Idea: randomised inflow + random-ish distorted cylinder + Eilmer shock-fitting
--> bow shock profile + ML
--> general relation for shape for grid gen?

• python 3
• Eilmer
• PyTorch

• FNN code is from Kyle Damm's shock_angle-ml, modified to suit

Mach number -> [2.0, 8.0]
temperature -> [300.0, 800.0] K

R1 constant, 1.0
R2, K1, K2 randomly set around 1.0

R1 = 1.0
R2 = 1.024010845383956037e+00
K1 = 5.559842733496153100e-01
K2 = 2.493385792650427979e-01
M = 7.985859177691415844e+00
T = 6.662489193564000516e+02

correct coefficients polynomial fit for shock in polar coordinates:

coeffs = [1.329053681099242601e+00, -2.381378425592319166e-03, 1.815951977252002747e-01, -7.257403682239993437e-02, 2.686813008048670737e-01, -1.739286381676704019e-01, 5.769665095911536562e-02]