We consider the dynamic optimization of chemical processes with changes in the number of equilibrium phases. Recent work has shown that transitions in the number of phases can be modeled as a mathematical program with equilibrium constraints (MPEC). This study generalizes the MPEC to consider dynamic characteristics. In particular, we describe a simultaneous discretization and solution strategy for dynamic optimization problems with complementarity constraints. These discretized problems are then solved with IPOPT-C, a recently developed barrier method for MPEC problems. Our approach is applied to two distillation examples. In the first, we consider the optimal startup of a binary batch distillation problem. In the second, we consider the dynamic operation of a cryogenic column for the separation of natural gas liquids. Both cases demonstrate the effectiveness of the approach on large scale MPEC problems.
Arvind U. Raghunathan, M. Soledad Diaz, Lorenz T.