Logic hybrid simulation-optimization algorithm for distillation design

Abstract

Please cite this article as: Caballero, J. A.,Logic hybrid simulation-optimization algorithm for distillation design., Computers and Chemical Engineering (2014), http://dx.doi.org/10. 1016/j.compchemeng.2014.03.016 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ! ! We take advantage of the best of algebraic models and process simulators. AbstractIn this paper, we propose a novel algorithm for the rigorous design of distillation columns that integrates a process simulator in a generalized disjunctive programming formulation. The optimal distillation column, or column sequence, is obtained by selecting, for each column section, among a set of column sections with different number of theoretical trays. The selection of thermodynamic models, properties estimation etc., are all in the simulation environment. All the numerical issues related to the convergence of distillation columns (or column sections) are also maintained in the simulation environment. The model is formulated as a Generalized Disjunctive Programming (GDP) problem and solved using the logic based outer approximation algorithm without MINLP reformulation. Some examples involving from a single column to thermally coupled sequence or extractive distillation shows the performance of the new algorithm.

Conceptual Design of Highly Integrated Processes – Optimization of Dividing Wall Columns

“…For the latter and especially for the separation of nonideal mixtures described by highly nonlinear thermodynamic models, special care for initialization and problem solution needs to be taken. For that purpose problem specific knowledge should be exploited as much as possible .…”

“…Instead of using a metaheuristic, Caballero [84] reports on a deterministic optimization approach that combines the logic-based outer-approximation approach [85] and a simulation model in Aspen Hysis . While advertising the advantages of the combined approach, exploiting the benefits of the available simulation models with the computational efficiency of the deterministic optimization approach, the necessary derivative information requires repeated evaluation of the simulation model with perturbated values for the DDoF [82].…”

“…While advertising the advantages of the combined approach, exploiting the benefits of the available simulation models with the computational efficiency of the deterministic optimization approach, the necessary derivative information requires repeated evaluation of the simulation model with perturbated values for the DDoF [82]. Since the convergence of the simulator presents a critical issue for successful application of this approach as well, Caballero [84] improves the convergence by avoiding the recycles in the simulation model, making use of the equivalent distillation sequence based on the approach of Navarro et al [65]. A similar approach was also used by Javaloyes-Antón et al [82] for the optimization with a metaheuristic.…”

A Hybrid Evolutionary–Deterministic Optimization Approach for Conceptual Design

“…This is especially important, since failure in the simulator convergence can result in a failure of the optimization 10 . The most prominent approach in EA and other population-based methods is the use of penalty functions, which deteriorate the fitness of infeasible individuals 30 .…”

Design of Dividing‐Wall Columns by Mixed‐Integer Nonlinear Programming Optimization

“…There a several implementations of MINLP algorithms in GAMS, but GAMS is a generic optimization platform without any unit operation models, physical property database, and flow‐sheeting capabilities. On the other hand, there are a number of applications that use Microsoft Excel as in the current paper or Matlab for the MILP part and link it with a commercial process simulator for the NLP part. The prerequisite for every successful MINLP solution is a stable and reliable NLP solution. Although sequential quadratic programming (SQP) algorithms for large flow sheets were developed and implemented in most commercial process simulators , they are not frequently used in practice.…”