On the Carbuncle Origins from Moving and Stationary Shocks

  • Keiichi Kitamura
  • Eiji Shima
Conference paper


Hypersonic flow computations still suffer from anomalous solutions such as a “carbuncle phenomenon” [1-3]. We still lack an accepted explanation for those anomalies, and we feel there is no single cause, nor is there any single cure. In the present study, we take the viewpoint that the shock anomalies are partly caused by the lack of mathematical expression for internal shock structure by the governing equations, and that they can be examined by numerical experiments. Quirk [1] introduced a benchmark test for numerical schemes on their responses to the captured (fast) moving shock. In this test, the shock took all the possible locations within a cell but instantly passed through them. Roberts [4] chose a more slowly moving shock which took 50 time steps to travel a single cell, and discussed a post-shock numerical noise propagation. Kitamura et al. [3] dealt with a stationary shock located within a cell with an initial shock position parameter ε=0.0, 0.1, ..., 0.9, i.e., the shock was placed at one of 10 possible locations in a cell. Their study discovered that any flux functions including FVS by Van Leer [5] are prone to carbuncles, though some of those methods had been believed to be carbuncle-free accoring to Quirk [1]. In other words, Quirk’s test for a moving shock was not enough to examine robustness of a numerical flux for a stationary shock. Moreover, in spite of those findings, the origin of the carbuncle remained a mystery. The present study will pursue its clue by bridging the gap between the works explained above, i.e., by varying the shock propagation speed from 0 of Kitamura et al. to 6 in Quirk’s choice and beyond. Following our earlier work, different flux functions by Roe [6], Van Leer [5], Liou (AUSM +  − up [7]), and Shima and Kitamura (SLAU [8]) will be used since they have different degrees of robustness against the shock.


Stationary Shock Hypersonic Flow Shock Strength Flux Function Single Cure 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Quirk, J.J.: Int. J. for Numer. Methods in Fluids 18, 555–574 (1994)MathSciNetzbMATHCrossRefGoogle Scholar
  2. 2.
    Pandolfi, M., D’Ambrosio, D.: J. Comput. Phys. 166, 271–301 (2001)MathSciNetzbMATHCrossRefGoogle Scholar
  3. 3.
    Kitamura, K., Roe, P., Ismail, F.: AIAA J. 47, 44–53 (2009)CrossRefGoogle Scholar
  4. 4.
    Roberts, T.W.: J. Comput. Phys. 90, 141–160 (1990)MathSciNetzbMATHCrossRefGoogle Scholar
  5. 5.
    Van Leer, B.: Lecture Notes in Phys. 170, 507–512 (1982)CrossRefGoogle Scholar
  6. 6.
    Roe, P.L.: J. Comput. Phys. 43, 357–372 (1981)MathSciNetzbMATHCrossRefGoogle Scholar
  7. 7.
    Liou, M.S.: J. Comput. Phys. 214, 137–170 (2006)MathSciNetzbMATHCrossRefGoogle Scholar
  8. 8.
    Shima, E., Kitamura, K.: Parameter-Free Simple Low-Dissipation AUSM-Family Scheme for All Speeds. AIAA J.(accepted for Publication), AIAA Paper 2009-136 (2009)Google Scholar
  9. 9.
    Anderson Jr., J.D.: Modern Compressible Flow with Historical Perspective, 3rd edn. McGraw Hill, New York (2003)Google Scholar
  10. 10.
    Van Leer, B.: AE520 Compressible Flow I. In: Fall 2006 Course Pack. University of Michigan (2006)Google Scholar
  11. 11.
    Kitamura, K., Shima, E., Nakamura, Y., Roe, P.L.: AIAA J. 48, 763–776 (2010)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Keiichi Kitamura
    • 1
    • 2
  • Eiji Shima
    • 1
  1. 1.JEDI CenterJAXASagamiharaJapan
  2. 2.JSPS Research FellowJapan

Personalised recommendations