Publication
Dynamic optimization can be used to determine optimal input profiles for dynamic processes. Due to plant-model mismatch and disturbances, the optimal inputs determined through model-based optimization will, in general, not be optimal for the plant. Modifier adaptation is a methodology that uses measurements to achieve optimality in the presence of uncertainty. Modifier-adaptation schemes have been developed for the real-time optimization of plants operating at steady state. In this paper, the concept of modifier adaptation is extended to transient plants such as batch processes. Two different schemes are proposed, and their performance is illustrated via the simulation of a semi-batch reaction system.