Summary The nuclear magnetic resonance (NMR) signal obtained from conventional oil, heavy oil, and bitumen formations can consist of both hydrocarbon and watersignals. Each NMR signal can further characterize both mobile and immobile fluids in the porous media. However, as the viscosity of the hydrocarbon phase increases and the NMR signal shifts toward lower relaxation times, the composite NMR signal for the hydrocarbon-bearing formation becomes very complicated. As the viscosity of the bitumen exceeds 100,000 cp (at natural conditions), the relaxation characteristics of bitumen become partially nondetectable by both the logging and laboratory NMR tools. As a result, the conventional methods of NMR detection fail to precisely recognize the hydrocarbon components. Laboratory NMR measurements of bitumen-bearing porous media under different temperatures were performed. This method delivered new information about bitumen reserves in situ. The results show that low-field NMR holds promise for the characterization of recoverable heavy oil and bitumen reserves. This new approach can be applicable for real-time monitoring of thermal extraction, including monitoring the efficiency of thermal recovery methods. Introduction NMR logging tools are currently used for determining reservoir properties such as porosity1,2 and permeability,1–4 as well as the presence of mobile and immobile fluids.5-8 Recent developments in NMR research offer tools for separating water, oil, and gas from the combined NMR signal.9–11 Very little is known about the use of NMR logging tools for the in-situ characterization of crude oils.1 With respect to heavy-oil and bitumen formations, NMR logging has not been very successful in characterizing crude oil (viscosity>100,000 cp). The reason for this is the fact that the spectra from most heavy-oil and bitumen formations cannot be adequately detected by the NMR logging tools. This is because the shortest relaxation times (t2's) of the spectra at normal temperature conditions (T 30°C) are lost. High-field NMR technology has solved such problems in the past but is not currently available to be used downhole.12,13 A fundamental objective of the research performed in our laboratory was to extend the use of NMR logging tools to heavy-oil and bitumen formations, particularly during thermal recovery projects. To this end, the NMR characteristics of these types of hydrocarbons in bulk volume and in porousmedia were investigated.8,14 The objective of this work was to investigate the NMR characteristics of these bitumens and heavy oils at elevated vs. ambient temperatures and to isolate the oil signal from the combined NMR spectrum of the formation. The hydrocarbon and water saturations were then determined. The possibility of increasing the quality of NMR data by increasing the signal-to-noise ratio and by proper reconstruction of the wholet2 spectra was also investigated. These objectives were achieved by performing a series of experiments, which addressed the following issues:Variable-temperature NMR spectra determination for bitumen-saturated cores to estimate different fluid components in porous media in situ.NMR characterization of brines, conventional oils, heavy oils, and bitumen in bulk volume at different temperatures.Estimation of the parameters of NMR tools and their applicability for monitoring thermal recovery processes. It must be noted that La Torraca et al.15 provided laboratory data that correlate NMR properties to viscosity of heavy oils ranging from <1,000 cp to >100,000 cp. Then they combined NMR log and conventional log data to predict the in-situ oil viscosity in two heavy-oil reservoirs. This work,15 although similar in nature to the work presented here, deals with oil reservoirs having 3 to 4 orders of magnitude less viscosity. Unfortunately, algorithms presented in the literature seem to collapse when applied to bitumen formations. Experimental Phase All field measurements for bitumen sands characterization were performed with a Schlumberger CMR-200™ logging tool. All measurements were done at natural in-situ conditions; the maximum recorded temperature was T=14°C.One example of these results is presented in Fig. 1. The entire NMR laboratory testing was performed with a custom-built Numar Corespec 1000™. This is a unique system with a separate temperature control for heating the magnet and the sample. The equipment was installed and tuned at the TIPM Laboratory and operates at a frequency of 1 MHz. All the samples were tested at different temperatures and at ambient pressure with a standard methodology developed for NMR log calibration. 16 All decay data were translated into NMR spectra with algorithms developed in-house and the NUMAR standard analysis packages 6 that are included with the spectrometer. Several sets of experiments were performed to address each of the issues mentioned earlier. Variable-Temperature NMR Spectra of Core. Variable-temperature NMR was used to determine the fluid components in bitumen- and water-saturated cores. The first set of experiments involved the testing of native state cores at different thermal conditions. Native-state bitumen-saturated plug samples were cut from full-size core using a liquid-nitrogen-cooled cutter. Testing started with NMR measurements at the following temperatures: 1°C, 6°C, 8°C, 12°C, 16°C, 22°C, 25°C, 30°C, 40°C,45°C, 50°C, 60°C, 65°C, 70°C, 75°C, and 80°C. For all measurements, a CPMG sequence was applied with interecho times of 0.3 and 0.6 ms. The NMR spectra were recovered. After completion of the temperature-cycle testing, the native-state bitumen-saturated plugs were saturated under vacuum with an aqueous paramagnetic solution (2N CuSO4, T2 1ms). The fact that water entered the cores under vacuum implies that some drying occurred during core handling. This procedure was performed to eliminate the water signal from subsequent NMR testing. The samples were measured at 30°C. Following the NMR data collection with the paramagnetic solution, the sand samples were cleaned using the Dean-Stark method (thus removing all bitumen) and resaturated with 2%NaCl brine. The brine-saturated samples were tested in the NMR again at the previous temperatures. This determined that the NMR spectrum of the sample was free of any bitumen effects, measuring only structural and mineralogical effects. X-ray diffraction (XRD) analyses were performed to determine clay type and concentration. In accordance with these analyses, a set of artificial samples(sand + clay) was constructed. All samples were saturated with 2% NaCl brine. The brine-saturated artificial samples were tested to determine the NMR spectra, again at the temperatures mentioned above. Examples of the obtained spectra set are presented in Figs. 2 through 4.

During production operations in heavy oil and bitumen formations where thermal recovery methods are applied, the fluids produced are often in the form of emulsions. This is also true in non-thermal recovery methods whenever oil and water are coproduced, but to a lower degree of severity. Conventional flow measuring devices are capable of measuring oil and water streams when they are segregated, but they fail when oil-in-water or water-in-oil emulsions form. Conventional methods are also not reliable when there are solids flowing in the stream. Low field NMR relaxometry was successfully tested as a tool for accurately measuring the oil and water content of such streams with and without emulsions present in the samples. The method was proved to be at least as good as conventional extraction methods (i.e., Dean-Stark). The technology was tested with both artificially and naturally occurring emulsified streams with accuracy better than 96﹪. This extremely encouraging result led to the design of an online NMR relaxometer for oil/water stream measurements under the conditions encountered in the production of heavy oil and bitumen. Introduction In the recovery of bitumen, viscosity reduction becomes important, both below and above the ground. The addition of a liquid diluent is thought to break down or weaken the intermolecular forces which create high viscosity in bitumen(1). The effect is so dramatic that the addition of even 5﹪ diluent can cause a viscosity reduction in excess of 80%; thus, facilitating the in situ recovery and pipe line transportation of bitumen. The knowledge of the bitumen-diluent viscosity is highly important, since without it, calculations in upgrading process, in situ recovery, well simulation, heat transfer, fluid flow, and a variety of other engineering problems would be difficult or impossible to solve. This paper presents the development of a simple correlation to predict the viscosity of binary mixtures of bitumen-diluent in any proportion. Experimental The data used for the development of the correlation was TABLE 1: Bitumen data at 30 °CDATA[C. Available In Full Paper. TABLE 2: Diluent data at 30 °CDATA[C. Available In Full Paper. obtained from Wallace et al.(2) and Wallace and Henry(3).The data consisted of a total of 99 points obtained from three bitumens and five diluents, respectively, listed in Tables 1 and 2. Each of these bitumen samples was diluted at 30 °CDATA[C to 5, 10, 25, 50 and 75 weight ﹪ diluent with each of the diluents. After mixing, the samples were reweighed, and any weight loss was attributed to solvent evaporation. The diluent weight fractions were adjusted accordingly, and the viscosities of the mixtures measured. For a detailed account of experimental procedures, refer to Wallace and Henry(3). Correlation Development Many correlations have been developed to predict the viscosity characteristics of bitumen-diluent mixtures(1-6). While several have been successful in making these predictions, most are cumbersome to use. Low Field Nuclear Magnetic Resonance (NMR) relaxometry techniques were developed in the laboratory to enhance and support comparable NMR logging tools that are currently used downhole.

No abstract

No abstract

Enhancing oil extraction from oil sands with a hydraulic fracturing techniquehas been widely used in practice. Due to the complexity of the actual process, modelling of hydraulic fracturing is far behind its application. Reproducingthe effects of high pore pressure and high temperature, combined with complexstress changes in the oil sand reservoir, requires a comprehensive numericalmodel which is capable of simulating the fracturing phenomenon. To capture allof these aspects in the problem, three partial differential equations, i.e., equilibrium, flow, and heat transfer, should be solved simultaneously in afully implicit (coupled) manner. A fully coupled thermo-hydro-mechanical fracture finite element model isdeveloped to incorporate all of the above features. The model is capable ofanalyzing hydraulic fracture problems in axisymmetric or plane strainconditions with any desired boundary conditions, e.g., constant rate of fluidinjection, pressure, temperature, and fluid flow/thermal flux. Fractures can beinitiated either by excessive tensile stress or shear stress. The fractureprocess is simulated using a node-splitting technique. Once a fracture isformed, special fracture elements are introduced to provide in-planetransmissivity of fluid. Effectiveness of the model is evaluated by solvingseveral examples and comparing the numerical results with analytical solutions.The model is also used to simulate large-scale laboratory hydraulic fracturingexperiments. Introduction Hydraulic fracturing technique has been a fast growing technology since itsfirst application in 1947. By 1988, more than one million hydraulic fracturingtreatments had been performed(1), and today this technique is one ofthe most important methods in enhancing oil extraction from wells. Hydraulicfracturing in oil and reservoirs plays an even more important role. Due to lowtemperature and low permeability of oil sand deposits and high viscosity ofbitumen, oil is virtually immobile(2). Hence, any attempt for insitu oil extraction should employ one of the following techniques: cyclic steamstimulation, in situ combustion, or hydraulic fracturing. Despite the fact that hydraulic fracturing technology has advancedsignificantly over the past fifty years, our ability to model the process hasnot changed as rapidly. As a matter of fact, this technique has been sosuccessful that in the past, designing the treatment with a high degree ofprecision was not of any interest. ut as the industry moved towardsapplications of very high volume/rate, and highly engineered and sophisticatedhydraulic fracturing treatments, the demand for more rigorous designs in orderto optimize the procedure have become more important. On the other hand, without a thorough understanding of the physical process and the factors thatare involved, our ability for an optimal design is limited. Modelling fluidflow combined with heat transfer in the reservoir has been used by the industryfor a long time, and the fracturing process was often designed based ontwodimensional closed-form solutions, such as Geertsma-deKlerk(3), or GdK in brief, and Perkins-Kern(4) and Nordgren(5), or PKN.