PMatrix in Numerics Structure
 

Hi people,

I am trying to understand SU2 jacobian matrix structure and components in depth. In numerics structure cpp file there is a Pmatrix and its inverse function. Why it is implemented? I checked where those functions called and see used in implicit application of jacobian matrix. I am aware that those are related to eigen-values and vectors of jacobian as in stated and derived in Hirsh book however what is the main reason to calculate? Any reference beside Hirsh will be appreciated.

Have a nice day.

-Ali

You have to use the diagonalized form instead of the Roe matrix. It may looks like that the Roe matrix A = P lambda P^{-1}. However, it is not in fact. Because you have the entropy correction which modifies the engenvalues lambda.