Design of Plastics Separation Systems under Uncertainty by the Sample Average Approximation Method

This paper is an application of the systematic design approach to the plastics separation systems synthesis under uncertainty. The problem is distinguished from the traditional chemical distillation processes in that, first there are many separation mechanisms can be used and a combination of them is a necessity and second there is a significant variation of feed composition leading to the design of a flexible system crucial. The problem is formulated as a stochastic MINLP, which is solved by the sample average approximation method proposed previously by the authors. The method can handle large uncertainty space and determine appropriate sample sizes needed, therefore making the systems design under uncertainty of realistic size possible.