Numerical and experimental observation of three-dimensional unsteady shock wave structures

On the Influence of Reynolds Number on the Mach Stem Height at Shock Reflection from a Circular Cylinder

“…The case of a cylindrical obstacle has been investigated in a number of studies, and experimental results indicate that the regular reflection pattern, identified by visual inspection of experimental images, appears to be maintained longer on the cylindrical surface when compared to the straight wedge case [5]. This finding is, however, at variance with recent numerical results [6,7,8]. They show that in the inviscid case the transition occurs at the same wall angle as for the straight wedge.…”

“…With the aforementioned ranges of p 1 and R, the Reynolds number Re (defined in the previous section) can vary between approximately 32,000 and 2×10 6 . In the tests by Takayama & Sasaki [5], p 1 was kept constant in the shock Mach number range investigated here, and Re varied as a function of cylinder radius between approximately 500,000 and 8.8 × 10 6 . Using two models in the same experiment has the advantage that such tests are fully free from any problems related to shot-to-shot repeatability of the shock tube as both models are subjected to the same shock wave.…”

“…Numerical simulations indicate that in its early stages the Mach stem may only be a few micrometers long [6], which is undetectable with typical optical visualisation systems, even if high-resolution recording material is used. Higher image magnification improves the spatial resolution and allows one to resolve smaller features, but while this approach reduces the aforementioned discrepancy between predicted transition point and first detection of a Mach stem, imaging limitations do not allow one to eliminate it [9].…”

Artificial Wind—A New Framework to Construct Simple and Efficient Upwind Shock-Capturing Schemes

“…The finite-volume code is based on a triangular unstructured grid with control volumes being constructed around each node. The respective formulas may be found in [14,16] and very closely resemble (76), (78), with the summation of fluxes for all faces (in 1D cases we have just two faces). In the original code an exact Riemann solver is used to compute fluxes at each face from the "left" and "right" values.…”

“…In Section 4 the intermediate statesŪ(ξ ) between U i−1 and U i were introduced using linear interpolation (see (16)). Let us now include an internal discontinuity at ξ = ξ * ; i.e.,…”