The identification of a dynamic simulation of a one-point high purity distillation column is investigated using different noise levels. The dynamic and stochastic components of a discrete transfer function are then identified and shown to be adequate using standard time series analysis techniques, provided the signal to noise ratio is sufficiently high.