Study of Performance of a Novel Stochastic Algorithm based on Boltzmann Distribution (BUMDA) coupled with self-adaptive handling constraints technique to optimize Chemical Engineering process

• Published in: Computer Aided Chemical Engineering Vol. 37; pp. 923 - 928
• Main Authors: Murrieta-Dueñas, R., Cortez-González, J., Hernández-Aguirre, A., Gutiérrez-Guerra, R., Hernandez, S., Segovia-Hernández, J.G.
• Format: Book Chapter
• Language: English
• Published: 2015
• Subjects: BUMDA; Distillation train; EDA’s; Optimization; Stochastic Algorithms; Distillation train; EDA’s; Stochastic Algorithms; Optimization; BUMDA
• ISBN: 9780444634290; 0444634290
• ISSN: 1570-7946
• DOI: 10.1016/B978-0-444-63578-5.50149-3

The optimal design of distillation systems is a highly non-linear, multivariable and multimodal problem. The rigorous model of distillation columns is represented by mass, equilibrium, sum and heat equations called MESH equations and phase equilibrium calculations (Thermodynamic model). Furthermore, it has several local optimums and is subject to several kind constraints such as, design and topology scheme constraints, and achieves targets of purity and recovery for each split component. In this paper, we propose the employment of a novel stochastic algorithm called Boltzmann Univariate Marginal Distribution Algorithm (BUMDA, Valdez, S. I. et al., 2013) coupled with self-adaptive handling constraints technique to optimize a well-known distillation process scheme. The optimization problem consists in minimizing the total reboiler duty in a distillation train to split a mixture made of four components. The BUMDA’s performance is compared with Differential Evolution (DE) due to the fact that this last algorithm is used frequently in the optimization of distillation columns. The results show that the BUMDA algorithm is better than the DE algorithm regarding effort computing, quality solution, and time used to find solution. The BUMDA algorithm is efficient, trusted, easy to use and of general applicability in any chemical engineering process.