[shekharc]

Hi, I have a heated flat-surface with water running over it parallely....a sort of forced convection.

The fluid water is assumed to have its physical properties varying with temperature, i.e. density, thermal conductivity etc. are temperature dependent.

To define the local Nusselt number at a position "x", what temperature should be appropriate for evaluation of the fluid properties?

Should it be the wall temperature T(x). Or, should it be the film temperature Tf(x), which is an average of the wall temperature T(x) and the far-field temperature T_infinity(x)?

In short:
Nu(x) = h(x) * x / k(x).
k(x) = Thermal conductivity at location x = k(T(x)).
T(x) = T_solid(x) or 0.5*[T_solid(x) + T_infinity(x)] ?

Thanks!