SPE 84886 Characterizing Partially Fractured Reservoirs by Tracer Injection

Fuad Qasem, SPE, Ridha B.C. Gharbi, SPE, and Muhammed I. Mir, Kuwait University
Copyright 2003, Society of Petroleum Engineers Inc. This paper was prepared for presentation at the SPE International Improved Oil Recovery Conference in Asia Pacific held in Kuala Lumpur, Malaysia, 2021 October 2003. 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, TX 75083-3836 U.S.A., fax 01-972-952-9435.

Saidi 1987). The difference arises from the two interacting paths (rock matrix and fractures) for fluid flow having totally different properties and communication with each other. Due to the complexity of NFR systems, the literature discussing characterization and modeling of NFR is scarce compared to non-NFR systems. Modeling Naturally Fractured Reservoirs. Research on fractured reservoir simulation has a long history. Performance prediction of naturally fractured reservoirs under uniform rock properties has been the subject of many publications during the last three decades. Currently, there are three different methods being used for simulation of NFR systems: (1) continuum approach, (2) discrete fracture approach, and (3) integrated approach. Conventional single continuum approach has been applied to simulate fractured reservoirs with fractures having small effects on the flow. Barenblatt et al (1960) first introduced the dual continuum (dual-porosity) approach. The authors considered naturally fractured reservoirs as two homogeneous, isotropic, overlapping continuums: the matrix blocks and fracture network. They assumed quasi-steady state flow from matrix to fracture. Warren and Root (1963) later extended this work. The authors proposed a simplified representation of fracture networks to be used in dual-porosity simulators. They assumed that secondary porosity is contained within an orthogonal set of equally spaced system of fracture networks, while the matrix blocks feed the fracture continuously. In addition, a no-flow condition between the matrix blocks was assumed. In order to characterize the relationship between the two porosity regions, Warren and Root (1963) introduced two dimensionless parameters: inter-porosity flow coefficient ( ), and dimensionless fracture storage (). In his study, Kazemi (1976) used the dual continuum approach to model NFR by a multi-layer system (dual-permeability model). The author assumed that fractures are thin layers of high conductivity, alternating with thicker layers of matrix blocks consisting of high storage capacity but low conductivity. The dualpermeability simulators allow matrix-matrix flow (Kazemi et al. 1976, Rossen 1977, Thomas et al. 1983). In a later study, Pruess and Narasimhan (1985) introduced the multiple interacting continua method (MINC), which is applicable to numerical simulation of heat and multi-phase fluid flow in multi-dimensional, fractured porous media. One of the disadvantages of the dual-continuum approach is that it assumes fractures to be distributed regularly and well connected. Whereas, real fracture systems are very irregular and very poorly connected (Chiles 1987, Laubach 1991, Lorenz et al. 1991).

Abstract There is considerable interest in the petroleum industry to characterize partially fractured reservoirs and to develop an increased understanding of the physics of fluid flow in these types of reservoirs. This is because fractured reservoirs have different behavior and there exist a large number of these reservoirs that are not fully developed. This paper presents a numerical simulation study that was performed to investigate the effect of rock properties on the tracer response in partially fractured reservoirs using a finite difference numerical simulator. These properties include fracture intensity, fracture porosity and matrix permeability. The functional relationships between these parameters and the calculated effective permeabilities are also investigated. Several images, each with different probability of fracture intensity, were generated randomly. Numerical simulations of single-phase tracer transport were then performed in each of the generated fractured models. Results show that the fracture intensity, fracture porosity and matrix permeability have a significant effect on the tracer response in naturally fractured reservoirs. Depending on the reservoir properties, the results also show that the flow in partially fractured reservoirs can be either matrix-dominated or fracture-dominated. The characteristics of each regime and the conditions for its occurrence are presented. Introduction There is a large number of oil and gas reservoirs that are naturally fractured. In fact, one may claim that all hydrocarbon reservoirs are naturally fractured reservoirs (NFR) to a certain degree. The question remains, however, whether or not these fractures form a fracture network that affects the fluid flow properties. The effect of fractures becomes important only when they occur with sufficient length of penetration, connectivity and spacing. The behavior of these types of reservoirs is considerably different than the conventional reservoirs (Aguilar 1980, Van Golf Racht 1982,

SPE 84886

Long et al. (1985) and Dershowitz (1988) presented an approach called discrete fracture flow models. In this approach, natural fractures are assumed as a system of interacting fracture segments instead of viewing them as a continuum. Because this approach considers flow only through the fracture space, the contribution of matrix permeability and isolated fractures is not accounted for. Besides this limitation, the models based upon this approach are limited in size, speed and level of sophistication required to simulate multiphase flow when compared to models based upon continuum approach. This approach is typically applicable for computation of flow through a single region containing a large number of connected fractures. To combine the advantages of both continuum approach and the discrete fracture network approach, integrated methods have been introduced (Oda 1985, , Lee et al. 1997, Lough et al. 1998, Jensen et al. 1998, Park et al. 2000, Dershowitz 2000, Lee et al. 2000, Sutopo et al. 2001). In this approach, a discrete fracture network model of the reservoir is first prepared. Then, either this model directly or the parameter distribution derived from this model is used to provide input grid parameters for simulators based upon continuum (single or dual) approach. The approach retains many of the advantages of continuum approach along with the realism offered by the discrete fracture network approach. Characterization of Naturally Fractured Reservoirs. Various sources for fracture data have been used: outcrop studies, seismic, well logging, pressure transient tests, and inter-well tracer studies. The data from all these sources is integrated to get a reliable description of fracture system both at field scale and at local reservoir cell scale. Reliable characterization of fractures are now possible by developing tools for merging seismic, borehole imaging, lithological and outcrop data together with the help of geological and geochemical rules. Accurate seismic data yield reliable models of large-scale fracture networks, whereas borehole imaging provides the actual fracture description along the wells, which enables a reliable statistical determination of fractures. Well logging data represent only properties measured at or near the well bore, so its application to characterize the fractures system in the reservoir must be done with caution. Pressure transient analyses have often been used to estimate the equivalent fracture permeability, fracture volume, and sometimes the shape factor of the fracture network around the well. Pressure interference tests can also indicate the global horizontal anisotropy of fracture permeability (i.e. fracture orientation). Radioactive and chemical tracers have been used for many years in ground water hydrology to analyze movement of water through porous formations. However, their use in geothermal and petroleum reservoir engineering is relatively recent (Jensen 1983). The literature on the flow of tracer in porous media can be divided into two main categories: direct and inverse methods. The direct method deals with tracer response behavior from the knowledge of pertinent reservoir and tracer parameters. The inverse method estimates the reservoir and tracer parameters from the interpretation of the tracer response.

Inter-well tracer studies provide valuable characterization of naturally fractured reservoirs. The applications of tracers to study naturally fractured reservoirs have been the subject of numerous studies (Wagner 1977, Tester et al. 1982, Ramirez 1993, Shinta et al. 1993, Daltaban et al. 1994, Ramirez et al. 1994, Sato et al. 1994, Zellou et al. 1995, Maroongoog et al. 1995, Deng et al. 1995, Wattenbarger et al. 1995, Almeida et al. 1996). Despite all these studies (only few cited here), the response of tracers in partially fractured reservoirs is yet to be investigated. In partially fractured reservoirs, the fracture network does not cover the entire reservoir volume. In other words, the fragment sizes are larger than the simulation grid block. Therefore, the primary objective of this work is to use numerical simulations to investigate the effect of fracture intensities on the tracer response in partially fractured reservoirs. A secondary objective of the work is to examine how fracture porosity and matrix permeability affect the tracer response in these types of reservoirs. A final objective is to study the functional relationship between these parameters and the calculated effective permeabilities. First, we present the methodology that is used for this study. Methodology Fracture Models. The conventional methods of simulating naturally fractured reservoirs consist of fully regular matrix network, surrounded by interconnected fractures. This approach is unsuitable for partially fractured reservoirs because of their high degrees of heterogeneities. In partially fractured reservoirs, fracture distribution is highly irregular. In this study, a random distribution of fractures is considered. A uniform random number generator was used to generate random points inside a two-dimensional field of 70x71 grid (Press et al. 1992) Several images, each with different probability of fracture intensity, were generated using a single realization. Fracture intensity is defined as the ratio of the number of grid blocks having fractures to the total number of grid blocks. The fracture models were generated to cover a wide range of fracture intensities from 0.1 to 0.9. A fracture intensity of 0.1 represents a nearly non-fractured medium while a fracture intensity of 0.9 represents a highly fractured medium (i.e. 90% of all grid blocks are fractured). Depending on the depositional environment, partially fractured petroleum reservoirs can have widely varied fracture intensities ranging from high to low numbers. Figures 1 and 2 show 2-D permeability maps with fracture intensities of 0.10 and 0.5, respectively. The white colored blocks represent the nonfractured ones, while the dark color blocks are fractured. Fluid Flow Models. Numerical simulations of single-phase tracer transport were performed in each of the generated 2-D models. One well was placed horizontally along one side of the reservoir, while the other was placed along the opposite side. An advanced black oil simulator IMEX (Users Guide, 2000), in dual permeability mode, was used for this purpose. We have used Gilman and Kazemi ( 1983) formulation for shape factor calculations. The 2-D areal model with x-y-z grid of 70x71x1 was found to be relatively insensitive to further mesh refinement. For all simulation runs, the x- and ypermeability values were assumed equal in each grid-block. Fracture permeability was kept constant at 1000 md. Fracturing rarely increases the porosity more than a few

SPE 84886

percent but may dramatically increase permeabilities to values of several darcies. The total porosity of each grid block is assumed constant. The relationship between total porosity (t), fracture porosity (f), and matrix porosity (m) is as follows:

t = f + (1 f ) m

The water is injected continuously and across the entire inlet end through the 70 grid blocks. Injection was at a constant rate of 100 bbl/day. Tracer injection schedule was 1 lbs/bbl of water injected for 10 days. Production, constrained by constant outlet pressure, occurred through the 70 grid blocks at the outlet end of the porous medium. Table 1 lists the simulation input data. A sensitivity study was carried out to investigate the effect of fracture intensity, fracture porosity and matrix permeability on the normalized properties of tracer produced. These include the normalized cumulative tracer produced and the normalized rate of tracer produced. The fracture porosity and the matrix permeability were varied from 0.01 to 0.06, and from 0.5 to 35 md, respectively. As mentioned earlier, the fracture intensity varied from 0.1 to 0.9. The total amount of tracer injected (1000 lbs) is used to normalize the cumulative tracer produced, while the tracer mass injection of 100 lbs/day was used to normalize the rate of tracer produced. Results and Discussion Figures 3a and 3b show respectively the effect of fracture intensity ( FI ) on the normalized cumulative tracer produced ( m D ) and on the normalized rate of tracer produced ( q D ), both of which are in dimensionless form. In these simulation runs, the fracture porosity ( f ) was assumed 0.01 and matrix permeability ( k m ) was 5 md. The figures show the effect of

FI on the tracer response is very significant. As FI decreases below a certain value ( FI 0.6 ), the response curves of Figure 3a gradually become elongated or more dispersive. This corresponds to a gradual flattening of the response curve in Figure 3b and a shift in the peak to a later time. In these instances, the matrix increasingly dominates the flow and the fractures play less important role as FI decreases. On the other hand, as FI increases to higher values ( FI 0.6 ), the shape of the response curves gradually becomes less stretched (Figure 3a). This corresponds to sharper and shortly delayed peaks for higher values of FI (Figure 3b). In these cases, the fractures become more and more dominant and the matrix play lesser role as FI increases. These results appear to imply that there is critical value of FI that sort matrix-dominated flow from fractureddominated flow. For the case of k m = 5 md , this value of FI is on order of 0.6. Therefore, depending on the fracture intensity, the flow in partially fractured reservoirs can be either matrix-dominated or fractured-dominated. Figures 4a-b and 5a-b show the results of the tracer response when the fracture porosity was modified to 0.01 and 0.06, respectively. All other parameters in these runs were

maintained constant. These figures can be compared to the base case run of Figure 3a-b, where the fracture porosity was 0.02. It can be noticed from these figures that an increase in fracture porosity delays the tracer break through slightly. For reservoirs with high values of fracture intensity, this delay becomes more significant. Another way of looking at these data is shown in Figure 6. The figure shows the results of tracer response variation with the fracture porosity for a given fracture intensity. As shown, the effect of fracture porosity is more significant for fracture intensity FI 0.6 and less significant for FI 0.6 . At a fracture intensity of 0.6, the response shows the formation of two peaks. In this instance, both the matrix and the fracture play a role. The earlier peak denotes the flow from the fracture, while the second peak is due to the flow from the matrix. Therefore, this confirms that fracture intensity in the order of 0.6 represents a transition from fracture-dominated to matrix-dominated flow for the case of k m = 5 md . Figure 7 shows the change in the peak arrival time at the production well with the fracture intensity, FI for fracture porosity of 0.02 and 0.06. As shown, for low fracture intensity (matrix-dominated flow), the peak arrival time is a decreasing function of fracture intensity. On the other hand, for fractured-dominated flow, the peak arrival time increases with fracture intensity. The point where the two curves meet indicates the transition from matrix-dominated to fractureddominated flow. This transition, however, is a function of fracture porosity. As the fracture porosity increases from 0.02 to 0.06, the transition from matrix-dominated to fracturedominated flow occurred at higher value of fracture intensity. In addition, the rate of change of the peak arrival time in the fracture-dominated region was higher for higher fracture porosity. Because of the low difference between the two values, the variation of fracture porosity from 0.02 to 0.01 did not make a significant effect on the peak arrival time. Figures 8a-b, 9a-b, and 10a-b show the results of the tracer response when the matrix permeability was modified to 0.5, 2, and 35 md, respectively, with fracture porosity maintained at 0.02. These figures should also be compared to Figure 3a-b, where the matrix permeability was 5 md. Except the matrix permeability, all other parameters for these runs were maintained constant. An increase in matrix permeability values delays the break through time. This effect is more pronounced in cases in which a variation in the matrix permeability changes the nature of the flow (i.e. matrixdominated or fracture-dominated). In other words, the matrix permeability plays a dominant role in deciding whether the flow is matrix-dominated or fracture-dominated. This point is much more clear to see in Figure 11, which shows the effect of matrix permeability on the tracer response for a given fracture intensity. As shown, for fracture intensity of 0.4, the matrix dominates the flow for all studied values of matrix permeability (0.5, 2, and 35 md). However, for fracture intensity of 0.9, the fractures dominate the flow for all cases. For reservoirs with fracture intensity of 0.6, the effect of varying the matrix permeability changes the flow to fracture-dominated for k m = 0.5 md , and to matrix-

SPE 84886

dominated for k m = 35 md . Therefore, higher matrix permeability shifts the transition from matrix-dominated to fracture-dominated flow to higher values of fracture intensity. Figure 12 shows the change in the peak arrival time at the production well with the fracture intensity for matrix permeabilities of 5 and 35 md. As shown, for the case of 5 md, the matrix dominates the flow up to a fracture intensity of 0.6. After this point, the fractures will take over and consequently; the contribution from the matrix becomes insignificant. For the case of 35 md, however, the matrix still dominates the flow for higher values of fracture intensity. This shows that an increase in matrix permeability has shifted the transition to higher value of FI . Further increase in the matrix permeability will yield matrix-dominated flow no matter what the value of the fracture intensity is. Data also show that a decrease in the matrix permeability below 5 md will shift the transition of matrix-dominated to fracturedominated flow to lower values of FI . Therefore, the transition is a function of fracture intensity, matrix permeability and fracture porosity of the reservoir. The effect of the reservoirs properties on the effective permeability ( k eff ) is also investigated. The effective permeability values for the various models having different fracture intensity, fracture porosity, and matrix permeability are presented in Figure 13. As shown, the effect of fracture intensity on k eff is quite significant. With an increase in fracture intensity, the portion of high permeability medium (fractures) is increased, and therefore the effective permeability is increased. For matrix-dominated flow, the rate of increase in k eff is low, whereas, for fracture-dominated flow, the rate is much higher. For the same fracture intensity, the effect of matrix permeability is to shift the effective permeabilities to higher values. This shift is more pronounced in cases where both the matrix and the fractures play a role (i.e. at or near the transition zone). It should also be noted that the change in fracture porosity plays no significant effect on the effective permeability values. Conclusion This study was aimed at investigating the effect of rock properties on the tracer response in partially fractured reservoirs using a finite difference numerical simulator. Properties included fracture intensity, fracture porosity and matrix permeability. Based on the results of reservoir simulations of single-phase tracer response, we conclude the following: 1. Fracture intensity, fracture porosity and matrix permeability have a significant effect on the tracer response in partially fractured reservoirs. 2. Depending on the reservoir properties, the flow in partially fractured reservoirs can be either matrixdominated or fracture-dominated. The formation of two peaks in tracer response (one for fracture and one for matrix) indicates the transition from matrix-dominated to fracture -dominated flow. 3. The transition from matrix-dominated to fracturedominated flow is a function of fracture intensity ( FI ),

the matrix permeability ( k m ), and the fracture porosity ( f ). 4. Higher matrix permeability and higher fracture porosity shift the transition from matrix-dominated to fracturedominated flow to higher value of fracture intensity and vice versa. The effect of fracture porosity is more significant for fractured-dominated reservoirs and less significant for matrix-dominated reservoirs.

5.

Acknowledgments The authors express their appreciation to Kuwait University Research Administration for financially supporting this work through a university research grant (EP 02/01). Nomenclature FI = Fracture Intensity k eff = Effective permeability

km mD qD
NFR

= = = =

Matrix permeability Normalized cumulative tracer produced Normalized rate of tracer produced Naturally fractured reservoirs

Greek Symbols t = Total porosity of a grid block,

f m

= =

Fracture porosity Matrix porosity

References Aguilar, R.: Naturally Fractured reservoirs, PennWell, Tulsa, Oklahoma, 1980. Almeida, A.R. and Cotta, R.M.: Analytical Solution of the Tracer Equation for the Homogenous Five-Spot Problem, Soc. Pet. Eng. J. (March) 31-38, 1996. Barenblatt, G. I., Zheltov, I.P., and Kochina, I.N.: Basic Concepts in the theory of Seepage of Homogeneous Liquids in Fissured Rocks, Priklandaia Matematica I Mechanica Academia Nauk, S.S.S.R., 24, No.5, 852-864, 1960. Barenblatt, G.I., Zheltov, I.P.: On the Basic Equations of the Single Phase Flow of Fluids Through Fractured Porous Media, Dokladi Akademii Nauk, S.S.S.R., 132, No.3, 542-548, 1960. Chiles, J.P.: ThreeDimensional Geometric Modeling of a Fracture Network, proceedings of the conference on Geostatistical, Sensitivity and Uncertainty Methods for GroundWater Flow and Radionuclide Transport Modeling, San Francisco, California, September 1987, editor Bruce E. Buxton, Batelle Press, (1987) 361. Deng, X. and Horn, R.N.: Description of Heterogeneous Reservoir Using Tracer and Pressure Data Simultaneously, paper SPE 30590 presented at the SPE Annual Technical Conference and Exhibition, Dallas, TX, October 22-25, 1995.

SPE 84886

Dershowitz, B., LaPointe, P., Eiben, T., and Wei, L.: Integration of Discrete Feature Network Methods with Conventional Simulator Approaches, Soc. Pet. Eng. Reservoir Eval. & Eng. 3 (2), April 2000. Dershowitz, W. S., Doe, T. W.: Practical Applications of Discrete fracture Approaches in Hydrology, Mining and Petroleum Extraction, proceedings of the International Conference on Fluid Flow in Fractured Rocks, Atlanta, Georgia, May 1988, 381. Gilman, J., and Kazemi, H.: Improvements in Simulation of Naturally Fractured Reservoirs, Soc. Pet. Eng. J., August 1983, pp. 695-707. Jensen, C.L., Lee, S.H., Miliken, W.J., Kamath, J., Narr, W., Wu, H., and Davies, J.P.: Field Simulation of Naturally Fractured Reservoirs Using Permeabilities Derived From Realistic Fracture Characterization, paper SPE 48999, presented at the SPE Annual Technical Conference and Exhibition held in New Orleans, LA, Sept. 27-30, 1998. Jensen, C.L.: Matrix Diffusion and Its Effect on the Modeling of Tracer Returns From the Fractured geothermal Reservoir at Wairakei New Zealand, Stanford Geothermal Program, SGP-TR-TR-71, Stanford, California, 1983. Kazemi, H., Merill, L., Porterfield, K. and Zeman, P.: Numerical Simulation of Water-Oil Flow in Naturally Fractured Reservoirs, Soc. Pet. Eng. J. (Dec.) 317-326, 1976. Kazemi, H.: Pressure Transient Analysis of Naturally Fractured Reservoirs, Soc. Pet. Eng. J. (December) 415462, 1969. Laubach, S.E.: Fracture Patterns in Low Permeability Sandstone Gas Reservoir Rocks in the rocky Mountain region, paper SPE 21853, presented at the Rocky Mountain Regional Meeting and Low Permeability Reservoirs Symposium, Denver, Colorado, April 1991. Lee, S.H., Durlofsky, L.J., Lough, M.F., and Chen, W.H.: Finite Difference Simulation of Geologically Complex Reservoirs with Tensor Permeabilities, paper SPE 38002, presented at the SPE Reservoir Simulation Symposium held in Dallas, TX, June 8-11, 1997. Lee, S.H., Jenson, C.L, and Lough, M.F.: Efficient Finite Difference Model For Flow in a Reservoir with Multiple length-scale Fractures, Soc. Pet. Eng. J. (Sept.) 268-275, 2000. Long, J.C.S., Gilmour, P., and Witherspoon, P.A.: A Model For Steady Fluid Flow in Random Three-Dimensional Networks of Disc-Shaped Fractures, Water Resources Research 21, 1105, 1985. Lorenz, J.C., and Hill, R.E.: Subsurface Fracture Spacing: Comparison of Inferences from Slant/Horizontal Cores and Vertical Core in Mesaverde Reservoirs, paper SPE 21877, presented at the Rocky Mountain Regional Meeting and Low Permeability Reservoirs Symposium, Denver, Colorado, April 1991. Lough, M.F., Lee, S.H., and Kamath, J.: An Efficient Boundary Integral Formulation for Flow Through Fractured Porous Media, Journal of Computational Physics 143, 462-483, 1998. Maroongroge, V., Saad, N., Pope, G.A., and Sepehrnoori, K.: Use of Inverse Modeling for Conditioning Geostatistical Models to vertical tracer Profiles, paper SPE 30592,

presented at the SPE Annual Technical Conference and Exhibition, Dallas, TX, October 22-25, 1995. Oda, M.: Permeability Tensor for Discontinuous Rock Masses, Geotechnique 35, 483, 1985. Park, Y.C. and Sung, W.M.: Development of FEM reservoir Model Equipped With Effective Permeability Tensor and its Application to the Naturally Fractured Reservoir, paper SPE 64793, presented at the SPE International Oil and Gas Conference and Exhibition held in Beijing, China, Nov. 7-10, 2000. Press, W.H., Teukolsky, S.A., Vetterling, W.T., and Flannery, B.P.: Numerical Recipes in Fortran: The art of Science Computing, 2nd Edition, Cambridge University Press, 1992. Pruess, K. and Narasimhan, N.T.: A Practical Method for Modeling Fluid and Heat Flow in Fractured Porous Media, Soc. Pet. Eng. J. (Feb) 14-26, 1985. Qasem, F.H.: Performance and Recovery Prediction in Heterogeneous Naturally Fractured Reservoirs Under the Solution Gas Drive Process, PhD Dissertation, University of Southern California, 1996. Ramirez, J., Samaniego, F., Rivera J., and Rodriguez, F.: Trace Flow in Naturally Fractured Reservoirs, paper SPE 25900 presented at the SPE Rocky Mountain Regional/Low Permeability Reservoirs Symposium, Denver, Colorado, April 12-14, 1993. Ramirez, S.J., Samaniego, V.F., Rodriguez, F., and Rivera, R.J.: Tracer-Test Interpretation in Naturally Fractured Reservoirs, paper SPE 28691, presented at the SPE International Petroleum Conference and Exhibition of Mexico, October 10-13, 1994. Rossen, R.H.: Simulation of Naturally Fractured Reservoirs with Semi Implicity Source Terms, Soc. Pet. Eng. J. (June) 210-210, 1977. Saidi, A.M.: Reservoir Engineering of Fractured Reservoirs, General Printing, Singapore, 1987. Sato, K. and Abbaszadeh, M.: Tracer Flow and Pressure Performance of Reservoirs Containing Distributed Thin Bodies, paper SPE 28444, presented at the SPE Annual Technical Conference and Exhibition, New Orleans, LA, September 25-28, 1994. Shinta, A.A. and Kazemi, H.: Tracer Transport in Characterization of Dual-Porosity Reservoirs, paper SPE 26636, presented at the SPE Annual Technical Conference and Exhibition, Houston, TX, October 3-6, 1993. Sutopo, Arihara, N., Sato, K.: Simulation of Naturally Fractured Reservoirs with Effective Permeability, paper SPE 68705, presented at SPE Asia Pacific Oil and Gas Conference And Exhibition held in Jakarta, Indonesia, 17-19 April 2001. Tester, J.W., Bivins, R.L., and Potter, R.M.: Interwell Tracer Analyses of a Hydraulically Fractured Granitic Geothermal Reservoir, Soc. Pet. Eng. J. (August) 537, 1982. Thomas, L.K., Dixon, N.T. and Pierson, R.G.: Fractured Reservoir Simulation, Soc. Pet. Eng. J. (February) 4254, 1983. Users Guide, IMEX, Advanced Oil/Gas Reservoir Simulator, version 2000, Computer Modeling Group, Ltd., Calgary, Alberta, Canada.

SPE 84886

Van Golf-Racht, T.D.: Fundamentals of Fractured Reservoir Engineering, Elsevier Scientific Publishing, Amsterdam, 1982. Wagner, O.R.: The Use of Tracers in Diagnosing Interwell Reservoir Heterogeneities Field Results, J. Pet. Technology (November) 1410, 1977. Warren, J.E. and Root, P.J.: The behavior of naturally fractured reservoirs, Soc. Pet. Eng. J., 3, 245-255, 1963. Wattenbarger, R. C., Aziz, A., and Orr, F.M.: HighThroughput TVD-Based Simulation of Tracer Flow, paper SPE 29097, presented at the SPE Symposium on Reservoir Simulation, San Antonio, TX, February 12-15, 1995.

Yi, T., Daltaban, T.S., and Archer, J.S.: Analysis of Interwell Tracer Flow Behavior in Transient Two-Phase Heterogeneous Reservoirs Using Mixed Finite Element Methods and the Random Walk Approach, paper SPE 28901, presented at the SPE European Petroleum Conference, London, UK, October 25-27, 1994. Zellou, A., Ouenes, A., and Banik, A.: Improved Naturally Fractured Reservoir Characterization Using Neural Networks, Geomechanics and 3-D Seismic, paper SPE 30722, presented at the SPE Annual Technical Conference and Exhibition, Dallas, TX, October 22-25, 1995.

Table 1. Simulation Input Data

Nx Ny Nz Injection rate Tracer mass injection Total tracer amount Matrix permeability Fracture porosity Block porosity Fracture permeability Fracture intensity 70 71 1 100 bbl/day 1 lb/bbl 1000 lbs 0.5 35 md 0.01 0.06 0.2 1000 md 0.1 0.9

SPE 84886

SPE 84886

0.8

0.8

0.6 mD 0.4
mD

0.6

0.4

0.2

0.2

0 0 0.5 PVI 1 1.5

0 0 0.5 PVI 1 1.5

FI:

0.2

0.4

0.6

0.9

FI:

0.2

0.4

0.6

0.9

Figure 3a. Effect of Fracture Intensity on tracer produced (f = 0.02, Km = 5 md)

Figure 3b. Effect of Fracture Intensity on tracer rate (f = 0.02, Km = 5 md)

0.3

0.8

0.2
0.6

mD
0.4

qD
0.1
0.2

0 0 0.5

PVI
0.4 0.6

1.5

0.5

PVI
0.2 0.4 0.6

1.5

0.2

0.9

0.9

Figure 4a. Effect of Fracture Intensity on tracer produced (f = 0.01, Km = 5 md)

Figure 4b.

Effect of Fracture Intensity on tracer rate (f = 0.01, Km = 5 md)

SPE 84886

0.8

0.6 mD 0.4

0.2

0 0 0.5 PVI 0.4 0.6 1 0.9 1.5

FI:

0.2

Figure 5a. Effect of Fracture Intensity on tracer produced (f = 0.06, Km = 5 md)

0.3

0.2

qD
0.1

0 0 0.5 0.2

PVI
0.4 0.6

1 0.9

1.5

Figure 5b. Effect of Fracture Intensity on tracer rate (f = 0.06, Km = 5 md)

10

SPE 84886

a) FI = 0.4
1

b) FI = 0.4 0.3

0.8 0.6

0.2 qD 0.1

mD
0.4 0.2

0 0 0.5 1 1.5

0
0 0.5 1 1.5

PVI
0.01 0.02 0.06 0.01

PVI
0.02 0.06

c) FI = 0.6
0.3

d) FI = 0.6

0.8

0.6

0.2
qD

mD
0.4

0.1
0.2

0 0 0.5 1 1.5

0
0 0.5 1 1.5

PVI

PVI
0.01 0.02 0.06

f :

0.01

0.02

0.06

e) FI = 0.9
1
0.3

f) FI = 0.9

0.8
0.2

0.6

mD
0.4

qD
0.1

0.2
0

0 0 0.5 1 1.5

0.5

1.5

PVI
0.01 0.02 0.06
0.01

PVI
0.02 0.06

Figure 6 Effect of f on Tracer Response for Various Fracture Intensities

SPE 84886

11

1 0.9 0.8 0.7

P e a kA rriv a lT im e ,P V I Matrix Dominated

0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.2 0.4

Fracture Inte ns ity Fracture Dominated

0.6

0.8

0.01

0.02

0.06

Figure 7: Effect of FI on Peak Arrival Time ( Km = 5 md )

0.3

0.8

0.2
0.6 mD 0.4

qD

0.1
0.2

0 0 0.5 0.40 PVI 0.60 1 0.90 1.5

0 0 0.5 0.40 PVI 0.60 1 0.90 1.5

Figure 8a. Effect of Fracture Intensity on tracer produced

(f = 0.02, Km = 0.5 md)

Figure 8b.

Effect of Fracture Intensity on tracer rate

(f = 0.02, Km = 0.5 md

12

SPE 84886

0.3

0.8

0.2
0.6 mD 0.4

qD
0.1
0.2

0 0 0.5 PVI 0.60 0.70 1 1.5

0 0 0.5

PVI
0.60 0.70

1.5

Figure 9a. Effect of Fracture Intensity on tracer produced

(f = 0.02, Km = 2 md)

Figure 9b.

Effect of Fracture Intensity on tracer rate

(f = 0.02, Km = 2 md)

0.3

0.8

0.2
0.6 mD 0.4

qD 0.1
0.2

0 0 0.3 0.6 PVI 0.4 0.6 0.7 0.9 1.2 1.5

0 0 0.5 PVI 0.4 0.6 0.7 1 1.5

Figure 10a. Effect of Fracture Intensity on tracer produced

Figure 10b. Effect of Fracture Intensity on tracer rate

(f = 0.02, Km = 35 md)

(f = 0.02, Km = 35 md)

SPE 84886

13

a) FI = 0.4
1

b) FI = 0.4 0.3

0.8

0.6

0.2 qD

mD
0.4

0.1

0.2

0 0 0.5
PVI

0
1 1.5

0.5
0.5

P VI 5

1
35

1.5

0.5

35

c) FI = 0.6
1
0.3

e) FI = 0.6

0.8
0.2

0.6

mD
0.4

qD

0.1

0.2

0 0 0.5 0.5
PVI

1 5 35

1.5

0.5 0.5

PVI

1 5 35

1.5

Km :

e) FI = 0.9
1 0.8
0.3

f ) FI = 0.9

mD

0.6 0.4 0.2 0 0 0.5

P VI 0.5 5 35
qD

0.2

0.1

1.5

0 0 0.5
PVI

1 5 35

1.5

0.5

Figure 11. Effect of Km on Tracer Response for Various Fracture Intensities (f = 0.02 )

14

SPE 84886

0.8 Peak Arrival Time, PVI

0.6

Matrix Dominated

0.4

Fracture Dominated

0.2

0 0 0.2 0.4 0.6 0.8 1 Fracture Intensity

km:

35

Figure 12. Effect of FI on the Peak Arrival Time (f = 0.02)

900 800 700 Effective Permeability, md 600 500 400 300 200 100 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

km = 35 md km = 5 md

Fracture Intensity

Figure 13. Effect of FI on the Effective Permeability (f = 0.02)