Conceptual design of inherently saferprocesses by genetic algorithms and case-based reasoning

An approach for synthesis of inherently safer chemical processes is presented. Process synthesis can be considered as an optimization task. Since the synthesis model is uncontinuous, ordinary optimization methods cannot be used, but a genetic algorithm is employed instead. In the genetic algorithm the structure of the process is represented as a string of integers, which describes the operations required and how they are connected. An inherent safety index is used as an objective function (fitness) in the genetic algorithm. To evaluate the process structure subindex a database of existing good and bad designs was collected from design recommendations and accident reports. By this way the experience from earlier designs can be reused in process synthesis and the probability of making again the same design mistakes is decreased.