A model-data based systems approach to process intensification

In recent years process intensification (PI) has attracted much interest as a potential means of process improvement to meet the demands, such as, for sustainable production. A variety of intensified equipment are being developed that potentially creates options to meet these demands. Their developments, however, are largely due to experiment based trial and error approaches and while they do not require validation, they can be time consuming and resource intensive. Also, one may ask, can a truly new intensified unit operation be obtained in this way? An alternative two-stage approach is to apply a model-based synthesis method to systematically generate and evaluate alternatives in the first stage and an experiment-model based validation in the second stage. In this way, the search for alternatives is done very quickly, reliably and systematically over a wide range, while resources are preserved for focused validation of only the promising candidates in the second-stage. This approach, however, would be limited to intensification based on “known” unit operations, unless the PI process synthesis/design is considered at a lower level of aggregation, namely the phenomena level. That is, the model-based synthesis method must employ models at lower levels of aggregation and through combination rules for phenomena, generate (synthesize) new intensified unit operations. An efficient solution procedure for the synthesis problem is needed to tackle the potentially large number of options that would be obtained. Here, established procedures for computer aided molecular design is adopted since combination of phenomena to form unit operations with desired objectives is, in principle, similar to combining atoms to form molecules with desired properties. The concept of the phenomena-based synthesis/design method for process intensification will be presented together with illustrations of its application through case studies. Aspects such as need for models, computer-aided framework, the work-flow and data-flow for the PI synthesis method will also be highlighted.