In-Situ Upgrading Process Offers Potential for Heavy-Oil Recovery

You have access to this full article to experience the outstanding content available to SPE members and JPT subscribers.

To ensure continued access to JPT's content, please Sign In, JOIN SPE, or Subscribe to JPT

In-situ upgrading (IU) is a promising method of improved viscous- and heavy-oil recovery. The IU process involves a reservoir being exposed to temperatures greater than 300°C long enough to promote a series of chemical reactions. In this work, the authors developed a numerical model of IU on the basis of laboratory experiences and validated results, applying the model to an IU test published in the literature. Simulation results for the cores submitted have shown that oil production and oil-sample quality were well-predicted by the model.


In this paper, results of a previously published experiment and those of an unpublished experiment (Experiment B) are modeled and discussed. To accomplish this, an in-house code is developed for coupling the kinetic model with the thermodynamic description, a process detailed in the complete paper. This code was able to represent both the reaction dynamics of bitumen-fraction decomposition and the phase distribution and production of pseudocomponents. The resulting model was used to simulate an IU field-scale test (using data from Shell’s Viking pilot).

Experimental Simulation

A 2D homogeneous radial model was used to simulate two IU experiments. Elemental gridblock size was 5 mm in both directions. The model size was 5×30 cells (Experiment A) and 5×138 cells (Experiment B). A total of 30 pseudocomponents was provided in the simulation model.

A sensitivity analysis was performed to identify the effect of some parameters in the IU model on the basis of experimental conditions in the experiments. This analysis was focused on less-certain properties related to the core and physical properties of oil pseudocomponents.

Experimental and Simulation Results. In the process of model development and adaptation, more effort has been dedicated to specify the thermodynamic model than the kinetic model. The final model is capable of successfully predicting liquid production at various test scales and conditions. To explain the mismatch observed between simulation and experimental results of solid generation during IU, the authors performed estimations of solid mass that would be generated under the IU conditions of both experiments in a closed system. The resulting amount of generated solids is the highest possible because the initial hydrocarbons were all retained inside a core. The estimated amount of pyrobitumen for Experiment A was 0.022 kg. This value is still lower than that reported for laboratory conditions (0.024 kg). With regard to Experiment B, taking into account that 35% of the original oil was evacuated by thermal expansion of fluids, the estimated mass of pyrobitumen was approximately 0.070 kg (for closed volume). This value is significantly lower than that reported by the laboratory for this experiment (0.090 kg). The most probable reason is a natural difference in the physicochemical conditions for measurements in kinetic tests and IU experiments.

Field-Scale Simulations

Simulation of Pilot Test. All information related to this project addresses the period 2006–2010. Shell’s Viking project was designed to determine the commercial viability on the IU process in bitumen. The project was implemented in the Bluesky formation in northwestern Alberta. The average temperature around the heaters was 350°C. Production began in early 2006 (after 15 months of heating). Two main production steps were involved

  • Pressurization caused by thermal expansion, from the initial reservoir pressure (35 bar) to approximately 120 bar
  • A slight pressure drawdown until no more oil was recovered

In 2007, a significantly upgraded production was reported, whereas oil production declination was reported in 2008.

The heated pattern size was 34×30×200 m. The total number of wells was 29, with three producer wells and eight observer wells. Seven heaters and 11 heater/producer wells were distributed in the pattern. The heating began in 2004 with a power of 900 W/m per heater; this was kept constant for 2.5 years before being progressively reduced.

Results of this test indicated that oil recovery was approximately 55% original oil in place (OOIP) and the cumulative oil quality was approximately 22 °API after 5 years of operation. The energy ratio reported was 4.2 J/J (output/input) at the end of the thermal process. This represents an energy consumption of 460 kW-h/BOE.

Model Description. A 2D homogenous model was used to simulate the IU field-scale test. To reduce the computational time, a simplified kinetic model was adapted to reproduce the full system of IU reactions. The original kinetic model comprised 23 oleic reactive pseudocomponents, four gaseous byproducts, and two solids (29 hydrocarbon components) taking part in 24 reactions. After applying an optimization algorithm, a reduced kinetic model using the activation energies as fitting parameters was proposed. This reduced kinetic model is composed of eight reactive pseudocomponents, four gases, and one solid. The total number of reactions was reduced to eight. Among the 13 pseudocomponents, the heavy-oil components were considered to be nonvolatile components and the generated gases were treated as noncondensable gases. Water is also included in the model.

Simulation Results. Some simulation results of the Viking project production are presented in this section. As observed in the field test, the oil production appears 12 to 15 months after the start of the operations. Oil production begins earlier in simulations than in the test. The difference in oil production is approximately 5% at the end of the test. Oil production stabilizes after 3.5 years of heating. The simulated recovery factor was approximately 55% OOIP, and it reached approximately 60% at the end of the thermal process according to field data.

Values of instantaneous °API higher than 19 were found after 2 years of heating. According to thermal-property specification for this model, approximately 30% of the total energy provided to the pattern can be expected to be lost after 5 years of operation.

The energy ratio was slightly higher in the simulations than in the reported field test during the first 3 years. This is because of the overestimation of produced oil in the simulations. At the end of 5 years of operation, the energy ratio was 4.3 J/J in the pilot vs. 3.7 J/J in the simulation model. The authors determined that the difference in this energy indicator came from discrepancies in the amount of energy provided by the heaters and underestimation of oil production between 3 and 5 years.

Study of IU Patterns With Reduced Energy Consumption. The possibilities of the energy efficiency improvement during IU in a small pattern were evaluated. Heater and producer distribution for two patterns was studied. A different heater distribution was specified in these patterns with respect to the Viking case. The first pattern included four (full) heaters, whereas the second included two heaters. The first pattern features heater density similar to that of the Viking pilot, while the second pattern has only two central heaters.

In general, cases with four heaters show lower energy efficiency than cases with two heaters. The case with 900 W/m and a heating period of 4 years (for two heaters) is the most-energy efficient case (10 J/J). Cases with four heaters resulted in high (final) cumulative °API gravity and a total recovery factor ranging from 52–70% OOIP. Surprisingly, the case with 900 W/m and only 4 years of heating period produces the highest oil recovery (70% OOIP). However, cumulative oil quality is poor (16 °API), which also can be explained by the lower average temperature in the pattern at the end of the heating period. Finally, the energy consumption is lowest among cases with four heaters per pattern.

This portion of the study indicates that the oil quality is proportional to the energy consumption and, thus, the average temperature reached in the reservoirs.

Discussion of Field-Scale Simulations. Fig. 1 summarizes the amount of produced fluids for the pilot test simulation and two IU experiments presented in this work. In addition, the amount of solid generated in situ and the residual oil in the reservoir are depicted in the figure. Values are presented by mass basis. Experimental results in both cores (0.15 and 0.69 m) indicated that the amount of produced oil varied between 56 and 61 wt%, whereas gas production ranged between 7 and 9 wt%. The amount of produced oil and gas estimated from simulation under the Viking conditions was similar to that obtained in the laboratory.

Fig. 1—Left and center: mass fraction of products and cumulative °API gravity obtained at the end of experiments. Right: mass fraction of products and cumulative API gravity obtained in the Viking test (simulation results) at 5 years.


This study made clear that obtaining a high oil quality during the IU process requires a high-enough density of heaters to provide necessary power continuously. In addition, according to the kinetic scenario, which includes the ­solid-residue deposition and hydrocarbon-gas release, the volume of upgraded oil is limited.


  • Two laboratory IU experiments, field-scale tests, and optimal well configurations were analyzed numerically using a dedicated reservoir simulator, an IU kinetic-model adaptation and coupling with multicomponent thermal equilibrium, and extensive adaptive grid refinement.
  • Three production stages were identified for the IU process at laboratory scale. At the first stage, hydrocarbon-fluid production was controlled mostly by thermal expansion. A second stage was driven by the developed IU process itself, and the third was an inert-gas injection from the top to displace the remaining liquid vertically from the core.
  • Using the developed IU model, the Viking field-scale IU test was successfully simulated. Good agreement in oil production and energy ratio was achieved.
  • A reduction in the number of heaters improved energy efficiency of the process. However, oil quality can significantly decline in this case.

For a limited time, the complete paper SPE 190695 is free to SPE members.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 190695, “Simulations of In-Situ Upgrading Process: Interpretation of Laboratory Experiments and Study of Field-Scale Test,” by A. Perez-Perez, SPE, M. Mujica, and I. Bogdanov, SPE, Computational Hydrocarbon Laboratory for Optimized Energy Efficiency, and A. Brisset, SPE, and O. Garnier, SPE, Total, prepared for the 2018 SPE Improved Oil Recovery Conference, Tulsa, 14–18 April. The paper has not been peer reviewed.

In-Situ Upgrading Process Offers Potential for Heavy-Oil Recovery

01 April 2019

Volume: 71 | Issue: 4



Don't miss out on the latest technology delivered to your email weekly.  Sign up for the JPT newsletter.  If you are not logged in, you will receive a confirmation email that you will need to click on to confirm you want to receive the newsletter.