Fracture Propagation during Water-Alternate-Gas Injection: A Practical Workflow Based on Wellbore Storage Analysis

Although, WAG has been studied and implemented widely for hydrocarbon recovery improvement, still there are some uncertainties regarding flow dynamics in the three-phase region. One of the most important uncertainty areas is modeling of three phase relative permeability in this region. Commercial simulators by using conventional three phase models cannot capture all the features and mechanisms behind the three phase flow. Hysteresis is one of the phenomenons, which is still remained controversial in three phase relative permeability models. Some of the current relative permeability models have accounted for this phenomenon and in others, hysteresis effect is lacking. Two phase models usually fail to predict the actual behavior of the hysteresis in three phase flow. Some of the three phase models also take in to account the features of hysteresis but not all of them. In the simulation process of water alternating gas, hysteresis modeling is a corner stone of the workflow. If this work step is not done properly, the results of the simulation study may not necessarily capture different aspects of the WAG injection scenario and the results could be mis-leading affecting the business decision. In this study, first an extensive parametric paper review and literature survey has been done on hysteretic two and three phase relative permeability models following by a comparative analysis. Parametric investigation is carried out and the advantages and disadvantages of the models are discussed. The model predictions versus selected experimental data are shown and the discrepancies are highlighted. The onset point to improve the model prediction is declared and finally to show the hysteresis effect on oil recovery a simulation case is done using a synthetic sector model. The discussion in this study show that by properly modeling the WAG process and taking into account the hysteresis phenomenon, the robustness of the results on incremental hydrocarbon recovery by WAG process can be significantly improved. It can be concluded that the simulation results using the commercial simulator can be mis-leading in term of the WAG injection performance if hysteresis modeling is not properly done and applied. Introduction Dependency of the global economy on oil and gas resources cannot be ignored. Although, these energy resources are finite and likely to decline, enhanced oil recovery has a significant role to improve the sustainable energy supply. With the ongoing advancement in technical methodologies and technology applications, EOR processes that are complex in nature and characteristic behavior, needs especial attention. Water alternating gas (WAG) is one of the EOR methods, which is increasingly becomes popular due to the successful field trials reported in the literature (Christensen, Stenby et al. 2001). WAG originally implemented to boost the recovery of conventional gas injection by using water cycles to control gas mobility. Although WAG process is widely applied worldwide, there are some uncertainty areas in the evaluation of WAG behavior like prediction of three phase relative permeability during the injection cycles.

This paper reviewed the interpretation of repeat fall-off tests acquired in two vertical pattern injectors operating in a carbonate reservoir undergoing full field development. Water Alternating Gas first pattern (WAG-1) started in August 2002 with a period of continuous gas injection until 2006 when the first water cycle was initiated. In the second pattern (WAG-2) water injection was initiated in June 1998 until September 2007 when the first gas cycle started. A few pressure fall-off tests were acquired during the monophasic injection phase mostly to verify well injectivity. After Water Alternating Gas (WAG) cycles started, pressure fall-off tests were usually acquired at the end of each three-month injection cycle with 1:1 WAG ratio. Analytical fall-off test interpretation relied on the use of the two-zone radial composite model. The apparent permeability thickness product was corrected with the Perrine formulation of multiphase mobility. Triphasic oil permeability was calculated using the modified Stone I model. Multiple fall-off test interpretations indicated that the two pattern vertical injectors behaved differently even though both being fractured. The difference in behavior was linked to different perforated intervals and reservoir properties. Gas and water injection rates were showing differences between both pattern injectors as a consequence. No major operational issue was reported during the three-year operation of both WAG patterns. During the WAG cycles, gas banks were found to be of a small inner radius and almost undetectable at the end of the subsequent water cycle. Changes in the pressure derivative slope at the end of the subsequent water injection cycle indicated most likely the creation of an effective mixing zone of injected gas and water in the reservoir. Numerical finite-volume simulation was required to account for potential injected fluid segregation and, the multi-layered and heterogeneous nature of the reservoir. Repeat saturation logs acquired in observation wells provided critical information on the saturation distribution away from the injection wells. Enhanced vertical sweep conformance through phase mobility control in the presence of strong reservoir heterogeneity was the major expected benefit from an immiscible WAG displacement mechanism. All available observations were reviewed and integrated using a history-matched reservoir simulation sector model with boundary conditions obtained from a full-field model.

Summary In this study, we show significant improvements over conventional pressure-transient analysis on the basis of nonlinear regression by using total least squares (TLS), which minimizes errors in both pressure and time simultaneously. To our knowledge, TLS has not been applied to pressure-transient analysis before in this sense. TLS regression is not an easy problem to solve mathematically, especially for nonlinear pressure-transient-model functions. In this work, we compare four different versions of TLS. We formulate a robust approximation of the TLS solution, which can handle a variety of tests and reservoir models yet does not compromise the performance of the analysis. We show that our technique reduces ambiguity in the estimation of parameters to a large extent, especially in the presence of noise in time. Using our TLS algorithm, we obtain much narrower confidence intervals on the parameter estimates of a variety of real data sets, compared to the conventional least squares (LS) approach. For synthetic data sets, we observe that the TLS estimates are often closer to the true values than estimates made with LS, especially for poorly determined problems. When the deviation is in pressure only, TLS and LS results are comparable. However, in the presence of deviations in time in addition to pressure, the performance of TLS algorithms is substantially better. We, therefore, expect that our technique will provide more accurate estimation of reservoir parameters, allowing for better forecasting of reservoir performance.

The use of the Diagnostic Fracture Injection Test (DFIT) technique as a means of pre-frac investigation has become relatively routine in the oilfield, particularly to understand the reservoir properties and then subsequently optimize the hydraulic fracture design. A key component of an effective DFIT is the performance of an effective After Closure Analysis (ACA) to assess the transmissibility of the formation and thereby allow for effective design. BP Oman is developing the Barik formation, within the Khazzan field, which is a low-permeability conventional tight-gas reservoir within Block 61 of the Sultanate of Oman. The reservoir is comprised of a series of tightly interbedded sandstones and shales, with substantial shale breaks between the principal sand lobes. During the Appraisal and Development well sequence to date, BP Oman have performed DFIT operations in over 50 vertical wells, within the Barik Formation. Each one of these wells was then subject to placement of a large (one million lb) hydraulic fracture treatment. Each treatment was then followed by a standard clean-up programme and when possible a PBU, with subsequent placement on production into the main gathering system. This paper seeks to demonstrate that there is unambiguous evidence of a coherent correlation between the petro-physical Barik open-hole logs, the transmissibility value (as estimated from the ACA), the conventional Pressure Transient Analysis (PTA) as well as the long-term production behaviour. Additionally, the paper will investigate the key aspects of the actual DFIT execution, the data gathering and the analysis that can impact the quality of the correlation. The paper will go on to demonstrate the most efficient methods of achieving the most accurate assessment of the formation transmissibility; that is both indicative and subsequently helpful for the fracture design and post-fracture productivity prediction. This paper successfully describes a 50 well, and growing, DFIT analysis programme and the suitability of the use of the results from the subsequently performed ACAs for forward planning and hydraulic fracture design. Providing a suite of useful and helpful insights, suggestions and recommendations; into how DFIT, for ACA, should be executed in the field; the paper adds an extensive case history to the industry database for future consideration.

A new physical modeling for two-phase wellbore storage due to phase redistribution

During early production of Kashagan Field, the surveillance program is critical for understanding connectivity within the reservoir. Pressure transient analysis (PTA) results in the rim facies of the Kashagan carbonate platform show well bore proximity to high permeability features. By integrating the PTA results in a fine-scale geologic model, the presence and magnitude of geologic features, including faults, karst bodies, and open fractures, can be evaluated as an explanation of pressure results. Good quality pressure transient data can be obtained from down-hole gauges during periods of production down-time. The character of the pressure-response can provide information to interpret reservoir properties such as permeability-thickness (kh) and the effect of geologic features in the vicinity of the well. In the Kashagan rim area, geologic features include seismically-visible karst features, seismically-interpreted faults, and open fractures that can be identified from wireline logs. Because fine-scale details of the flow properties could not be differentiated in the full field simulation model, a fine-scale sector model of the rim area was constructed using Petrel ™ software. By integrating the surveillance and geologic data, the subsurface team can make several key observations. Rim wells with cavernous karst features contain kh values up to two orders of magnitude higher than stratigraphically-equivalent platform interior wells. Wells that produce from open fractures in the rim are commonly adjacent to seismically visible faults and karst geobodies, and the distance from the well to the seismically-visible geologic feature is similar to the distance estimated from the PTA results. At present, the kh interpretations from PTA are the only direct estimates of permeability for the large geologic features in the Kashagan rim. In a fine-scale 3D sector model, the permeability of the geologic objects, including faults, karst geobodies, and open fractures is statistically distributed using the PTA results. The fine-scale sector model demonstrates the value of geologic and surveillance data integration in order to understand PTA results. By establishing a relationship between the distance to geologic objects and PTA-based permeability estimates, a powerful predictive model can be developed to better represent the flow along the rim and guide the placement of future drill-wells.

Water alternating gas injection process (WAG) is a tertiary oil recovery technique commonly used to enhance the displacement efficiency of the residual oil which cannot be recovered during the primary and secondary recovery processes. In this study, experimental tests were conducted in a transparent water wet microporous model packed with glass beads and initially filled with crude oil. Alternate cycles of gas and water were injected to characterize the dynamics of three-phase flows and its associated oil recovery. The displacement of the residual oil at the end of each WAG injection cycles were studied in detail using the state of the art image processing techniques. It was observed that, the continuous cycles of WAG injection have helped the saturated residual oil to spread across the porous region and effectively displaced the residual oil that remained inside the water wet porous media after the initial water flooding. Obtained experimental results helps to understand the WAG process in detail and further it helps to validate future numerical models related to this process.

Permanent storage of carbon dioxide requires accurate estimation of trapped gas saturation and reservoir storage capacity. Hydrogen and gas storage in subsurface reservoirs involves injection and withdrawal cycles, which necessitates the accurate estimation of hysteresis in relative permeability (kr) and capillary pressure (Pc) data. Cyclic injections in the form of Water Alternating Gas (WAG) injection is an effective Enhanced Oil Recovery (EOR) method for improving volumetric sweep efficiency in the reservoirs. When simulating all these processes, subsurface reservoir engineers rely on commercial software to estimate trapped gas saturation and cyclic hysteresis in kr and Pc data. However, few researchers have investigated the reliability of the available hysteresis models implemented in commercial software for this purpose. To this aim, a thorough sensitivity analysis was performed to investigate the performance of three-phase hysteresis models available in the literature for simulating the measured data of cyclic experiments. The Water Alternating Gas (WAG) experiment used in this study was performed at near-miscible conditions on a mixed-wet sandstone sample, and the cyclic injection was started with drainage. The results show that when simulating WAG experiments performed at low injection rates, using a hysteresis model developed by Aghabozorgi and Sohrabi (2019) provides the most accurate results for pressure drop, fluid saturations, and gradual oil production. However, the model developed by Larsen and Skauge (1998) can be only used for experiments reaching maximum gas saturation in their first three-phase imbibition cycle, mostly representing the limited grid blocks near the wellbore. Even in this scenario, the dependence of the model on Land’s initial-residual relationship and trapping coefficient is questionable. This study highlights the possible pitfalls of using commercial software for estimating trapped saturation and hysteresis in relative permeability and capillary pressure. It also provides useful suggestions for improving the accuracy of performed simulations. The presented results are vital for the oil and gas experts as they can demand improvements in current approaches and promote constant growth and development of commercial software.

Summary Secondary oil recovery for the Jay/Little Escambia Creek (LEC) Fields Unit will exceed initial estimates by 27 × 10(6) bbl [4.3 × 10(6) m3] because of innovative reservoir management based on a comprehensive surveillance program and detailed reservoir description data. Infill program and detailed reservoir description data. Infill drill wells have accounted for 76 × 10(6) bbl [12 × 10(6) m3] of oil production. The mature waterflood was phased into a tertiary recovery project in 1981, and early phased into a tertiary recovery project in 1981, and early performance is generally consistent with the planning performance is generally consistent with the planning study, which predicted that 47 × 106 bbl [7.5 × 106 m3] of tertiary oil will be recovered. Introduction The Jay/LEC field waterflood was initiated in 1974. Waterflood performance has been extremely good, as indicated by the estimated primary plus secondary reserves of 346 × 10(6) bbl [55 × 10(6) m3] being surpassed in May 1983 with the field producing 40,000 B/D [6400 m3/d] of oil at the time. Ultimate recovery from primary and secondary is now expected to be 373 × 10(6) STB [59 × 10(6) stock-tank m3] or 51 % of the original oil in place (OOIP). The major contributing factor to this success was effective reservoir management, which originated from results of reservoir description and surveillance programs. Following 8 years' technical evaluation, a fieldwide miscible gas displacement project began in Jan. 1981 using available field methane gas. High-pressure N2 injection began in Dec. 1981 when this cheaper gas and high-pressure compression facilities became available. By mid-1984, 2.9% hydrocarbon pore volume (HCPV) of the 20% HCPV design of miscible gas had been injected. Current injection operations use a total of 61 × 10(6) cu ft [1.7 × 10(6) m3] N2 and 172,000 B/D [27 000 m3/d] water in a water-alternating-gas (WAG) scheme. As expected, water injectivity has decreased significantly but presents no serious operational problem. A total of 29 spinner surveys have been conducted problem. A total of 29 spinner surveys have been conducted in 18 injection wellsto define the initial gas profiles andto determine the effects of injection pressure and duration of the gas injection cycle on the profiles. Initial nitrogen breakthrough occurred in Dec. 1982 and 14 of the 67 currently producing wells have experienced nitrogen production. Oil production response from these wells has been identifiable but varies significantly between wells. This paper discusses the performance of the highly successful waterflood and describes how, in the later life of the flood, it was phased into the tertiary oil recovery project. project. Reservoir and Fluid Properties The Jay/LEC field is located in the Florida panhandle and south Alabama. The oil accumulation is in the Smackover carbonate and Norphlet sand formations. Oil occurs mostly in the dolomitized portions of the Smackover. The formation is found below 15,000 ft [4572 m] and has an average thickness of about 350 ft [107 m]. All wells drilled in the field, including the infill wells, were cored conventionally through the Smackover formation. Original reservoir pressure was 7,850 psi [54 MPa] about 5,000 psi [35 MPa] above the saturation pressure of 2,830 psi [20 MPa]. No free hydrocarbon gas pressure of 2,830 psi [20 MPa]. No free hydrocarbon gas has been released from the reservoir oil, since minimum pressure was 4,800 psi [33 MPa]. Table 1 presents a pressure was 4,800 psi [33 MPa]. Table 1 presents a summary of rock, fluid, and reservoir properties. Waterflood Performance Following unitization and flood start in March 1974 a comprehensive surveillance program was implemented. Surveillance data were combined with the abundance of reservoir description information to provide a highly successful reservoir management program for this complex carbonate waterflood. Surveillance data were obtained from several production logging tools to determine initial and subsequent vertical flood conformance. Additionally, radioactive tracers, reservoir pressure surveys, and routine well test data were used to define areal conformance. Several programs to improve recovery evolved from this effort. Surveillance techniques and resulting programs are treated in more detail in Refs. 1 and 3. The programs are treated in more detail in Refs. 1 and 3. The infill drilling program, which began in 1977, is particularly significant. To mid-1984, a total of 37 infill particularly significant. To mid-1984, a total of 37 infill wells were drilled, and the average well spacing was reduced from 160 acres [647 × 10(3) m] per well to 112 acres [453 × 10(3) m]. Table 2 presents the updated results of this program. Infill wells have accounted for 76 ⨯ 10(6) bbl [12 × 10(6) m3] of oil production and currently are producing 17,200 B/D [2700 m3/d] oil. JPT P. 261

Summary The gas huff ’n’ puff (HnP) technique holds great potential for enhancing oil recovery from shale reservoirs following primary recovery. Both CO2 and hydrocarbon (HC) gases are commonly considered for HnP processes, each offering distinct advantages and disadvantages. In this study, we propose a novel experimental approach to systematically evaluate the enhanced oil recovery (EOR) potential of different injectants in shale cores and the potential carbon storage capacity for CO2 HnP. The HC gas mixture utilized consists of 72% methane (C1), 16% ethane (C2), and 12% propane (C3), closely mirroring the composition of available HC gas in oil fields. In each test, shale cores were first saturated with the dead oil, and then high-pressure methane diffusion was utilized to make in-situ live oil. Primary recovery was performed before HnP cycles. Next, six cycles of gas HnP were conducted with a huff pressure of 6,000 psi, a soaking time of 24 hours, and a puff pressure of 2,000 psi. The huff pressure was selected to be much higher than the calculated minimum miscibility pressure (MMP) to enable the development of miscibility of injected gases with the oil in the matrix. The EOR performance of each injectant was evaluated by analyzing oil recovery through nuclear magnetic resonance (NMR), core mass measurements, and effluent fluid collected at each step. The produced gas and liquid effluent were carefully collected and analyzed by gas chromatography (GC) analysis. The experimental results revealed that both CO2 and HC gas enhance oil recovery from the studied shale cores. After six cycles of HnP process, oil recoveries reached 84% for CO2 and 74% for HC gas. Oil recovery efficiency was about 58 STB/MMscf of gas injected. NMR results confirmed that most of the injected oil could be extracted after six cycles of gas injection. However, incremental oil recovery decreased significantly with each subsequent cycle for both gases. In addition, oil produced in later cycles generally had a higher fraction of heavy HC components, as evidenced by density measurements and GC analysis. CO2 can be stored in the pore space of the shale matrix, with up to 1.56 kg/ft³ of the shale bulk volume if the production is ceased in the final HnP cycle.

Summary The influence of after flow on transient pressure responsehas long been recognized in well test analysis. The type-curve approach based on the simplified model of aconstant wellbore storage coefficient has attempted to assessafterflow effects from pressure data alone. Theintroduction of a new generation of production logging tools thatcan measure bottomhole flow (BHF) and pressure (BHP)simultaneously allows the afterflow in the well to bemonitored at the surface. Consequently, the well test canbe more rigorously interpreted by using a new form ofvariable rate analysis, and the shut-in time required toallow the detection of a straight line on the appropriatesemilog plot [e.g., Homer, or Miller-Dyes-Hutchison(MDH)] is much reduced. The analysis of the pressureand flow data transmitted from the production loggingdevice may be carried out at the wellsite in real time so that the test can be continued until the desired objectives are achieved. A field example shows how continuous BHF measurement during the wellbore storage period can improve the interpretation of a short-duration build up test. Introduction The constant rate solution of the flow equation in aninfinite reservoir is the basis of the conventional analysisof well tests. Even if the reservoir is bounded, the analysis applies while the reservoir is infinite acting-i.e., untilthe pressure change reaches the outer boundary. For the conventional analysis, it is essential to maintain constantrate, and it is well known that this is difficult to achieve. The best way to maintain constant rate is to make it zero, hence the wide use of buildup tests. Because of the compressibility of the fluid in thewellbore, the BHF does not adjust instantly to thesurface flow after a change of rate is imposed at the wellhead. This phenomenon is called "afterflow" in a buildup testor, more generally, "wellbore storage effect. "Duringthis time, the sandface rate is not constant, and early datacannot be used in the conventional (i.e., Homer or MDH)analysis. Furthermore, outer boundary effects maydevelop before the wellbore storage effect disappears, inwhich case the conventional analysis does not apply. Hence, the necessity of interpreting early data. This interpretation is possible through a more generalsolution of the flow equation, with a variable sandfacerate as inner boundary condition. The rate can be obtainedfrom a direct measurement or by other means. The firstattempt known to us of direct measurement in a builduptest was presented by Gladfelter et al. The afterflow ratewas obtained by measuring the rise of the liquid level inthe wellbore. The analysis used a semilog plot of the"corrected pressure," which is the ratio of the pressure dropto the sandface rate variation. Another method is to estimate the sandface rate froma mathematical model. Van Everdingen and Hurstpresented an exponential function of time as an presented an exponential function of time as an approximation of the sandface rate. In the currently acceptedtype-curve analysis, it is estimated from the derivative ofpressure with respect to time, on the assumption that the pressure with respect to time, on the assumption that the wellbore storage coefficient and the surface rate areconstant. A comprehensive review of this technique ispresented in Ref. 4. presented in Ref. 4. Recently, wireline tools capable of simultaneousmeasurement of sandface rate and pressure (as well asfluid temperature and density) were introduced. The firstbuildup tests recorded with these tools in oil and gas wellsalmost invariably revealed abnormal situations such asvery long afterflow or backflow and phase segregationin the wellbore. Such observations justifiably may leadus to question the validity of sandface rate estimated frompressure data alone. We present an analysis of the pressure data alone. We present an analysis of the wellbore storage concept and show briefly how a direct sandface measurement could improve the type-curve analysis. Our main purpose is to propose a new analysis basedon simultaneously measured sandface rate and pressure. The technique makes use of a rate-convolved time function similar to the superposition time function of Odehand Jones. The wireline aspect of the measurement offers the possibility of analysis in real time. We presentthis analysis for a buildup test in a homogeneous reservoir. Shortly after the original presentation of this work, several authors presented more comprehensive applications of direct sandface rate measurements. Fetkovich andVienot introduced their rate-normalized type curves, Kucuk and Ayestaran used a deconvolution process toidentify the reservoir model, and Stewart et al. appliedthe rate-convolved time function in nonhomogeneous reservoirs. JPT P. 143

Multiphase transient analysis of horizontal wells during CO2-EOR

The models that describe distributed temperature behavior in a reservoir appear as advection-diffusion equations with forcing terms that include compressibility effects due to Joule-Thomson and adiabatic expansion. The model lends itself to solution techniques like Operator Splitting which decomposes a multiphysics, multiscale problem into components representing different simple physics and scales, where each component can be solved independently at each time step using any appropriate technique for the physical process. Unlike diffusion, advection does not destroy information and it is nonsmoothing and preserves sharp boundaries. By Method of Characteristics (MOC), a process can be estimated as it is advected downstream, from an initial data line, along characteristic paths at some characteristic velocity. This idea has useful implications on what information is carried by the advection processes that occur in a reservoir, such as heat advection. In this paper, the results of solving the temperature model by Operator Splitting are presented. This approach utilizes the fact that temperature signatures, generated by spatial pressure changes in the reservoir due to Joule-Thomson and adiabatic expansion effects, are transported downstream in time from the initial line of generation to the wells. Along the corresponding characteristic curves, this transport is described by the characteristic velocity, and time to reach the well. Importantly these curves are independent of the pressure model but are dependent on the convective thermal velocity. The result is that reservoir structures that generate significant spatial pressure changes produce temperature signatures that are transported along the characteristic solution lines from the point of generation to the wells. Back calculations using the arrival times and and corresponding velocities give the location of the reservoir structures. This technique was applied to the calculation of damaged radius, stimulated radius and inner boundaries of a multiring reservoir system, using temperature measurements and showed that near wellbore properties that are masked by wellbore storage in pressure transient analysis can be revealed by the slower moving temperature field. A proposal for new well test design to utilize this approach is also presented.

Acidizing wells after completion or even after some period of production is a standard practice in naturally fractured reservoirs. Such a practice allows easier and faster removal of mud materials from the fractures intersecting the well. In some reservoirs, acidizing is performed deep into the formation for the purpose of enlarging the width of fissures and enhancing their interconnectivity. This results in a composite system with two zones, each having different characteristics, including storage capacity ratio, interporosity flow coefficient, and permeability. Composite systems may also result from fluid injection. High injection of fluids causes larger opening of fractures near the wellbore (due to higher pore pressure), which results in an increase of fracture permeability around the well. A model for a composite naturally fractured system has been developed by assuming pseudo-steady state fluid transfer from matrix to fractures, and taking into consideration the effects of: wellbore storage, near-wellbore formation damage and skin at the interface of the two zones. The inner zone is considered to be finite, but the outer zone can be bounded or infinite. The loglog plots of the pressure and pressure derivative as a function of the storage capacity and permeability ratios of the two zones provide unique characteristics, which were used to derive equations for calculating various parameters of the two zones, e.g. radius of the inner zone, mechanical skin, and skin at the interface. From the interpretation of the loglog plots of the pressure and pressure derivative data, using Tiab's Direct Synthesis (TDS) technique, we have found that:the contribution of the matrix to the flow can occur in the inner zone or the outer zone, depending on the product of the interporosity flow factor and radius of the inner zone;When the ratio of the dimensionless radius of the two zones is less than ten, the naturally fractured reservoir behaves like a homogeneous system;For storage capacity and permeability ratios greater than unity, the pressure derivative curve has a hump, which may appear before or after the trough; the coordinates of this maximum point may be used to calculate the permeability of the outer zone in the absence of the second radial flow line; andthe presence of a thin skin zone at the interface influences the starting time of the radial flow line of the outer zone. Several numerical examples are included in the paper. Introduction Naturally fractured reservoirs have been studied intensively during the last 30 years in geological and engineering fields. Transient pressure analysis has received particular attention. In naturally fractured reservoirs, the unique pressure behavior due to the double porosity medium and their interaction has lead to the development of a variety of sophisticated models. These models differ from each other in a way the matrix/fracture fluid transfer is handled. The aim of these models is to evaluate the special flow characteristic and provide useful information from interpretation of pressure data. One of the earliest models was developed by Pollard to quantify skin resistance from pressure buildups and used to evaluate acid treatment in fractured limestone. Not only the skin, but also the fracture and matrix pore volume could be determined from the graphical analysis. The pressure differential in the well was assumed to be in response to three distinct flow regions connected in series. The first region is the damaged or improved region in the fracture near the well (skin region). The second region considers the pressure loss due to the flow resistance within the fractures, and the last one is the pressure response between the matrix and the fracture system. The flux between the matrix and fractures and between fractures and wellbore was assumed to be under pseudosteady state condition. Warren and Root 19 modified Barenblatt et al.'s solution to account for fracture compressibility. They present a comprehensive and applicable solution to the double porosity, naturally fractured problem.

The effectiveness of heat injection into a target formation has a great impact on the efficiency of bitumen and heavy oil recovery and energy savings under many steam heating processes such as the startup phase of SAGD (Steam Assisted Gravity Drainage) (Butler 1991). However, this parameter is hard to calculate due to many unknown variables such as variations in operational conditions and steam saturation along the horizontal wellbores, heat return rates, and losses to the vertical section above the target formation. This paper proposes a new technique to estimate cooling time and formation thermal diffusivity by using thermal transient analysis (TTA) along the horizontal wellbore under a steam heating process. A novel concept of a heating ring is also introduced to measure the heat storage in the heated bitumen at the time of testing. Heating ring can be considered analogous to a drainage area in a conventional pressure transient analysis. The proposed cooling time and formation thermal diffusivity calculated along the horizontal wellbore can be used to assess the effectiveness of the conduction heating. Cooling time in this paper is defined as the theoretical time required to cool the heated formation to the initial formation temperature. A longer cooling time indicates a higher net heat gain in the formation while the calculated thermal diffusivity is used to predict wellbore conditions and the type of fluid saturation along the horizontal wellbore. Thus, a combination of cooling time and formation thermal diffusivity can be employed to assess the effectiveness of heat injection during various steam heating processes. By knowing the effectiveness of each heating scenario, the process can be selected and optimized not only to save heat energy and steam consumption but also to enhance bitumen recovery. This paper is limited to the new heating operation processes.