So, the advection term is non-linear and therefore usually being treated explicitly. Now, there is a confusion of mine.
Consider the case where N-S eqn has been discretized as
(I-L)*u+C_t*u_t=u_t-1/rho * Dx*p, where I is the identity matrix, L is the standard Laplacian operator for momentum space, C_t is the convection term which has been constructed via previous timestep, u_t is the velocity at the previous timestep. Does Rhie-Chow interpolation deals with the coefficients, which is including the convection term or not?
So I want to know is
A=I-L or A=I-L+C_t
Then this leads to the modified explicit convection expression of C_t*u_t+1.
Previously, I was doing with A=I-L and this seems to blow up the solution but had not verified if the wrongdoing was from the numerical coding of mine or the fault of the theory, where the coefficient has not been properly set.
Thank you.