On the application of a nature-inspired stochastic evolutionary algorithm to constrained multi-objective beer fermentation optimisation. (4th January 2018)

• Record Type: Journal Article
• Authors: Rodman, Alistair D.; Fraga, Eric S.; Gerogiorgis, Dimitrios
• Is Part Of: Computers & chemical engineering. Volume 108(2018)
• Page Start: 448
• Page End: 459
• Publication Date: 2018-01-04
• Subjects: Dynamic optimisation -- Nature-inspired optimisation -- Multi-objective optimisation -- Stochastic optimisation -- Solution representation -- Beer fermentation; Chemical engineering -- Data processing -- Periodicals; 660.0285
• DOI: 10.1016/j.compchemeng.2017.10.019
• Languages: English
• ISSNs: 0098-1354

Abstract

Graphical abstract: Highlights: The nature-inspired Strawberry algorithm has been employed for beer fermentation optimisation. The effect of solution representation on stochastic optimisation efficiency has been analysed. The Pareto front of optimal fermentor temperature manipulation profiles is computed and visualised. The profiles which increase ethanol as well as reduce batch time have great industrial importance. The potential to achieve a batch time reduction of over 12 h at higher ethanol is determined. Abstract: Fermentation is an essential step in beer brewing, often acting as the system bottleneck due to the time-consuming nature of the process stage (duration >120 h), where a trade-off exists between attainable ethanol concentration and required batch time. To explore this trade-off we employ a multi-objective plant propagation algorithm (the Strawberry algorithm), for identifying temperature manipulations for improved fermentation performance. The methodology employed successfully produces families of favourable temperature profiles which exist along the Pareto front. A subset of these output profiles can simultaneously reduce batch time and increase product ethanol concentration while satisfying constraints on by-products produced in the fermenters, representing significant improvements in comparison with current industrial practice. A potential batch time reduction of over 12 h has been highlighted, coupled with a moderate improvement in ethanol content.