TL;DR
This paper compares two numerical modeling approaches for Alkaline-Surfactant-Polymer (ASP) flooding: an "effective" model and a more detailed model coupled with a geochemistry package. Both models show reasonable agreement with experimental core flood data, but the coupled model offers a more in-depth understanding of the complex chemical interactions driving ASP performance, making it valuable for detailed interpretation and upscaling.
What Problem Does This Paper Address?
Enhanced oil recovery (EOR) techniques like Alkaline-Surfactant-Polymer (ASP) flooding are crucial for maximizing oil extraction from mature reservoirs. However, accurately predicting the performance of ASP floods is challenging due to the complex interplay of chemical reactions, phase behavior, and fluid flow within the reservoir. To design and analyze ASP pilots and generate reliable field forecasts, robust and scalable modeling workflows are essential. A critical step in developing such workflows is to validate numerical models against experimental data, specifically ASP core flood experiments. This paper addresses the need for selecting the "right" ASP model by comparing the predictive capabilities and insights offered by two distinct modeling approaches when history-matched against experimental core flood data.
Methods
The paper presents and compares two numerical modeling approaches for ASP flooding:
- Effective ASP Model: This model treats alkali, surfactant, and polymer as chemical "tracers" in the aqueous phase. It simplifies saponification by assuming a fixed acid concentration in the oil and a direct reaction with alkali. Interfacial tension (IFT) reduction is modeled based on salinity and the ratio of soap to surfactant, using empirical correlations derived from phase behavior experiments. Polymer viscosity is modeled using a correlation that accounts for concentration and salinity, with shear-thinning effects considered. Surfactant and alkali adsorption are modeled using a Langmuir isotherm.
- ASP Model Coupled to PHREEQC: This approach extends the effective model by coupling it with the PHREEQC geochemical modeling software. This allows for a more detailed representation of the chemical reactions occurring during ASP flooding, including cation exchange, saponification, and mineral dissolution. Acids and soaps are modeled explicitly as tracers, and their partitioning between the oleic and aqueous phases is considered. This model provides a more mechanistic understanding of alkali consumption and retardation, as well as the complex interactions driving the ASP process.
Both models were history-matched against the same ASP core flood experiment.
Datasets
The primary dataset used in this study is from a single, vertically oriented Bentheimer core flood experiment. The core was 30 cm long and 5 cm in diameter. Before chemical flooding, it was waterflooded to a remaining oil saturation of 50%. The chemical flood involved injecting 0.3 pore volume (PV) of an ASP slug followed by 2.5 PV of a polymer solution. Production data, including oil recovery, pressure drop, and effluent concentrations of alkali, surfactant, and polymer (inferred from viscosity), were collected.
Key Insights and Ideas
- Model Complexity vs. Insight: While both the "effective" and the PHREEQC-coupled ASP models can reasonably match experimental core flood data (oil recovery, pressure, and chemical breakthrough), the coupled model provides a deeper mechanistic understanding.
- Saponification and Geochemistry Matter: The PHREEQC-coupled model's ability to explicitly model geochemical reactions, particularly saponification, leads to a more accurate representation of alkali consumption and retardation, which can significantly impact the timing and effectiveness of the ASP process.
- IFT Modeling Nuances: The paper highlights the importance of accurate interfacial tension modeling, which is linked to salinity and the complex interplay of soap and surfactant. The coupled model's more detailed approach to these interactions can improve predictions.
- Parameter Interpretation: The "effective" model relies on "effective" parameters that are tuned to match data. In contrast, the parameters in the PHREEQC-coupled model are more directly linked to real physical and chemical processes, facilitating better interpretation and upscaling.
- Screening vs. Detailed Analysis: The "effective" model is suitable for screening and when experimental data is limited due to its simplicity and lower data requirements. The PHREEQC-coupled model is better suited for detailed interpretation of core floods or pilot tests, especially when underlying chemical processes are critical.
Key Results
- Both the "effective" ASP model and the PHREEQC-coupled ASP model provided reasonable matches for oil recovery and pressure drop in the core flood experiment.
- Chemical breakthrough times and production profiles were generally well-matched by both models.
- The PHREEQC-coupled model showed improved matching of cumulative oil production and oil production after chemical breakthrough compared to the "effective" model. This improvement is attributed to more accurate modeling of soap and surfactant interactions and salinity effects.
- The coupled model was able to match more detailed experimental data, such as pH and carbonate/bicarbonate concentrations, which are not considered in the "effective" model.
- For the coupled model, increasing surfactant adsorption improved the match of surfactant breakthrough.
- The paper notes that for the "effective" model, Corey exponents for the chemical flood were higher than for the ASP model coupled to PHREEQC to achieve a good match.
Limitations and Future Work
- Limitations of the "Effective" Model:
- Simplifying assumptions are made, leading to "effective" parameters that are tuned rather than directly representing physical processes.
- Alkali and surfactant "adsorption" are simplified representations of complex consumption and retardation mechanisms.
- Does not explicitly model micro-emulsion phases.
- Limitations of the PHREEQC-Coupled Model:
- More complex and requires more experimental input data.
- Computationally more intensive.
- Future Work:
- Research is ongoing to better understand the applicability range of ASP models.
- Development of ASP models that explicitly model the micro-emulsion phase is underway and being implemented in Shell's in-house simulator.
- The paper suggests that the PHREEQC-coupled model's ability to link matched parameters to real processes is important for upscaling from core to pilot scale.
Summary updated