Probabilistic design approach to build the litheness in an integrated process scheme

摘要

Flexibility in manufacturing processes is essential not only to handle a variety of feed stocks or to produce multiple product grades but also to accommodate process or catalyst inventions to improve productivity or quality. While designing in litheness we should ensure the robustness of the process output. In this paper, through a case study of biodiesel manufacturing, we summarize a novel technique which uses a probabilistic design approach using meta-models to evaluate the process intensification/integration scheme in a bio-refinery context. In our methodology first we develop individual process models for a number of processes intensification alternatives where waste or non-edible vegetable oil are converted to biodiesel. These alternatives include the use of reactive distillation and a supercritical process. We integrate these process intensification schemes along with other necessary process steps to establish a plausible manufacturing process. Simultaneously we formulate high fidelity meta-models to describe the various responses of the reactive distillation column w.r.t the changes in the process conditions (or decision variables) in a given design space as an example. This simplification allows us to setup a probabilistic design using commercially available meta-heuristic optimization tools.