Simulation of Undeclared Operations at Gas Centrifuge Enrichment Plants

A model was developed to simulate enrichment plant operations that deviate from optimal cascade design and operate contrary to declarations. A base case was determined from the MSTAR’12 code, which assumes an ideal cascade to optimize separative work. A number of scenarios were considered for undeclared operation, such as using enriched feed (batch recycle), side streams, and other modifications to standard operations. These scenarios do not involve retooling the cascade piping, but rather they assume that machine connections are unperturbed and that all stages remain as designed. Modeling these processes with a code that assumes optimization of separative work could be misleading, as the undeclared operations will undoubtedly be less efficient. This work has developed models to describe such undeclared operations based on the assumption that stage cuts and numbers of machines in each stage are unchanged. When compared to calculations using ideal cascade models, substantial differences are noted for some scenarios. For example, consider a cascade designed to produce 3% enrichment from natural feed. Three successive replications of batch recycle yield 72% enrichment under the ideal model but 89% when the realistic model is applied. These examples are developed for the Rome centrifuge, a hypothetical machine whose description is in the public domain. Using simulations, deviations from operational parameters (flows and assays) in undeclared scenarios can be compared with parameters under declared operations. Several scenarios are discussed with the intention of improving unattended monitoring capabilities.