Minimising environmental impact using CBR: an azeotropic distillation case study

This paper describes a Case Based Reasoning (CBR) system which aids the design of process sequences aimed at the separation of ternary mixtures containing azeotropes. The solution of such tasks is a non-trivial process particularly since many factors including environmental impact must be considered. CBR is an appropriate technique for design since it is rare that a completely new problem is encountered and thus solutions for previous problems found to be environmentally acceptable may be adapted to suit the current situation. The system is able to store, categorise and retrieve appropriate design cases for this class of problem. Also, the system stores indexed design cases within a library. These are organised according to an existing classification system for azeotropic mixtures which enables appropriate cases to be retrieved rapidly and efficiently. Additionally, it is possible for the system to produce an abstract description of the RCM and separation process, so that the designer may more easily apply the library case to the problem in hand. This description is also the first step towards the automation of the case adaptation step. Finally, a case study is presented which illustrates the use of the system.