Synthesis Of Multi-component Distillation Sequences For Nonsharp Separation System

Based on stochastic optimization strategy,a systematic formulation methodology is proposed for synthesizing thermal coupling distillation systems,allowing more than one middle components as the distributing components between a pair of key components in the non-sharp split.Meanwhile,the synthesis problems that allow coexistence of simple columns,prefractionator columns,fully and partial thermal coupling columns,is studied to obtain the best distillation process structure with the mininum total annual cost.Based on the proposed methodology,the distillation system process synthesis with large-scale combinatorial feature and non-linearity can be solved efficiently.1.In order to represent and manipulate the distillation configuration structures,a new coding procedure is proposed in the form of array.First,consider an Ncomponent mixture,and the components {C1,C2,…,Cn,…,CN} are ordered based on their volatilities.Thus,an ordered nature numbers(1,2,…,N)can be used as the codes to name components in the feed.Then,an array [i,j](1?i<j?N)is used,in which “i” represents the light key component and “j” represents the heavy key component,to represent a separate task.Theoretically,an array can represent any kind of split(non-sharp and sharp)without limiting the unmber of distribution components in a non-sharp separation.2.With the application of a binary sort tree approach,a robust flow sheet encoding and decoding procedure is developed so that the problem formulation and solution becomes tractable.Besides,the proposed encoding procedure can not only guarantee the feasibility of any separarion structure,but also enable the direct application of conventionally evolutionary procedure for separation sequence transformation.3.A themal coupling distillation sequence model with non-sharp split,allowing at most two middle components as the distributing components between a pair of key components,is bulit in this paper.Meanwhile,the synthesis problem is formulated as a mixed-integer nonlinear programming(MINLP)problem.In addition,there are two kinds of decision variables in the optimization problem: continuous variables of individual columns which include operating pressure,the recoveries of the pair of key components and reflux ratio,discrete variables of the flow sheet structure which include thermal coupling structures and separation sequences.4.An improved simulated annealing approach is adopted to solve the MINLP problem.The algorithm can not only optimize both continuous and discrete variables simultaneously within a same annealing scheme,but also avoid the complexities of formulation and handling of an explicit MINLP model.