SPE翻译(应用软件对油藏产出水进行计算分析的新方法).doc

SPE 108702New Insights and Applications of Soft Computing on Analysis of Water ProductionFrom Oil ReservoirsLeonid Sheremetov,Ana Cosultchi,Ildar Batyrshin,and Jorge Martnez-Mu?oz,Instituto Mexicano del Petrleo,and Sergio Berumen,Pemex Exploracin y ProduccinCopyright 2007,Society of Petroleum EngineersThis paper was prepared for presentation at the 2007 International Oil Conference andExhibition in Mexico held in Veracruz,Mexico,2730 June 2007.This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s).Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s).The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE,P.O. Box 833836,Richardson,Texas 75083-3836 U.S.A., fax 01-972-952-9435.AbstractDuring the development of reserves of oil fields high water production is not uncommon problem.However, the practice of Oil & Gas production has witnessed many cases with poor field results and low efficiency of operational efforts to mitigate the water production problem. One of the most common reasons is a lack of a precise diagnosis to identify the problem that considers the analysis of well and reservoir integrated information. That is why the key factor in the success of a water control treatment relies in a reliable and precise diagnostics identifying the origin of the water production problem. This leads to select an adequate treatment to effectively mitigate the problem.This paper describes a new methodology of water shut-off (WSO).This methodology is implemented through a new generation computerized tool, SMART-Agua, which uses soft computing techniques(fuzzy expert systems and perceptual data mining).The system was designed to manage the diagnosis of 14 potential problems commonly found both at wells and reservoirs. The solution analysis phase offers a technical and an economic analysis of the different chemical or mechanical applications and evaluates the potential for additional productivity if the water cut is reduced. Additional modules for solution design (gels, mechanical solutions and surface separation facilities) generate the parameters of the most efficient solutions instead of selecting a particular product from the vendor. Applications of this new methodology to several oil field cases were used to validate the methodology, and results of two such cases are discussed in the paper. These cases are focused on a carbonate fractured oil reservoir and a carbonate reservoir under waterflooding.IntroductionThe growing interest and enthusiasm in the petroleum industry for the intelligent engineering systems like intelligent wells and intelligent fields have increased significantly the demand for more powerful, robust and intelligent tools. The interest in decision support for the operational activities lays in the fact that in petroleum companies, senior personnel daily have to solve problems based on extensive data analysis and their experience gained through years of field work. In recent years, hybrid intelligent systems(HIS)integrating different artificial intelligence techniques from expert systems to fuzzy logic have attracted the attention of the Oil & Gas industry, by their capacity to handle successfully the complexities of the real world like imprecision, uncertainty and vagueness1-4. In spite of the fact that the petroleum industry is usually being criticized for moving at a snails pace in embracing information technology, it was one of the first in applying artificial intelligence technologies. Expert systems like DIPMETER ADVISOR and PROSPECTOR are often mentioned as examples of early classic systems. Nevertheless, an analysis of commercially available intelligent software, indicates that although there are some applications,this type of systems have still not been put widely in the market5.For example in the area of water control,two expert systems exist from service companies, XERO and WaterCase developed by Halliburton and Schlumberger respectively to provide assistance during the diagnosis and treatment selection phases of water control6,7.But these systems are primarily of internal use for choosing a product from that vendor,and are not open for scrutiny or modification of the rule bases. This work describes a hybrid intelligent system that uses fuzzy logic, the techniques of the expert systems and of data mining to support the petroleum engineers in the diagnosis of the problems of water shutoff, one of the challenging problems of oil production. In many cases, the innovating WSO technologies can mean a reduction of the costs and an increase in the hydrocarbon production. In order to properly diagnose a problem and find most suitable and efficient solution, one requires the intervention and the experience of the petroleum engineers. SMART-Agua (Agua stands for Water in Spanish)makes use of these experiences acquired for naturally fractured reservoirs(typical in Mexican scenes) integrated within a hybrid intelligent tool.Water production problemWater production represents an imperative and complex problem affecting the lifespan of most hydrocarbon wells 8,9and, consequently, any potential solution and its success depends on the opportune and reliable diagnosis of the problem. If one considers that the world-wide water production is of approximately 210 million barrels per day which accompanies to the 75 million barrels per day of petroleum, it would be possible to say that many companies have practically become water producers9,10.Since the systems of handling of water are expensive-the cost is considered between 5 to more than 50 cents per barrel of water -in a well that produces petroleum with an 80%of water cut, the cost of the handling of the water can ascend to$4 per produced petroleum barrel.Reservoir rocks normally contain both petroleum hydrocarbons and water but, sources of water may include flow from above or below the hydrocarbon zone, flow from within the hydrocarbon zone, or flow from injected fluids and additives resulting from production activities. This water becomes produced water when hydrocarbons are produced and these fluids are brought to the surface as a mixture. Nonetheless, water-driving mechanism is the beneficial aspect of an aquifer.The physical and chemical properties of produced water vary considerably depending on the field location, the geological formation, and the type of hydrocarbon being produced11.If water flooding operations are conducted, these properties and volumes may vary even more dramatically as additional water is injected into the formation. Water production may also provoke additional problems such as: a)corrosion of tubing string, casing or their connections; b)obstruction of the tubing by scale; c)increase of hydrostatic pressure, etc. For example, it is challenging to minimize the amount of produced water by using blocking devices or chemicals, but the strategies depend upon the problem location, type of hydrocarbon production (natural or waterflooding) type of formation(i.e. anisotropy),the results of previous treatments, etc. This requires synthesizing the information from different data sources like geological model of the field, perforation reports, well and production logs, fluids analysis, records of maintenance of oilfield facilities, etc.Solution descriptionThe aim of SMART-Agua system is to offer technical support to field operators in order to diminish the time spent for diagnosis and to increase the success of the selected solution. According to the water control process diagram(Fig.1),in order to diagnose the problem, the data comprising several groups such as: borehole features, production histories, production logs, reservoir characterization, water analysis and past-treatments data are required. At the diagnostics phase, there are two stages: a preliminary one and a confirmation step. The underlying methodology for both of those was constructed over the experiences of experts in the problems of water control. The discrimination between the different water production problems (the preliminary diagnostics) is based on 93 properties and data collected from the field. For a particular case, usually there is no unique diagnosis but few problems that can be ranked by score according to their likelihood of incidence. Hence there lies the need for confirmation of diagnosis. Usually, at this second stage, it is required to apply additional tests, run updated logs and analyze current produced water. All these activities provide the system with updated information which will confirm(or discard)problems found during the preliminary diagnosis.The stage of solution analysis and selection also consists of several steps. The first one consists of evaluating the recommended treatments for the selected problem. Treatment selection is based upon treatment objectives; these are classified in three groups: those that minimize the amount of produced water that reaches the surface, those that recycle or re-inject produced water, and those that involve disposal of produced water. Application of the adequate solution represents challenges and costs which may increase with a wrong diagnosis. The evaluation of solutions is performed and adjusted to: the available facilities, characteristics of the formation, results of water analysis and the objective of the treatment as stated above.Upon deciding what solution to apply, it is required a detailed design as the next step in the process. This task includes the parametric design of the most efficient solution along with a review of available products and equipment for water control and management that meet the requirements of the suggested solution. Since the justification of a treatment in any well depends upon the expected increase in the hydrocarbon production, it is needed to quantify the cost of the treatment of water control by means of an analysis of risk components using the methods of quantitative risk analysis. This analysis is made in the phase of economic evaluation of the proposed technical solution.Additionally, the results from each treated case, either success or failure, are processed in order to feed two databases: the knowledge base and a historic record of cases for future reuse.Statement of Theory and DefinitionsIn order to develop a straightforward strategy and methodology for water excess diagnosis and solution analysis (described in more details in this section), experts from different problem domains were consulted. These domains included in particular: reservoir, formation and petrophysical models, NODAL analysis, pressure transient test, logging tools, conning simulation, gels and polymers, economic evaluation of separation facilities and equipment, risk analysis, etc. Experts experiences along with other additional information obtained from available bibliographical sources were structured in order to relate the problems with their symptoms as explained below. Further on, the relationships were coded in the form of fuzzy production rules thus forming three related knowledge bases (preliminary, confirmation and solutions), one for each stage of the diagnosis and solution analysis process.Classification of problems of excessive water productionMost excess-water-production problems are classified using different criteria depending on the authors interest 9,12-19. That is why,although the water production problems are widely described in the literature9,we briefly describe them emphasizing the considerations used in this application.For classification purposes there are used:the water entrance (linear or radial),the water control treatment strategy,or the type of formation(fractured or not).SMART-Agua uses location of the problem as a primary classification parameter and includes 14 problems.Each of the these problems can be classified as(Table 1):a)close to the well bottom,b)within the formation,c)between injector and producers,d)at the injector and,e)at the well completion.The problems are defined as follows.Moving oil-water contact.This problem occurs in wells producing from homogeneous formation with associated aquifers,which is a normal situation for all wells where water is co-produced with oil and the pressure depletion promotes a slow entrance of bottom water.This problem can be recognized if the well produces below the critical flow rate. Coning or cresting.This problem is identified if the well produces above critical flow rate from homogeneous formation with associated aquifers.Coning occurs in vertical producers and depends on the well pressure depletion,while cresting occurs in horizontal wells.Water channeling.This case occurs when the formation has permeability contrasts and impermeable barriers,which allows water to mobilize towards the producer by flowing though high permeability strata(channels).Water channeling through fractures.This problem occurs in naturally fractured formations which allow water phase entrance.This problem can occur also for waterflooding.In such case,it is favored by fracture nets crossing both,injector and producer,but if such nets are not directly crossing the well,then water entrance may occur in the future.Barrier breakdown.Fracturing of naturally impermeable barriers allows water entrance.Fracturing may be the consequence of fracturing workovers or pressure excess during production or completion.Stimulation out-of-zone.Any stimulation or cleaning workover may contribute to open channels or fractures into a fractured formation.Poor areal sweep.When waterflooding is used in anisotropic formation containing high permeability strata (channels)water starts flowing preferentially through these channels.Channel through high permeability strata.A similar problem,but the formation may not necessarily be anisotropic. The presence of impermeable barriers,however,blocks crossing flow between strata.Channel trough high permeability strata with crossflow. When the impermeable barriers are break down,crossing flow between strata occurs.Flow behind casing.A poor adherence of the cement to the formation side means channels,caverns or fractures,which allow water to accumulate behind casing.Usually occurs when casing was poorly cemented or the cement job was damaged during production or repairing jobs and especially,when the formation contains sand or clay.Casing leaks.This problem occurs as a consequence of corrosion or pipe misalignment when the casing has small or large aperture breaches.Casing corrosion in injector well.The problem is similar to casing leaks,although,in this case,injected water may displace behind casing instead of the formation front. Gravity segregation in injector well.This problem happens in injector wells due to the presence of fractures.Thus, injected water is gravitationally displaced to the well bottom instead of the formation front.Problems classified as close to the well bottom are mostly associated to naturally or artificial lift producer wells:Moving oil-water contactConing or crestingWater channelingWater channeling through fracturesProblems located within the formation are related toprevious workover with negative results on the oil production:Barrier breakdownStimulation out-of-zoneProblems located between injector and producer wells aremore common during waterflooding:Poor areal sweepWater channeling through fracturesChannel through high permeability strataChannel trough high permeability strata with crossflowProblems located at the Injector well are typical for thesewells:Gravity segregation in injector wellCasing corrosion in injector wellProblems located at the well completion are common to allproducers:Casing leaksFlow behind casingTherefore,when the analyzed well is a producer,there are 12 problems to be evaluated,although they are different when the well produces naturally or through artificial lift than when it produces by waterflooding.When the well is an injector, only two problems can be present.Methodology of preliminary diagnosisDue to the space limits,the preliminary diagnosis phase is discussed in more details than the other phases.Different facts (also called resources of the problem domain)intervene in the elaboration of the preliminary diagnosis as shown in Table 1.In SMART-Agua,formation characteristics are obtained from input data like:mineral composition,permeability, permeability distribution model(e.g.Dykstra-Parson)and interpretations of logs(e.g.sonic,gamma ray).A homogeneous formation is related to two problems;in comparison,a heterogeneous formation is present in six problems as follows:a channel-type formation defines the problem Water channeling,an anisotropic formation is related to Gravity segregation and Casing corrosion in injector well, a fractured formation is related to Water channeling through fractures,Stimulation out-of-zone and Gravity segregation. Anisotropy and the presence of fractures are involved in the diagnostics of:Stimulation out-of-zone and Gravity s