Open guillaumetousignant opened 4 years ago
Seems to work with -D instead of D_hat (WHY??) and no fluxes for a while. There is a disparity at boundaries because there should be fluxes.
Maybe has something to do with the fact that a function should be multiplied by 1/sqrt(1 - x^2) when integrating?
Maybe the derivative matrix is not correct?
Textbook p. 31 maybe has something, it's describing the coefficient for Chebyshev derivatives.
Chebyshev polynomials should be drop-in replaceable with Legendre ones. Right now it doesn't work, the curves go crazy pretty quick.