Synthesis Of Multicomponent Distillation System Based On Rigorous Optimization

The synthesis problem of distillation configurations can be divided into three subproblems :(1)determine the search space of distillation configurations;(2)generate or represent the distillation configurations;(3)optimize the configurations,and choose the optimal distillation configuration.However,on the one hand,the search space which was considered in most of studies is not reasonable(some disadvantage distillation configurations are included,but the good ones are lost).On the other hand,in order to simplify the complexity of the synthesis problem and achieve effective solution,most of existed methods were based on short-cut models,which could not guarantee the reliability of optimal design.In this study,a novel method for generating distillation configurations is proposed to deal with the above problems so that a completer and more reasonable search space can be considered.By combining matrix method with encoding method,the present method is simple and easy to implement.Different from the most of the existed methods,the present method generates the distillation configurations based on the columns directly,rather than determining separation tasks of each stream in configurations.Therefore,some valuable configurations which could not been generated or lost in existing methods can be obtained.Through rigorous simulations and optimization,these configurations are demonstrated having good performance under certain conditions.In addition,a simple criterion is also proposed to effectively remove the configurations in which the product streams will be polluted(cannot meet the product purity requirements).Therefore,the process of generating and synthesizing distillation configurations can be speeded up by removing unreasonable configurations according to the criterion.On the basis of the novel generation method,a method to automatic synthesis of distillation structure is further proposed in this study.This method can be divided into two layers.The distillation configurations are generated in the outer loop,while the evaluation and optimization are in the inner loop.Compared with the existed methods,the present method is more robust and faster with a more complete search space is considered.