Heuristic procedure for synthesis of separation system for multicomponent mixtures using fuzzy logic

Abstract

In this study, the task is to develop a heuristic procedure for the synthesis of systems for the separation of multicomponent mixtures, which would take into account the uncertainty of the factors of the separation processes and the multivariance of the solution.
 The authors reviewed the current state of the existing methods for the separation of mixtures. Many works devoted to the synthesis of effective systems for the separation of multicomponent mixtures leave unsolved the problem of taking into account the uncertainty of many factors of the separation processes. In the presented work, it is proposed that a heuristic approach is applied using fuzzy logic.
 The proposed procedure is based on the choice of a separation sequence that takes into account the value of the separation coefficient and the difference in boiling points of two adjacent key components. The proposed strategy is based on an estimate of the likelihood of each rule for each possible separation. The search for solutions has five stages. First, the mixture to be separated and the available equipment are determined. After that, an ordered list of all separation options is drawn up with possible values of the economic costs of their implementation. Next, the mass load of possible separation is estimated and the separation with the minimum value of the separation ratio is selected. The heuristic is that the heavy separation is done last. And, at the last step, the most economical separation scheme is selected.
 The above strategy at the stage of quantitative assessment is subject to the value of the threshold S (boiling point difference). This point can be used to determine the quasi-optimal solutions to the problem of sequencing according to the rules of relaxation of separation. 
 For this fuzzy heuristic method for determining cost-effective separation sequences for mixtures, appropriate software has been developed. In addition, since the procedure uses the theory of fuzzy sets, the Fuzzy Logic Toolbox package of the Matlab was also used to solve the problem of synthesizing schemes for separating hydrocarbon mixtures. The proposed procedure was successfully tested in both software implementations using the example of separating a five-component mixture. 
 The main difference between the proposed approach and other approaches based on fuzzy logic for the synthesis of separation systems is the gradation of working rules and possible quantification of a set of rules, which eliminates conflicts between the rules.