An Adaptive Refined Grid Search Strategy for Assessing Operational
Flexibility and Application on Refrigerant Selection
Abstract
A novel adaptive refined grid search strategy is developed for
representative characterization of process feasible region boundaries
and accurate estimation of its hypervolume. In particular, a linked list
data structure adopted from the field of computer science is used to
maintain the grid connectivity information. A uniform perturbation
scheme is used to refine the search only near boundaries. The volumetric
flexibility index FI_V is calculated directly from a summation of
feasible hypercubes in the grid, without the need to apply shape
reconstruction techniques. The proposed adaptive grid search strategy
can capture complex region shapes with reduced sampling costs and
without randomness for better reproducibility. Operational flexibility
is optimized traditionally at a process scale. A case study on
refrigerant selection is presented to demonstrated that the developed
strategy could be combined within a computer-aided molecular design
framework for operational flexibility optimization in molecular scale.