- 3:17 PM
304a

Rigorous and Exact: the “What Not to Teach” of Property Estimation

Kenneth R. Cox, Chemical and Biomolecular Engineering, Rice University, 6100 Main Street, Houston, TX 77005-1892

This paper highlights some of the challenges of teaching product design concepts to undergraduate students, while building on the fundamental foundations provided by lower level courses. The treatment of physical property estimation is illustrative. As with so many other concepts, the detailed methods themselves are often identical to those applied in process design. Yet, the application of those methods are very different for the two types of design. Chemical product design and chemical process design do have a lot in common, but students should be taught to view property estimation differently for the two applications.

Much of effective design involves the strategy of developing alternatives, then selecting the best alternative for a particular application. If we had to develop a detailed process simulation for each alternative choice of equipment in a chemical process design, we would severely limit the number of alternatives we could consider due to time and resource limitations. We are able to greatly increase the number of equipment alternatives we can consider when the initial judgments are made with simplified calculations based on heuristics and rules of thumb. A good example of this is sequencing a train of separation devices where only distillation and sharp separations are considered. A process involving just ten products would force us to consider thousands of possible sequences. For such a process, simplified calculations in the initial stages of design become a necessity. Once the preliminary decisions are made based on simpler calculations and experience, then detailed simulations can be employed to advantage for the final optimizations.

This same general approach can be applied to product design. The difference is that we are no longer selecting equipment for a process flow diagram. Instead, in product design we generally are selecting particular functional groups in molecules or molecules in formulations. Since we start with so many possibilities, we find it useful to use experience based heuristics and empirical approximations as the starting points for our design. There are generally too many possibilities to eliminate in the early stages of a design to become overly concerned with ten-digit accuracy and setting up timeconsuming calculations. Once we narrow the possibilities, we can then combine more rigorous calculations with laboratory studies for our final optimizations.

We will present several examples involving both formulation and molecular selection. In each case, parallels will be drawn to techniques already employed in the practice of chemical process design. Yet, in each case we will see that the students must evolve beyond expectations they developed during the study of lower-level courses in the curriculum.