A systematic tuning procedure of thermodynamic models, for realistic simulation of multicomponent distillation trains, is presented. The procedure allows a straightforward identification of dominant model parameters and tuning conditions. It is demonstrated that the sequential nature of the distillation train splits, generates a unique matrix of dominant parameters for the multicomponent separation process. The methodology is illustrated by tuning an equation of state to the fractionating section of a large ethane based ethylene plant. The tuned simulator predictions are compared with industrial test data.