Abstract:This paper is concerned with a space-time adaptive numerical method for instationary porous media flows with nonlinear interaction between porosity and pressure, with focus on problems with discontinuous initial porosities. A convergent method that yields computable error bounds is constructed by a combination of Picard iteration and a least-squares formulation. The adaptive scheme permits spatially variable time steps, which in numerical tests are shown to lead to efficient approximations of solutions with localized porosity waves. The method is also observed to exhibit optimal convergence with respect to the total number of spatio-temporal degrees of freedom.
What problem does this paper attempt to address?
The paper aims to solve the problems caused by the nonlinear interaction between porosity and pressure in nonlinear porous media flow, especially in the case of discontinuous initial porosity. Specifically, the paper focuses on the nonlinear interaction in transient porous media flow, which may lead to the formation of porosity waves. These problems are particularly important in geothermal engineering applications, such as the formation of porosity waves due to high temperature when simulating ascending magma.
### Main contributions of the paper:
1. **Adaptive space - time method**:
- An adaptive space - time numerical method is proposed to solve the transient porous media flow problem with discontinuous initial porosity.
- This method combines Picard iteration and least - squares formulation and can generate convergent solutions with computable error bounds.
2. **Space - time adaptive scheme**:
- It allows for spatially varying time steps, which has been shown to be effective in approximating local porosity waves in numerical tests.
- The method exhibits optimal convergence in terms of the total number of space - time degrees of freedom.
3. **Model description**:
- A transient porous viscoelastic model is considered, which can be regarded as a generalization of the previously introduced model.
- The model equations are as follows:
\[
\frac{\partial \phi}{\partial t}=-(1 - \phi)\left(\frac{b(\phi)}{\sigma(u)}u + Q\frac{\partial u}{\partial t}\right),
\]
\[
\frac{\partial u}{\partial t}=\frac{1}{Q}\left(\nabla\cdot\left(a(\phi)(\nabla u+(1 - \phi)f)\right)-\frac{b(\phi)}{\sigma(u)}u\right),
\]
where \(a(\phi)=a_0\phi^n\), \(b(\phi)=b_0\phi^m\), \(a_0, b_0>0\), \(n, m\geq1\).
4. **Innovation of the numerical method**:
- Compared with existing numerical methods, such as the finite difference method and the adaptive wavelet method, the method proposed in this paper shows better convergence and efficiency when dealing with discontinuous initial porosity.
- Through adaptive meshing, local features can be captured more accurately near discontinuity points without requiring extremely high mesh resolution.
5. **Theoretical analysis**:
- The convergence of the algorithm is proved, and a posteriori error estimates are provided.
- The effectiveness and optimal convergence rate of the method are verified through numerical experiments.
### Conclusion:
By proposing an adaptive space - time method, the paper effectively solves the problem of nonlinear interaction between porosity and pressure in porous media flow, especially in the case of discontinuous initial porosity. This method is not only strictly proven theoretically but also shows good performance in practical applications.