Abstract
This paper investigates the vanishing dynamic capillarity limit for scalar conservation laws in the context of two-phase flow through heterogeneous porous media. By incorporating a third-order mixed spatio-temporal derivative to represent dynamic capillary pressure, the authors establish a rigorous convergence proof to a weak solution using compensated compactness. Furthermore, a robust numerical scheme is developed to capture non-classical shock waves and standing waves at permeability interfaces, demonstrating that small-scale physical effects significantly influence large-scale flow structures.
Introduction
Modeling fluid transport in heterogeneous soil or rock is fundamental to petroleum engineering and hydrology. These processes are mathematically represented by scalar conservation laws. However, when the medium's permeability ($K$) is discontinuous, the resulting equations admit multiple weak solutions. To identify the physically relevant solution, researchers analyze the "vanishing limit" of regularized models that include small-scale effects like diffusion or capillarity. This paper specifically addresses the complexities introduced by dynamic capillary pressure models in the presence of discontinuous flux functions.
The Vanishing Dynamic Capillarity Approach
While static capillary models are standard in porous media flow, they often fail to account for the temporal dynamics of moving fluid fronts. This study focuses on a dynamic capillary pressure model, which adds a time-dependent component to the physics. The authors define a regularized Cauchy problem that incorporates diffusion, dispersion, and dynamic capillarity (represented by a third-order mixed spatio-temporal derivative). The primary challenge is proving that as the capillarity parameter ($\nu$) approaches zero, the sequence of approximate solutions converges to a stable, physically consistent weak solution of the original conservation law.
Mathematical Analysis and Convergence
The core theoretical contribution of this work is the establishment of a convergence theorem (Theorem 1.1). Using the method of compensated compactness, the authors provide a rigorous proof that the limit of vanishing dynamic capillarity exists and constitutes a distributional solution to the conservation law. This provides a mathematical foundation for using dynamic capillarity as a selection criterion for entropy solutions in systems where rock properties change abruptly.
Numerical Methodology
To validate the theoretical findings, the authors developed a semi-discrete finite difference scheme. Time integration is handled via a third-order Runge-Kutta method. The scheme is designed to be robust, resolving sharp transitions and complex wave structures without requiring an excessively fine grid. The model was evaluated using standardized Riemann problems:
- Homogeneous Case: Flow in a uniform medium ($K=1$).
- Heterogeneous Case: Flow across an interface where rock permeability jumps (e.g., from 1.0 to 1.2).
Key Insights
Non-Classical Shocks
One of the most significant findings is the presence of "non-classical" shock waves. Unlike standard shocks that follow the Lax entropy condition, these waves feature intermediate states and are a direct result of the dynamic capillary effects. The paper demonstrates that these are not numerical errors but essential physical features of the model.
Small-Scale Sensitivity
The research emphasizes that the final macroscopic solution depends explicitly on the specific type of small-scale physics assumed. Ignoring these "tiny" effects is impossible, as they dictate the fundamental structure of the large-scale flow.
Standing Waves in Heterogeneous Media
In cases with discontinuous permeability, the model successfully predicts "standing waves"—discontinuities that remain fixed at the interface between different rock types. These waves satisfy the Rankine-Hugoniot conditions, ensuring flux continuity across the medium's boundary.
Results and Performance
The proposed numerical scheme proved highly efficient, capturing the intermediate states of non-classical shocks even at a relatively coarse resolution of 200 mesh points. Furthermore, the authors demonstrated that rock discontinuity does more than create a standing wave; it fundamentally alters the propagation speeds of both classical and non-classical shocks throughout the entire system.
Limitations and Future Directions
While the current proof and numerical experiments provide a strong foundation, they are limited to one spatial dimension. Real-world reservoirs are three-dimensional and involve significantly more complex fluid chemistry and phase changes. Future work will focus on extending these methods to multi-dimensional porous medium flows and characterizing the vanishing dynamic capillarity limit in higher spatial dimensions.
Summary updated