These datasets are a variety of compressible turbulent channels performed in the spirit of Coleman, G. N., Kim, J., Moser, R. D., "A numerical study of turbulent supersonic isothermal-wall channel flow." Journal of Fluid Mechanics 305 (Dec. 1995), 159-183. DOI: 10.1017/s0022112095004587.

This page is a supplement to the 2014 dissertation Reducing Turbulence- and Transition-Driven Uncertainty in Aerothermodynamic Heating Predictions for Blunt-Bodied Reentry Vehicles by Rhys Ulerich. Refer to Sections 3.1 and 3.2 of that document for the form and nondimensionalization of the governing compressible Navier--Stokes equations. Refer to Section 5.3 for background on the simulations, in particular some known limitations of the data. Please consider citing the thesis if you find this data useful towards publishing research.

Final ensemble averages as well as each *in situ*
sample are provided in a compressed HDF5 archive. For example,
dataset `/bar_T` contains instantaneous planar-averaged
temperature profiles while the `mu` HDF5 attribute
of `/bar_T` contains the ensemble profile. Temporal
traces of the L^{2} and root-mean-square of conserved
state are available in associated log files. The end of
the log also contains performance information for the code Suzerain
which produced this data. Spectra are *not* normalized
but *are* comparable with Figure 1 from Coleman et al.
The scenarios are labeled according to the code units used—
please refer to the quantities of interest below for centerline Mach
numbers and friction Reynolds numbers.

Uncertainties were computed using the reference implementation of
the autoregressive technique described within "Estimating
uncertainties in statistics computed from direct numerical
simulation", by Todd A. Oliver, Nicholas Malaya, Rhys Ulerich
and Robert D. Moser, Phys. Fluids 26, 035101 (2014), DOI: 10.1063/1.4866813. Standard
errors are stored as attributes on each Reynolds-averaged dataset.
For example, attribute `mu_sigma` of dataset `/bar_T`
is the pointwise standard error associated with the ensemble average
in attribute `mu`.

- Reynolds 3000, Mach 0.1
- HDF5 (337 MB, MD5 40ded1a60f8d65a7282c5afaa8d43e10)
- Log (2.4 MB)
- y
^{+}≈10 spectra: k_{x}, k_{z} - y≈1 spectra: k
_{x}, k_{z} - Reynolds 3000, Mach 0.5
- HDF5 (340 MB, MD5 65561dbb6918e828afed9cf31a61f600)
- Log (2.5 MB)
- y
^{+}≈10 spectra: k_{x}, k_{z} - y≈1 spectra: k
_{x}, k_{z} - Reynolds 3000, Mach 1.5
- HDF5 (260 MB, MD5 73ee5929a5fc4b726e6095ffe1ed2589)
- Log (1.9 MB)
- y
^{+}≈10 spectra: k_{x}, k_{z} - y≈1 spectra: k
_{x}, k_{z} - Reynolds 3000, Mach 3.0
- HDF5 (335 MB, MD5 f9147b944f908af9b4dfc86b000df89a)
- Log (2.4 MB)
- y
^{+}≈10 spectra: k_{x}, k_{z} - y≈1 spectra: k
_{x}, k_{z} - Reynolds 5000, Mach 0.1
- HDF5 (272 MB, MD5 1581eee7c8e4ff4a227f67cdf8357235)
- Log (1.8 MB)
- y
^{+}≈10 spectra: k_{x}, k_{z} - y≈1 spectra: k
_{x}, k_{z} - Reynolds 5000, Mach 0.5
- HDF5 (284 MB, MD5 12658b01075e5acd49098f519fb11918)
- Log (1.8 MB)
- y
^{+}≈10 spectra: k_{x}, k_{z} - y≈1 spectra: k
_{x}, k_{z} - Reynolds 5000, Mach 1.5
- HDF5 (323 MB, MD5 25d4944b8729aa7308f65302279172fe)
- Log (2.1 MB)
- y
^{+}≈10 spectra: k_{x}, k_{z} - y≈1 spectra: k
_{x}, k_{z} - Reynolds 5000, Mach 3.0 (coarse)
- HDF5 (315 MB, MD5 5113239ac101f2b7c357fc70f8f49721)
- Log (2.0 MB)
- y
^{+}≈10 spectra: k_{x}, k_{z} - y≈1 spectra: k
_{x}, k_{z}

Quantities of interest taken from ensembles over at least 11 flow
throughs follow. The code units employed are per the scenario name.
For example, the `coleman3k01` case used `code_Re` =
3000 and `code_Ma` = 0.1 with the latter involved in computing
the centerline Mach number `Ma_c` from velocity `u_c`
and sound speed `a_c`. Scenarios marked with a star (*)
used somewhat coarse near-wall resolution as compared against
Coleman et al. though their streamwise and spanwise directions were
better resolved. Beware values report far too much precision given
sampling uncertainty.