Global optimization of multicomponent distillation configurations: Global minimization of total cost for multicomponent mixture separations

A novel stochastic optimization method to efficiently synthesize large‐scale nonsharp distillation systems

“…It should be pointed that they treated the continuous variables ( p , ξ LK , ξ HK , and r ) as constant. In order to be consistent with the optimization conditions of Jiang et al, 17 we fixed p , ξ LK , ξ HK , and r as constant, and the best TCDS obtained by the proposed method is presented in Figure 9(A), which had a minimum TAC of 2,540,365.66 $/year. The probability that the proposed method could obtain the best TCDS was estimated by testing the method 500 times; for each test, all initial independent variables were randomly generated, and an optimized solution was collected.…”

“…As stochastic optimization method is featured by a certain probability of reaching the global optimal solution, the ability of the proposed method to ensure global optimality was tested first. To this end, our proposed method was compared with the enumeration‐based deterministic global optimization method of Jiang et al, 17 which synthesized a global optimal TCDS for the separation of a five‐component alcohol mixture. Their data for the separation problem are given in Table S1 in the Supporting Information.…”

“…To guarantee global optimality, global minimization algorithms 16,17 were presented recently, where all distillation configurations were enumerated first by the matrix method 18 . Thereafter, an NLP problem for each configuration was solved using BARON 14 .…”

Global minimization of total exergy loss of multicomponent distillation configurations

“…However, for many cases, the resulting MINLP could not be solved to global optimality due to various convergence difficulties . Despite various attempts to tackle these issues, including proposing an alternative algorithm based on a modified version of logic‐based outer‐approximation algorithm and decomposing the optimization problem into subproblems followed by introducing an iterative optimization procedure to solve them, the MINLP‐based approach still fails to guarantee global optimality …”

“…A number of strategies, such as parallelization and bound tightening techniques, can further bring down the computational time significantly. Later, Jiang et al extended the GMA framework and introduced an enumeration‐based global optimization algorithm for determining the minimum cost of multicomponent distillation configurations. The objective of this article is to develop the first general global optimization algorithm that minimizes the total exergy loss (i.e., maximizes thermodynamic efficiency because the minimum work of separation in Equation is a constant if feed and product specifications are known) for any regular‐column distillation configuration.…”