedly, result in both cost savings and improved dol-lar revenues. This paper aims at highlighting the need for a smart system forwell perforating design, completion, and problem diagnosis and at describingthe many different processes involved in developing such an intelligent tool.

2 Special Perforating Jobs

The selection of the perforating technique is also influenced by the need andtype of the well stimulation process and/or sand exclusion method. Perfo-rations play a crucial role in achieving a successful fracturing/acidizing andsand control well treatment.

Perforation design for a well which will be hydraulically fractured is usu-ally controlled by the requirements to place the stimulation treatment [16].Key parameters are the number of perforating shots per foot, perforations'depth, and perforations' phasing and orientation. Fracture initiation delin-eates the communication path between the wellbore and fracture plane [17].The initiation process is dependent on the orientation of the perforationswith the far field stress and the local pore pressure builds up relative tothe injection pressure [18]. Nonplanar fracture geometries such as multiple


Computer Methods in Water Resources IV 47

strands, reoriented, and T-shaped (Fig.l [17]) have a negative impact onthe potential to successfully stimulate the pay-zone. Typically, the objectiveis to ensure the existence of a single fracture, with optimum width, thatpropagates perpendicular to the minimum horizontal stress. In addition, nouniform criteria exist within the industry for defining optimum perforationdensity and phasing. Operators are using a variety of techniques based ontheir experience and on some adopted rules-of-thumb.

• Reorientation0 Multiple fracture(away from wellbore)

• Single• T-chaped0 Multiple

VHmin• Multiple (at well bore)• Reorientation

Figure 1: Possible Fracture orientations from Horizontal Well. [17]

Perforating for sand control requires different guidelines and serves com-pletely different purposes than perforating for hydraulic fracturing or acidiz-ing. Where sand control is needed it can be assumed that there will notbe a permanent, open-perforation tunnel beyond the cement sheath. Theperforation channel behind the casing will most probably be filled with pack-ing material. Designing perforations for pre-packed gravel packing demandsshort cylindrical perforation openings (Fig.2 [3]). Perforations are the ba-sic problem in this case. Crushed formation particles and metallic particlesfrom the jet charge act to restrict the placement of packing material into theperforation tunnel. Everything must be aimed at obtaining large-diameterclean perforations. For gravel packing, large-diameter perforations are moreimportant than perforation length, provided that the perforations effectivelycommunicate with the formation (reservoir).



SCREENLINER ̂ ̂

GRAVEL PACK PERFORATION

FORMATION SAND

Figure 2: Gravel-packed perforation tunnel. [3]

According to the nature of the reservoir, the well may be either completedwith formation stimulation, such as hydraulic fracturing and/or acidizing, ornaturally with sand-exclusion and control. Each completion method requiresdifferent approach to perforating and sometimes even demands an unortho-dox/unconventional (heuristics-based) perforating procedural guidelines.

3 Knowledge-Based Perforating Constraints

It is well known that many human and numerous equipment constraints playmajor roles in resulting a less than optimal perforating job.

Some of the prevalent human causes for failed/unsatisfactory perforat-ing results probably are (1) a lack of understanding of what is required forsuccessfully perforating a well, (2) a rather widespread practice of awardingperforating jobs on the basis of price rather than job quality [12], (3) poorperforating design, (4) poor selection or poor application of a good selectionof perforating fluid [7], (5) a lack of awareness of the well environmentalfactors and of existing or anticipated wellbore conditions.

Because the perforation operation is equipment intensive, there are equip-ment constraints: (1) casing deformation or damage/fracture can result frominadequate use of perforating guns [19]; (2) plugging perfs with shaped-chargedebris [13]; (3) inadequate gun clearance [20]; (4) damaged primacord; (5)poor quality booster; (6) poor quality or aging main explosive [21]; and (7)



poor quality and improper perforating charge.In addition to problems associated with system geometry (shot density,

perforation dimensions, perforation phasing), impairements sometimes resultfrom the perforating process itself. In conventional jet perforating systems,holes are made through casing and cement and no material is removed. Theperforation hole volume is obtained by compaction of material around thehole. The crushed zone around the perforations essentially increases the ver-tical resistance to flow and, accordingly, adds to the vertical skin. Because ofthis damage, common perforating practices provide only 80% of productivityof an ideal undamaged perforation [5]. In addition, whenever a workover op-eration is needed, kill fluid is pumped into the well to control it. The invasionof filtrate from the kill into the perforations causes a deposition of filter cakeat the face of the perforation tunnel and reduces its permeability [22]. Ac-cordingly, it is very essential to identify the type/degree of occured damageand to use proper damage removal/treatment technique(s).

King and Buckley [21] conducted a study on cases of formation damageoccuring upon initial perforating completion and continuing past standardperforating breakdown attempts or small matrix acidizing jobs. A commondenominator in several of these cases appeared to be the performance ofperforating charges and its support system.

Well completions should be designed to minimize workovers and recomple-tion jobs. For optimum design of a completion one should choose the mosteffective perforating system, charges, guns, and shot density that will resultin a productivity ratio, PR, of near-one (production flow of the perforatedinterval equals open hole production of the same interval), even after 50% ofperforations have plugged off [23]. One should then select the most effectiveperforating pressure that the formation and/or well mechanical system willpermit in an effort to achieve damage-free, near-zero skin, completion. Inshort, perforating is a knowledge-intensive subject and a problem domainwhere expert heuristics needs to be considered in almost every step of thedesign and completion process.

4 Recognition/Prediction Problems

A large number of variables (wellbore pressure; formation strength; near-wellbore reservoir anisotropy, laminations, or natural fractures ([24], [25]);chemical incompatibilities between completion and formation fluids; perfora-tions geometry; geothermal effects; perforations plugging; casing and cementdamage; etc.) affect the perforation completion process. Therefore, perfo-rated completions seldom perform as expected. It is imperative that eachperforation job be evaluated to confirm its efficiency and to decide uponthe remedial actions needed. "Accurate" estimates of the productivity aregenerally required to predict well performance and to evaluate completion



success. Several studies ([14], [26] - [30]) have investigated the relative ef-fect of the parameters which define a perforated completion system on theproductivity of the well using experimental, analytical or numerical mod-els. However, current experimental setups do not (and cannot) simulateactual perforated-wellbore completion conditions; analytical treatments areextremely difficult (because of the spiral distribution of perforations in thewellbore axial direction and the presence of the wellbore itself) even withsimplifying assumptions; and numerical techniques, while they provide usefulinsight into predicting the productivity, are complex, expensive and mostlyproprietary. Probably, the best evaluation of any perforating completion isits performance during production. This performance is affected by manyfactors (human, operational, reservoir, etc.) that are not necessarily a partof the completion process. After a number of completions, however, trends inperformance can be established. This kind of problems accordingly suggeststhe appropriateness of the artificial neural network technology.

The importance of good depth control can never be over-emphasised.Remedial measures to correct problems caused by perforating off-target arecostly, time consuming, difficult to analyse, frustrating, and to the leastmisleading. The problem is misleading to the extent that many other reasonsare often suggested to explain well's poor performance caused by shootingoff-depth. The accepted method of ensuring accurate perforation depth isto run a cased-hole log (Casing Collar Locator, CCL) and correlate it withopen-hole measurements (Gamma Ray-Neutron, GR/N, logs). In additionto the correlation problem, corrections needs to be made on both (open-hole and cased-hole) measurements to adjust for vertical displacements oflogs measure points. If either measurement-correct ion is made in the wrongdirection, the resulting error will be twice the amount of the displacementbeing corrected.

Accordingly, while numerous heuristic-based constraints are involved inthe perforating process and suggesting the approprietness of this topic as anexpert-system application; there are some other perforating problems thatcan best be solved using the Artificial Neural Network (ANN) technology.Namely the following two problems: (1) depth control to avoid any off-targetperforating, and (2) perforating job evaluation and well performance predic-tion.

5 Summary and Conclusions

Knowledge-based systems perform reasoning using domain-specific rules, facts,heuristics and "tricks-of-the-trade" for a well defined and narrow problemdomain. They are especially useful for interacting with the user to definea specific problem and bring in facts peculiar to the problem being solved.Neural networks, with their remarkable ability to derive meaning from com-



plicated or imprecise data, are used to extract patterns and detect trendsthat are too complex to be noticed by standard computing techniques. Neu-ral computing can offer the advantage of execution speed once the networkhas been trained. Since Neural Networks are best at identifying trends andpatterns in data (pattern recognition), they are also well suited for predictionor forecasting needs.

Selecting the proper perforating technique requires full understanding ofthe pressure environment in the well, a decision on whether or not the wellwill be stimulated and how, whether or not a sanding treatment is needed andof what type, whether this is a new completion or a workover, and whetherthe well is to be used for production or injection. While shape and geometri-cal distribution of perforations, degree of perforation damage and perforatingefficiency strongly influence the final outcome; the selection of proper per-forating hardware is essential for the optimization of well productivity (orinjectivity). In short, the process relies heavily on heuristics and know-howand requires proper planning and execution.

It is generally recognized that the productivity analysis of perforated com-pletions, because of the complex 3D flow into a spiral system of perforations,is not easily amenable to analytical treatments. "Friendly" numerical meth-ods, on the other hand, are scarce, mostly proprietary, complex and to theleast time consuming. Moreover, the importance of good perforating depthcontrol can never be over-emphasized. Remedial measures to correct prob-lems caused by perforating off target are always costly and oftenly misleading.Accordingly, this domain involves also complex productivity-prediction andserious depth-recognition problems.

The proposed hybrid "intelligent" system is based on alternate and im-proved methodologies adopted from the AI (Artificial Intelligence) technol-ogy. These methodologies are: fuzzy expert systems and artificial neuralnetworks. Expert systems and artificial neural networks are well establishedthat in fact represent complimentary approaches. The logical, perceptive,heuristic, and mechanical nature of expert systems very well compliment thenumeric, associative, self-organizing, biological nature of neural networks.Fuzzy logic and expert systems, on the other hand, are natural combinationsbecause fuzzy concepts are naturally presented in the form of rules in fuzzyexpert systems.

The itemized list of the proposed-system objectives includes: (1) decisionon the completion type (cased-hole or open-hole) and accordingly, on theneed for perforating; (2) decision on perforation completion objectives (nat-ural, sand-control, or stimulation) (3) classification of applicable perforatingtechniques; (4) selection of appropriate perforating procedure(s); (5) designof the perforating job (number of shots per foot, perforations' depth, perfo-rations' phasing and orientation, etc.); (6) selection of effective perforatingsystem, charges, and guns; (7) selection/design of perforating completionfluid; (8) re-completion (of perforations) in workover wells; (9) identification



of type/degree of perforation damage and recommendation on proper dam-age removal/treatment technique(s); (10) sensitivity analysis and perforationeconomics; (11) recognition of correct perforating depth required to avoid off-target perforating; and (12) evaluation of the perforating job and predictionof well performance.

References

[1] Behrmann, L.A., and Halleck, P.M., Effects of Wellbore Pressure onPerforator Penetration Depth, SPE-18243, Proc. 63rd Annual TechnicalConference and Exhibition of the SPE, Houston, TX, Oct. 2-5, 1988,7pp. Also: SPE Reprint Series, No.31: Perforating, SPE, Richardson,TX, 1991, pp.98-104.

[2] Allen, T.O., and Roberts, A.P., Production Operations: Well Comple-tions, Workover, and Stimulation, v.2, Chapter 7: Acidizing, and Chap-ter ̂ '.Hydraulic Fracturing, Oil and Gas Consultants, Inc., Tulsa, OK,Third Edition, 1989, 350pp.

[3] Penberthy, W.L. Jr., and Shaughnessy, C.M., Sand Control, SPE Serieson Special Topics v.l, Henry L. Doherty Series, Richardson, TX, 1992,98pp.

[4] Allen, T.O., and Roberts, A.P., Production Operations: Well Comple-tions, Workover, and Stimulation, v.2, Chapter 4: Sand Control, Oiland Gas Consultants, Inc., Tulsa, OK, Third Edition, 1989, 350pp.

[5] Suman, G.O. Jr., Ellis, R.C., and Snyder, R.E., Sand Control Handbook,Gulf Publishing Co., Houston, TX, Second Edition, 1983, 89pp.

[6] Bonomo, J.M., and Young, W.S., Analysis and Evaluation of PerforatingCleanup Methods, JPT, March 1985, pp. 505-510.

[7] Patton, L.D., and Abbott, W.A., Well Completions and Workovers,Part 8 - The Systems Approach to Perforating, Petroleum Engineer In-ternational, April 1980, v.52, no.5, p.44 (6)

[8] Saucier, R.J., and Lands, J.F., A Laboratory Study of Perforations inStressed Formation Rocks, SPE-6758, JPT, Sept. 1978, pp.1347-1353.Also: SPE Reprint Series, No.31: Perforating, SPE, Richardson, TX,1991, pp.76-82.

[9] Halleck, P.M., Saucier, R.J., Behrmann, L.A., and Ahrens, T., Reductionof Jet Perforator Penetration in Rock Under Stress, SPE-18245, Proc.63rd Annual Technical Conference and Exhibition of the SPE, Houston,TX, Oct. 2-5, 1988, 15pp. Also: SPE Reprint Series, No.31: Perforating,SPE, Richardson, TX, 1991, pp.105-119.



[10] Behrmann, L.A., and Halleck, P.M., Effect of Concrete and BereaStrengths on Perforator Performance and Resulting Impact on the NewAPI RP-43, SPE-18242, Proc. 63rd Annual Technical Conference andExhibition of the SPE, Houston, TX, Oct. 2-5, 1988. Also: SPE ReprintSeries, No.31: Perforating, SPE, Richardson, TX, 1991, pp.83-97.

[11] Porter, W.L., and Satterwhite, B., Evaluation of Well Perforator Per-formance, JPT, Dec. 1976, pp.1466-1472.

[12] Allen, T.O., and Roberts, A.P., Production Operations: Well Comple-tions, Workover, and Stimulation, v.l, Chapter 7: Perforating Oil andGas Wells, Oil and Gas Consultants, Inc., Tulsa, OK, Third Edition,1989, pp.310.

[13] Suman, G.O. Jr., Perforations — A Prime Source of Well PerformanceProblems, JPT, Apr. 1972, pp.399-411. Also: SPE Reprint Series, No.31: Perforating, Richardson, TX, 1991, pp.165-177.

[14] McLeod, H.O., The Effects of Perforating Conditions on Well Perfor-mance, JPT, Jan. 1983, pp.31-39. Also: SPE Reprint Series, No.31:Perforating, SPE, Richardson, TX, 1991, pp. 132-140.

[15] Twyford, L.R., and Tyler, T.N., Impact of Perforation Variables on WellProductivity at Prudhoe Bay, SPE-12765, Proc. 1984 SPE CaliforniaRegional Meeting, Long Beach, CA, Apr. 11-13, 1984.

[16] Pearson, C.M., Bond, A.J., Eck, M.E., Schmidt, J.H., Results of Stress-Orientd and Aligned Perforating in Fracturing Deviated Wells, SPE-22836, Proc. 66th Annual Technical Conference and Exhibition of theSPE, Dallas, TX, Oct. 6-9, 1991, pp.195-202.

[17] Abass, H.H., Wilson, J.M., Venditto, J.J., and Voss, R.E., StimulatingWeak Formations Using New Hydraulic Fracturing and Sand ControlApproaches, SPE-25494, Proc. Production Operation Symposium, Ok-lahoma City, OK, March 21-23, 1993, pp.797-808.

[18] Behrmann, L.A., and Elbel, J.L., Effect of Perforations on FractureInitiation, SPE-20661, Proc. 65th Annual Conference and Exhibition ofthe SPE, New Orleans, LA, Sept. 23-25, 1990, pp.459-472.

[19] Godfrey, W.K., and Methven, N.E., Casing Damage Caused by Jet Per-forating, SPE-3043, Proc. SPE Annual Meeting, Houston, TX, Oct.4-7,1970.

[20] King, G.E., and Packard, B.R., The Cause of Blown Port Plugs, SPE-18988, Proc. SPE Joint Rocky Mountain Regional/Low PermeabilityReservoirs Symposium and Exhibition, Denver, CO, March 6-8, 1989,pp.577-584.



[21] King, G.E., and Buckley, P.D., Effect of Storage Conditions and Time ofStorage on Performance of Perforating Charges, SPE-21656, Proc. Pro-duction Operations Symposium, Oklahoma City, OK, April 7-9, 1991,pp.247-253.

[22] Han, L., Peden, J.M., Ford, J., Effects of Various Parameters on Perfo-ration Plugging and Perforation Cleanup, SPE-31090, Proc. FormationDamage Control Symposium, Lafayette, LA, Feb. 14-15, 1996, pp.185-205.

[23] Colle, E., Increase Production With Underbalanced Perforation,Petroleum Engineer International, July 1988, pp.39-42.

[24] Tariq, S.M., Ichara, M.J., and Ayestaran, L., Performance of PerforatedCompletions in the Presence of Anisotropy, Laminations, or NaturalFractures, SPEPE, Nov. 1989, pp.376-384. Also: SPE Reprint Series,No.31: Perforating, SPE, Richardson, TX, 1991, pp.60-68.

[25] Ichara, M.J., and Cannon, D.E., The Effect of Perforating on the Per-formance of Shale-Laminated and Crossbedded Formations, SPE-16205,Proc. Production Operations Symposium, Oklahoma City, OK, March8-10, 1987.

[26] Locke, S., An Advanced Method for Predicting the Productivity Ratio ofa Perforated Well, JPT, Dec. 1981, pp.2481-2488. Also: SPE ReprintSeries, No.31: Perforating, SPE, Richardson, TX, 1991, pp.124-131.

[27] Todd, B.J., and Bradely, D.J., Perforation Geometry and Skin Effectson Well Productivity, SPE-15029, Proc. Permian Basin Oil and Gas Re-covery Conference, Midland, TX, March 13-14, 1986. Also: SPE ReprintSeries, No.31: Perforating, SPE, Richardson, TX, 1991, pp.49-59.

[28] Tariq, S.M., Evaluation of Flow Characteristics of Perforations Includ-ing Nonlinear Effects with the Finite-Element Method, SPEPE, May1987, pp.104-112; Trans. AIME, v.284. Also: SPE Reprint Series, No.31:Perforating, SPE, Richardson, TX, 1991, pp.40-48.

[29] Perez, G., and Kelkar, B.C., A New Method To Predict Two-Phase Pres-sure Drop Across Perforations, SPE-18248, Proc. 63rd Annual TechnicalConference and Exhibition of the SPE, Houston, TX, Oct. 2-5, 1988,pp.659-672.

[30] Karakas, M., and Tariq, S.M., Semianalytical Productivity Models forPerforated Completions, SPEPE, Feb. 1991, pp.73-82. Also: SPEReprint Series, No.31: Perforating, SPE, Richardson, TX, 1991, pp.141-150.


Documents

1 Perforating: An Overview€¦ · less optimal historical methods in perforating is a hybrid intelligent system integrating a fuzzy expert-system and an artificial neural-network