Operation and design optimisation of industrial low-density polyethylene tubular reactor for multiple objectives using an evolutionary algorithm-based strategy
Multi-objective optimisation (MOO) of a low-density polyethylene (LDPE) production in a tubular reactor is performed for two problems with three different objectives: maximisation of monomer conversion and minimisation of operating cost for problem 1; maximisation of productivity and minimisation of...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Published: |
Springer Nature
2023
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/106517/ http://dx.doi.org/10.1007/s41660-023-00308-z |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Multi-objective optimisation (MOO) of a low-density polyethylene (LDPE) production in a tubular reactor is performed for two problems with three different objectives: maximisation of monomer conversion and minimisation of operating cost for problem 1; maximisation of productivity and minimisation of operating cost for problem 2. As a precaution against a run-away in the tubular reactor, an inequality constraint for the reactor temperature is also imposed. The multi-objective evolutionary optimisation algorithms (MOEA), namely the Pareto envelope-based selection algorithm II (PESA-II), the multi-objective evolutionary algorithm based on decomposition (MOEA/D), and the strength Pareto evolutionary algorithm II (SPEA-II), are used to execute the optimisation problem with the Aspen simulator as a model-based optimisation for LDPE in a tubular reactor. Prior to that, model validation and a variables selection methodology based on the Pearson correlation coefficient (PCC) are devised for the selection of the appropriate decision variables for the MOO. The final inputs for MOO’s decision variables are the jacket flowrate of zone 5, initiator 2, and the length of zone 5. Performance matrices including hyper volume, spacing, and pure diversity are employed to select the most effective MOEA method. Based on the results of the comparison study, the most effective MOO strategies were SPEA-II for problem 1 and MOEA/D for problem 2. This is due to the fact that the discovered solution set provided the most precise, diverse, and appropriate in homogeneity allocation points along the Pareto front (PF). |
|---|