A NEW ALGORITHM FOR SELECTION OF SEPARATION TECHNOLOGIES

This paper presents the development of a new algorithm implemented in a rule-based system to select user defined 'n' good separation sequenc es for multicomponent mixtures. The procedural language, Q'NIAL, was c hosen to embed the rule base and to achieve the necessary control stru cture for addressing the needs of automated reasoning. The algorithm w as tested using four sample problems involving gas, liquid, and solid components. The separation sequences suggested by the algorithm were t he same or similar to those found in the literature. Computational tim e is dependent upon the number of sequences desired and also on the nu mber of separation technologies to be considered.