Noticing that SI units for the dynamic viscosity (μ) are N·s/m2, Pa·s or kg/(m·s), it seems advisable, for the sake of dimensional consistency, to correlate its reciprocal (m·s/kg) with the mass fractions of the components of the polymer solution, in order to predict (estimate) its dynamic viscosity (μsolution) after that of its generic i-components (μi). Such correlation can well take the form: 1/μsolution = ∑ fi·wi/μi. Here, wi stands for the mass fraction of the i-component, while fi are dimensionless coefficients that can be fitted from the available experimental data, or can be taken as equal to the unit if no such data is available. Last case is that of a predictive mixture rule: 1/μsolution ≈ ∑ wi/μi. This polymer solution may possibly include both polymer(s) and (non-polymeric) solvent(s). Please note, however, that the concept of 'intrinsic viscosity' (*) of a solute, which embeds Mark-Houwink equation (**), is usually preferably defined based on the volume fraction of the solute (i.e. polymer) in the polymer solution, rather than on its mass fraction, even if, in practice, mass to volume concentration often substitute for the a priori preferable volume fraction.

(*) https://en.wikipedia.org/wiki/Intrinsic_viscosity

(**) https://en.wikipedia.org/wiki/Mark%E2%80%93Houwink_equation