Multi-period optimization for the design and operation of a flexible power-to-methanol process
Andrea Maggi, Jens Bremer, Kai Sundmacher
Abstract
The increasing share of renewable resources in the context of energy transition scenarios requires new methodologies for the design and operation of chemical production facilities, which must adapt to the unsteady nature of their power supply. In this contribution, a highly flexible, fully electrified Power-to-Methanol process, supplied with unsteady wind power generated within the system boundaries, is designed by means of a large-scale NLP multi-period optimization for profit maximization. The problem is constrained by detailed models of interconnected units, feasibility conditions, and discretized power loads (periods) associated with their probability of occurrence. External power must be integrated into the plant to sustain feasible operations when the renewable input is not sufficiently available. Results show that the price at which the external power is purchased determines whether the resulting flexible-plant configuration is competitive with a comparable plant, optimized for steady-state operations ensured by large hydrogen or electricity buffers. Intermediate configurations represented by small buffers and semi-flexible operations constitute an important compromise for future applications of this novel approach.