| Knowledge-based tools are used to design separation sequences where non-key components distribute. Specifically, frame-based, rule-based, and assumption-based truth maintenance systems are used. The design process utilizes information for classification, opportunistic decisions, alternative generation and evaluation, and modification guided by constraint satisfaction. By applying the appropriate information at the appropriate stage of the design process, the process and final design are more readily understandable by the designer. The use of non-sharp separators where non-key components can distribute, agree with the intuition that the minimum separation and blending should take place. An apparent lower bound on separation purity is defined and these values and the Fenske equation are used to calculate a minimum number of stages and model the non-key component distribution. The minimum number of stages is used to opportunistically produce a partial sequence ordering, which in turn is used to automatically produce alternatives. The structure of these sequences, the modelled distribution, which is an overestimation, and the product specifications, are used to guide the separation purity adjustment. |
