Optimal design of an isotope distillation unit NLP solution for minimum tritium inventory

Abstract

Abstract In the frame of thermonuclear fusion, the cryogenic distillation system which is a subsystem of the fusion fuel processing system, is required for separating hydrogen isotopes and for recycling protium, deuterium and tritium to the fusion torus. This cryogenic distillation process, mainly composed of columns, has to be designed under safety constraints. The criterion that has been accepted is the tritium inventory in the distillation cascade. Our task is thus to design the cascade of distillation columns under minimum of tritium inventory criterion and purity and production constraints. Considering that the structure (arrangment of columns) of the cascade is fixed, the problem can be solved with a NLP procedure coupled to a shortcut method. Here are discussed some relevant aspects of the NLP formulation problem, and results concerning an actual cascade are presented. Perspectives concern the implementation of a MINLP procedure for optimal design in structure and operating conditions of an isotope separation process.