Shock Dynamics under Non-conservative Finite Volume Discretizations in One Dimension

  • Guglielmi, Riccardo (Politecnico di Milano)
  • Rausa, Andrea (Politecnico di Milano)
  • Re, Barbara (Politecnico di Milano)
  • Guardone, Alberto (Politecnico di Milano)

Please login to view abstract download link

Accurate shock capturing is a fundamental requirement for numerical methods in compressible CFD, as even small errors in shock speed or position may lead to significant global inaccuracies. A crucial aspect to ensure correct shock propagation is the conservative nature of the numerical discretization. However, in practical CFD applications, non-conservativity may arise for several reasons. For instance, when the numerical solution has to be transferred between different computational grids due to adaptation or re-meshing, the use of interpolation techniques that do not exactly preserve mass renders the method globally non-conservative. Alternatively, the choice of non-conservative variables in the solution of the governing equations leads to different weak solutions, with a different shock speed prescribed by the Rankine-Hugoniot relations. In this work, we investigate the impact of non-conservativity on shock dynamics using one-dimensional model problems for conservation laws. A fully conservative finite volume formulation is taken as a reference, from which the conservative shock position $X_S^{C}$ and speed $V_S^{C}$ are derived. Non-conservative effects are then introduced through different mechanisms, leading to the non-conservative shock position $X_S^{NC}$ and speed $V_S^{NC}$. By means of a geometrical analysis, we derive analytical relations to estimate $X_S^{NC}$ (or $V_S^{NC}$) from the knowledge of $X_S^{C}$ (or $V_S^{C}$). These expressions depend only on the global mass discrepancy and on the solution jump across the shock. Numerical results confirm the theoretical predictions and demonstrate that non-conservative effects can lead to systematic and measurable errors in shock tracking. The present analysis highlights the critical role of conservation in shock-capturing methods and provides a diagnostic framework for assessing the reliability of numerical schemes in the presence of shocks.