--> Abstract: Predicting Oil Quality - Simulating Reservoir Alteration Processes, by C. C. Walters, H. Freund, S. R. Kelemen, A. L. Braun, and L. M. Wenger; #90091 (2009)

Datapages, Inc.Print this page

Predicting Oil Quality - Simulating Reservoir Alteration Processes

Clifford C. Walters1, Howard Freund1, Simon R. Kelemen1, Ana L. Braun2, and Lloyd M. Wenger2
1ExxonMobil Research & Engineering, 1545 Route 22 East, Annandale NJ 08801
2ExxonMobil Upstream Research Company, PO Box 2189, Houston, TX 77252

As petroleum exploration and production move into deeper waters, higher temperature and higher pressure reservoirs, and basins with significant political instability, accurate risk assessments must consider not only the volume of recoverable reserves, but the quality and value of these reserves. We considered that the available models and methods used to predict petroleum quality were inadequate to meet these challenges. In most cases, reservoir alteration processes, which can greatly impact oil quantity, producibility, and value, are not modeled and pre-drilled predictions of hydrocarbon quality are based on geologic and geochemical correlations with previously tested fluids.

ExxonMobil has developed the Petroleum Quality Simulator (PQS) to improve the pre-drill predictions of oil and gas quality and quantity and our understanding of reservoir charge history. The basic premise of PQS is that the alterations of petroleum composition that occur in reservoir rock results from physical, chemical, and biological processes that can be modeled within an integrated framework. Furthermore, the molecular composition of petroleum must be expressed in sufficient detail such that the chemical alterations involving all reservoir processes can be modeled.

PQS is initialized using the results from basin simulations to define the geologic history of a reservoir. In PQS, a reservoir is defined as a single "tank" that may represent one or more grid elements. The results from basin simulations, which may be tempered by additional expert knowledge, provide the temperature, pressure, and salinity of the reservoir over geologic time. These sources also provide input on the timing, duration, quality, and quantity of petroleum charge and leakage events.

The compositions of the charging fluids are defined by a molecular representation involving 1 to over 10,000 components. Even at this degree of compositional specificity, only single component gases (C1-C5, CO2, H2S, and N2) are defined without compromise. The composition of all other petroleum fluids is approximated by representative species.

Approximately ~4000 of these components define the <1100°F b.p. fraction that consist mostly of homologous series of identified or representative hydrocarbon or NSO-cores with varying degrees of alkylation. Isomer forms are not distinguished individually, though exceptions can be made for specific biomarkers of interest. The remaining species form the residuum and are represented by linked multicore-structures that collective exhibit measured chemical and physical properties of the residuum but individually may not actually exist in petroleum.

Obviously, no basin simulator provides such detailed compositional information. To address this, a library was compiled containing the detailed compositions of "primary" fluids generated and expelled from specific source environments or organic facies (e.g., marine shale, carbonates, lacustrine, ect.) at different levels of thermal maturity. "Primary" fluid compositions are based the laboratory measurements of well-characterized produced petroleum samples that subjectively appeared best to represent fluid from a specific source with no reservoir alteration. To initialize the fluid charge in PQS, the user selects a library fluid that most closely matches the source facies and maturity indicated by basin simulation. A fluid charge also may be expressed as a mixture of library fluids, adjusted to match any available data from test fluids, and/or saved fluids that were altered in a prior PQS simulation.

Fluid charges and leakages may be entered either as instantaneous or continual events. Quantities may be specified in absolute units, but more typically is expressed as a percentage of the total reservoir volume. Once a reservoir is filled, additional gas charges mix with accumulated gas and volumetrically displaces any liquid. Additional liquid charges may be set to bypass or mix with existing liquid than then spill the excess. Reservoir geometry is approximated by simple geometric shapes (e.g. dome, cylinder, cone) where the height and radius is specified by user input.

With the reservoir geohistory defined, the user selects the Equation of State and lumping scheme that will be used to calculate gas-liquid phase behavior and the time-step under which PQS will run. Phase behavior may be calculated by calls to a commercial program, Multiflash, or an ExxonMobil propriety program. Expert users also have the option of changing recommended default parameter values used in the biodegradation and TSR transforms.

Execution of the simulation involves sequential application of reservoir alteration transforms. These transforms include: phase behavior under changing temperature and pressure conditions and hydrocarbon mixing (gas/liquid separation, asphaltene precipitation), biodegradation and water washing (oil-water partitioning), thermal cracking, and thermochemical sulfate reduction (TSR, under current development).

Gas-liquid phase behavior is performed at the start and end of all other transforms. Although the cumulative effects of the alteration transforms can be inspected on each component, this is rarely done. Since the chemical and physical properties of each component is specified, these properties may be calculated for the whole fluid or using weighted averages or, in the case of viscosity, an non-linear summation. Various lumping schemes are embedded that allows the user to query the evolving fluid compositions in terms of bulk properties (e.g. API gravity, viscosity) or chemistry (e.g., % sulfur, TAN, % saturates).

Asphaltene precipitation is based on Scatchard-Hildebrand Theory where the solubility of individual components is compared to the total fluid composition. The solubility parameter of a component is temperature dependent and asphaltene precipitation may occur simply by changes in reservoir temperature, through the addition of incapable fluids (light hydrocarbons), or reservoir depressurization. A decrease in pressure may result in an increase in the partial molar volume of dissolved gases, lowering the effective solubility parameter of the total fluid. Under these conditions asphaltene precipitation continues as long as the gases remain in solution. Once precipitated, asphaltenes are assumed to represent an insoluble solid phase.

Biodegradation is modeled as a first order reaction. The rate is controlled by reservoir temperature, salinity, lithology, and geometry, which influences the nature of oil/water interactions. The total amount of altered oil is calculated first. The model assumes that the residuum components are resistive to biodegradation. The relative proportion removed from each <1100°F lumped homologous series is determined by its relative susceptibility to biodegradation, which was determined empirically by comparing the compositions of produced oils that had been naturally biodegraded to varying degrees. Once the amount removed from each compositional lump is determined, embedded rules determines the amount of removed of each component within the homologous series. Products of biodegradation include fatty and naphthenic acids, which are modeled as a portion of their consumed hydrocarbon precursors, and the gases CH4, CO2, H2S, which may be modeled as retained or expelled from the total reservoir fluid.

An example of modeled biodegradation on oil composition is shown. The relative composition, expressed as lumped chemical fractions, is plotted versus the amount of oil consumed. Overlaid are laboratory-measured compositions of genetically related oils that have undergone reservoir biodegradation. More than 50 wt% of the oil is consumed before saturated biomarker ratios would be affected (> 3 on the Peters-Moldowan biodegradation).

The thermal cracking transform is based on a model developed for visbreaking, a high-temperature, short contact refinery process. Detailed compositions of reactant oils and product fluids were measured at varying time-temperature conditions and a kinetic model was developed that is able to express the observations. Although developed for the downstream, the visbreaking thermal cracking model appears to work very well under geologic time-temperature conditions suggesting that in-reservoir cracking also is purely a thermal process with no significant geocatalytic influences. Evolved coke becomes insoluble pyrobitumen that is inert to further alteration.
PQS has proven useful not only as a forward model to predict oil quality, but also as a reverse model where the composition of altered petroleum is known. Under early phases of exploration and development, the geologic models used for input into PQS are uncertain. By varying these input parameters and comparing the oil compositions of modeled against recovered samples, the charge history can be more tightly constrained. We have found that this reverse modeling is fairly robust and a unique solution requires the fitting to only a few bulk and chemical properties (e.g., API Gravity, % sulfur, TAN, and n-alkane distribution).

PQS is a stand-alone research tool. The compositional model and calculations are far too demanding to be integrated within current basin or reservoir simulators. However, next generation simulator programs and computing hardware may be able to accommodate the full model or it may be possible to construct a simplified version of PQS that captures most of the composition and process information.

Unnumbered Figure.

 

AAPG Search and Discovery Article #90091©2009 AAPG Hedberg Research Conference, May 3-7, 2009 - Napa, California, U.S.A.