Computational performance of aggregated distillation models

Compartmental and aggregated modeling is used to derive low-order (reduced) dynamic models from detailed models of staged processes. In this study, the aggregated modeling method of [Lévine, J., & Rouchon, P. (1991). Quality control of binary distillation columns via nonlinear aggregated models. Automatica, 27, 463] is revised with the objective of deriving computationally efficient models for real-time control and optimization applications. A simple implementation of the original method not requiring the specification of compartments is presented. The resulting DAE models are converted into ODE models by pre-solving and substituting the algebraic equations resulting from the reduction procedure, which is the key step to increase simulation speed. To study this, the performances of several full and reduced distillation models, with and without base-layer controllers, are compared using different numerical integrators. It is found that while the reduced DAE models are computationally not advantageous, the reduced ODE models decrease the simulation time by a factor of 5–10.