Superstructure optimization-based process synthesis is generally regarded as theoretically powerful; however, it has not been widely used in practice since it typically results in large-scale non-convex Mixed-Integer Non-Linear Programs (MINLP) which are very hard to solve effectively. To address this limitation, we propose a framework leading to substantially simpler formulations through the replacement of complex first-principle unit models by compact and yet accurate surrogate models. We show how all the relevant variable relationships established by a unit model, can be expressed in terms of a subset of the original model variables. We discuss how this subset of variables can be identified, and we present a method to develop high quality surrogate models through artificial neural networks. Finally, we propose a tailored surrogate model reformulation to incorporate binary variables that allow activation/deactivation of particular units within the superstructure model. An example is presented to illustrate the application of the proposed framework.
All Science Journal Classification (ASJC) codes
- Chemical Engineering(all)
- Computer Science Applications
- Process optimization
- Process synthesis
- Surrogate models