Choosing turbulence parameters
 

Which approach is better for the turbulence conditions - eddy viscosity ratio, intensity If we don't know what values to use.

Either inputting some values that suit the physics (or) Using expressions
to gather them as the solution progresses for instance - volumeAve(eddyviscosityratio)@DomainName

This question seems to miss the point a bit. For the common 2 eqn turbulence models you have two terms to define - a turbulence energy component (ie have intense the turbulence is) and "something else", where the something else could be a length/time scale, dissipation rate or something else.

So you need to define these two parameters for turbulence. Then just define them in the way which matches what you are trying to model best.