[oliver.glahn]

Hello,
does anyone else uses the Ferziger method to evaluate the iteration error?
I wanted to use the method described in „FURTHER DISCUSSION OF NUMERICAL ERRORS IN CFD“ from 1996 to quantify my iteration error based on target values.
See the attached image of the formula for the part of linear convergence.

I find it very difficult working with it and interpreting the error.
If my solution is about to converge and the difference between the iterations decrease, my eigenvalue (calculated as shown in the image) tends to be <1 but very close to 1. This would result a negative and very small denominator for the error calculation.
Therefore I am wondering in what circumstances it could be used, or if I have an interpretation mistake.
Furthermore I am wondering what are the recommended ranges of such calculated error and why is it related to the eigenvalue based on iteration differences instead the absolute value, as we usually do for relative errors?
Does anyone know how to work and interpret the results of such method?
Thank you very much
Oliver